0

Chapter 3 – Inequality and Social Progress

1

Coordinating Lead Authors:[1] Giovanni Andrea Cornia, Rebeca Grynspan, Stephan Klasen, Luis-Felipe Lopez-Calva, Nora Lustig

2

Lead Authors:[2] Augustin Fosu, Sripad Motiram, Flora Myamba, Andreas Peichl, Sanjay Reddy, Eldar Shafir, Ana Sojo, Ingrid Woolard

3

Contributing Authors:[3] Shai Davidai, Michael Förster, Rahul Lahoti, Rainer Thiele,

4

Word count: [72,648]

5

Abstract: [Abstract 200 words]

6

Summary

7

Inequality has received increasing prominence in social science debates over the past 20 years.  The attention on inequality issues focused in the 1980s and 1990s particularly on industrialized countries.  This was largely related to the sharply rising income inequality observed in many OECD countries since the early 1980s which was well-documented using available data. In industrialized countries, more recently debates have also arisen concerning inequality in education and health inequality as well as wealth inequality.  New information on the top of the income and wealth distribution (which is usually poorly captured in standard household survey instruments) extended the debate to including top incomes and wealth holders in the overall assessment of inequality.

8

In developing countries, there has been a long-standing interest and debate in inequality issues, relating, for example, to Kuznet's famous inverse U hypothesis (or rising and then inequality in the development process).  Until the early 1990s, however, comparable and consistent data to assess inequality trends across countries and over time were largely lacking.  With the publication of many more, more reliable, and more frequent household surveys, this picture changed dramatically, allowing inequality for the first time a detailed assessment of levels and trends in.  Analyses of inequality in developing countries showed that within country income inequality was rising in many parts of the developing world, esp. since the 1980s.  Similarly, some literature has emerged on health and educational inequality. 

9

With inequality data available for all groups of countries, the 1990s and 2000s also saw the emergence of a debate on global income inequality, which is, of course, related to both inequality within as well as between countries. The rising prominence of inequality can also be seen in the recently concluded negotiations on the global Sustainable Development Goals which, in contrast to their predecessor, the Millennium Development Goals, now includes a specific goal to reduce 'inequality within and among countries.'  Similarly, much greater attention is placed on inequality at international organizations such as the World Bank, the IMF, or the OECD. 

10

This chapter will survey the social science literature on inequality as it has emerged in the past three decades.  While we will briefly mention global inequality, the chapter will largely focus on inequality within countries.  Inequality between countries (esp. in the income dimension) is largely an issue of differential economic growth rates between countries, a subject covered in detail in chapter 4.  We will discuss inequalities in many dimensions, but inevitably a greater emphasis will be on income inequality, simply because there is more information on this dimension and much of the literature has focused on it.  The chapter will not only dwell on measures of inequality, its trends and determinants, but also include a detailed discussion on the impact inequality has on the well-being of people, both now as well as in the future (and thus affect, if you will, social progress).  It will also study deep drivers of inequality before discussing policy issues that could affect inequality. 

11

The chapter is organized as follows: After presenting the key messages below in the remainder of this introduction, section 2 will discuss different concepts and measures of inequality.  Section 3 will investigate the question why inequality matters, while section 4 will provide a brief overview of global inequality and its trends and drivers.  Section 5 will assess trends in within country inequality, section 6 will summarize the literature on decomposing changes in inequality into different components which can be seen as immediate drivers of inequality change.  Section 7 will examine deep drivers of inequality, while section 8 will focus on policy issues affecting inequality.   

12

1. Introduction

13

There are many (often interacting) inequalities and different forms of inequality to consider.  While much of the literature has focused on economic inequality, usually measured by income, there are many more dimensions to consider.  They include inequality in freedoms, in opportunities, in central capabilities such as the ability to be healthy, educated or socially integrated, in political power, social standing, among many others.  In addition, one can distinguish between inequality between households, and within households (e.g. by age and gender), as well as between inequalities between individuals/households (vertical inequality) and inequality between social groups (horizontal inequality).  One needs to also distinguish static from inter-temporal assessments of inequality, including the issue of mobility. When it comes to measuring inequality, one needs to differentiate between uni-dimensional and multidimensional measures of inequality, objective versus subjective measures, absolute versus relative inequality, inequality versus polarization, and different indicators of inequality which emphasize different aspects.  Thus, much of the heterogeneity in findings in the literature relate to the different concepts and dimensions of inequality that are being analyzed. 

14

1.1 Economic inequality

15

High income and wealth inequality retards social progress (intrinsically and instrumentally) in terms of improvements in welfare for people and promotion of social cohesion.  While some social and economic differentiation is tolerable and even desirable, high inequality in resources, opportunities, or capabilities runs counter to most theories of justice who see such inequalities as inherently 'unfair'.  In addition, inequality reduces overall well-being in a society, is perceived to be problematic by the vast majority of most societies, is particularly a burden on the well-being of the poor, and is associated with greater intergenerational immobility.  Instrumentally, high inequality increases poverty, it lowers the impact of economic growth on poverty reduction, promotes social conflict, disaffection, and protest, affects behavior that can trap poor people in a state of poverty, and economic inequalities often promote social and political inequality, as well as inequalities in health and education.  High levels of inequality are also associated with lower subsequent economic growth, while there is no consensus (and robust results) on the impact of income redistribution on economic growth.  

16

1.3 Trends in inequality

17

Within-country income inequality has been rising in many countries since the 1980s and now contributes a significantly larger share of global inequality (while between-country inequality stabilized or has fallen somewhat since the 1980s, depending on definitions, concepts, and data); since the late 1990s, trends in within-country income inequality have been more heterogeneous.  In particular, in Asia they tend to have stabilized at higher levels (or rising more slowly in South Asia), in Latin America they have reduced substantially (from very high levels), and in Africa trends differ greatly by country.  In transition countries, inequality levels have also stabilized at high levels (or risen more slowly), while in OECD countries, changes in inequality were generally smaller since 2000 than before.  Findings on inequality in non-income dimensions point generally to smaller levels of inequality, falling global inequality in health and education (particularly since about 2005), and substantial heterogeneity in inequality trends at the national level, often also related to average levels of health and education achievements.  It is important to note that there are considerable uncertainties and debates about these findings regarding levels and trends in inequality.  While they partly relate to different notions of inequality being studied, they also relate to poor quality, irregular, incomplete, inconsistent, and incomparable data continue to plague debates about inequality.  These data gaps and problems exist in all countries, but are particularly severe in developing countries, and are particularly problematic in capturing the very rich and the very poor and some forms of income and wealth.

18

1.4 Causes of inequality

19

Key drivers and causes of inequality (between and within countries and groups) can be distinguished between deep drivers and more immediate determinants.  The drivers and determinants are often country-specific so that only rather general statements can be made.  This already suggests that no overarching trend (such as globalization or technological change) can account for all or even most of inequality dynamics, but country-specific situations, policies, and institutions matter.  In OECD countries, increases in inequality are immediately related to a sharply widening earnings distribution, with the earnings and employment prospects of poorly educated people (particularly of men) stagnating or falling while earnings and employment prospects at the top of the education and income distribution increased substantially.  More fundamental causes relate to skill-biased technological change, increased trade in labor-intensive manufacturing products with emerging countries, the rise of top earnings in a growing financial sector, labor market policies (esp. regarding unionization, minimum wages, and low wage sectors), and the declining redistributive role of the state.  Difference in the importance of these factors can contribute to explaining the substantial variation in inequality trends within OECD countries.  In developing countries, inequality trends are affected by the earnings distribution of employees, but also strongly affected by regional inequality and rural-urban gaps. Increased trade with rich countries has not (as expected) led to reduced inequality, but often served to increase it, related to the regional concentration of export-oriented industries as well as the rising skill-intensity of trade in developing countries that has come with outsourcing and the splitting up of global value chains.  In addition, education, health and fiscal policies in developing countries have shaped inequality and contributed to rising inequality in the 1980s and early 1990s dominated by structural adjustment policies.  The substantial decline of inequality in Latin America was related to favorable economic conditions and a set of policies, mostly enacted by left governments, on taxes and fiscal redistribution, labor markets, and social protection.  Many of these trends in inequality are related to more deep-rooted factors that change more slowly.  They include strong path dependency and reproduction of inequality, strong social stratification in many societies causing persistent group-based inequalities, deep-seated norms and preferences regarding redistribution, the role of social movements, the strong link between economic and political inequality, and demographic dynamics with large inequality implications. 

20

1.5 Policy and politics issues  

21

While global economic forces (such as increased trade, technological change, capital flows, and migration) affect inequality, domestic policies, institutions, and circumstances also play an important role in affecting inequality.  Also, global economic forces, international policies, and national policies can interact to influence inequality.  Possible country policy agendas to affect inequality also depend greatly on country circumstances and the nature and dynamics of inequality.  In general there is substantial scope for affecting inequality by improving pro-poor investments in health, education, infrastructure, capital, and assets, by increasing the tax/GDP ratio which particularly in many developing countries is very low through the use of broad-based consumption taxes, resource taxes, among others, by increasing the progressivity of the tax/transfer system, by labor market policies to improve employment and earnings of the low end of the labor market, and by social protection policies that promote inclusion of the poorer end of the distribution.  Horizontal inequality can also be tackled through anti-discrimination as well as preferences for disadvantaged groups, although the record of such policies is mixed.  Particularly in rich countries, there is also substantial scope for policy to affect top incomes, related to tax policies, regulations of earnings, and transparency of earnings.  The potential for implementing such policies will depend on the domestic political economy which is affected by the size and voice of the middle-class, the nature of political alliances, and the role of popular and social movements, among others.  The scope of international policies to affect within-country inequality is generally more limited, with trade, aid, migration, and intellectual property policies more likely to affect growth than inequality.  But international cooperation on combating tax avoidance and evasion by wealthy individuals and multinational companies, international coordination in regulating financial markets, as well as more orderly and less costly international migration regimes could all contribute to lowering inequality. 

22

2. Different concepts and measures of inequality

23

2.1 Inequality of what? Outcomes, opportunities, capabilities, freedoms, primary goods?

24

In his influential work Inequality Reexamined, Amartya Sen (1992) argues that equality has been assessed by comparing persons on various aspects, and the measurement of inequality and our assessment of it depend upon the particular aspect that we have chosen. Sen terms these aspects as “focal variables” and claims that even positions like libertarianism, which appear to be inegalitarian at first glance, are concerned with equality on some dimension (e.g. equal liberties). Focal variables themselves may be a combination of multiple entities (e.g. different kinds of freedoms put together) and Sen makes a case for distinguishing such “internal plurality” from differences among various focal variables. Human beings differ in terms of their characteristics (e.g. gender, race) and in terms of their endowments, ownership etc. Given these differences, privileging equality in one domain may result in tolerating inequality in another, e.g. equality in incomes may result in inequality in well-being given that different individuals may have different needs and since these incomes may translate into different bundles of commodities. Sen argues persuasively that, in light of these considerations, the question of what the focal variable should be (“equality of what?”), is important. He then proceeds to evaluate different philosophical traditions (utilitarianism, Rawlsian justice) to make a case for focusing on the capabilities of individuals to achieve desirable functionings.

25

In practice, the choice of focal variables has been dictated by several considerations, including the imperatives of a particular context/policy, availability of data, ease of measurement etc. A disproportionate focus has been given to inequality among individuals and households – termed as vertical inequality – and this literature has largely focused upon income and consumption-based inequality (Stewart 2002, 2010). Given this, the literature on the measurement of income and consumption-based inequality is the most well-developed one. Focus has also been given to wealth, and studies have attempted to measure wealth inequality at both the individual and household levels (e.g. Davies and Shorrocks 2000; Jayadev et al. 2007). The measurement of wealth inequality raises certain issues and problems that do not arise in the context of income and consumption, e.g. the wealth of individuals can be negative, which precludes the use of indices like Theil; in some contexts, wealth is concentrated in assets like buildings and jewelry, which are notoriously difficult to capture (see Jayadev et al. (2007), which presents a discussion).

26

Two other inequalities that have received some attention have been in the domains of education and health. Although techniques and concepts used in the measurement of income inequality have been applied here, the issues, nature of the data, limitations of data are different (see e.g. O‘Donnell et al. (2008) on health, and Ferreira and Gignoux (2011a) on education). To illustrate, the Indian National Sample Surveys on consumption expenditure and employment, which are very widely used, contain only categories of education (e.g. primary, secondary etc.) and not years of education. Hence, unlike in the case of income, consumption and wealth, one is dealing with a variable that is not continuous and bounded. Similarly, some of the health data is not only discontinuous, but also subjective and raises specific concerns (see Agrawal (2010) for a discussion).

27

One domain that has received considerable attention in recent times is that of opportunities. There has been substantial interest in the nature and extent of inequality of opportunity, particularly in the space of incomes, i.e. inequality of opportunity as a driver of incomes and earnings. The main inspiration for this agenda has been the work of John Roemer (e.g. 1998, 2008). Broadly speaking, Roemer has argued that what individuals achieve depends upon two sets of factors: those beyond their control (circumstances), and their efforts. It is unjust to hold individuals responsible for factors beyond their control (i.e. their circumstances) and societies should therefore try to reduce these disadvantages to the extent possible (or compensate individuals for suffering from such inequalities in circumstances). This formulation is simple, intuitively appealing, and (most importantly) amenable to empirical analysis. As a result, a large and growing body of literature has emerged that has tried to estimate the inequality of opportunity in various countries (e.g. de Barros et al. 2009; Checchi and Perragine 2010; Ferreira and Gignoux 2011b; Singh 2011). Some of the circumstance variables that have been identified are parental education, caste, gender, religion, rural-urban location etc. Both parametric (regression-based) and non-parametric approaches have been suggested (see Singh (2011) and Kanbur and Wagstaff (2014) for details).[4]           

28

Even as the empirical literature on inequality of opportunity is gaining ground, there are serious philosophical questions and problems that have been unresolved. Kanbur and Wagstaff (2014) present a discussion of these issues. Briefly put, some of these concern the difficulties in separating justifiable and unjustifiable sources of inequality, the treatment of luck, accounting for talent, and inequality of opportunity of children. To take one example, studies hold people responsible for their luck. This is problematic because one has to distinguish between “brute luck” (e.g. the case of someone who is involved in a road accident for no fault of his/her own) and “option luck” (e.g. the loss in the stock market to a calculating investor) – an argument can be made that it is unfair to treat both of these in a similar manner, and hold people responsible for the former. Moreover, the treatment of absolute poverty needs to be addressed better – it is morally reasonable to argue that the concerns of the truly destitute need to be addressed even if their destitution is a result of their actions and/or risks that they have taken.  There are also questions of how to interpret differences in information access or cognitive limitations that might affect outcomes and might erroneously be considered effort. 

29

Despite the above limitations, the agenda of inequality of opportunity is promising and has the potential for influencing policies that promote equity. Circumstance variables could include (and have included) membership of groups, and in this sense, the literature on inequality of opportunity has shed light on differences among groups. However, philosophically, and to a large extent in its empirical application, the thrust of this literature has been on individuals. This is also true of the welfare economics tradition inspired by Sen’s capability approach.  The question of group-based, or horizontal, inequality is addressed in more detail below.

30

2.2. National inequality, global inequality, regional/local inequality

31

The frame of reference from which to study inequality matters greatly for measurement, assessment of the relevance of inequality, and policy issues.  Inequality in whichever dimension (e.g. income, wealth, education, health, etc.) can be studied in a very local context, within a small geographic area, a particular group of people, or even within households (for example, when studying intrahousehold gender inequality, e.g. Klasen, 2007; Klasen and Lahoti, 2016).  At the other extreme, one can study global inequality.  For people's well-being, behavior, and sense of equal treatment, local inequality may matter more than global inequality; but improved access to information may increase the relevance of national and global inequality for one's own assessment of well-being (e.g. Lohmann, 2015); for much of policy-making which tends to be done at the national level, however, within-country inequality is of particular importance.  And most analyses of inequality focus on national level analyses.  But when studying within-country inequality, inequality between groups or regions is often considered to be an important dimension to consider as well (see discussion below on horizontal vs. vertical inequality).  Moreover, to assess perceptions of inequality, it is often important to discern the relevant reference group that people use for comparisons which may differ according to context (Boyce et al. 2010;  Stutzer, 2004).

32

When considering global inequality, Milanovic's (2005) influential classification into three types of global inequality is very useful.  The first type refers to inequality between countries, treating each country as a single, equal-weighted observation.  Such an analysis is useful to study inequalities in living standards across countries, differentials in economic power, and international inequality from a policy perspective where countries are usually the unit of observation and action (e.g. at UN fora).  From a policy perspective, the only way to reduce global inequality using this concept in, say, the income dimension, is for the poorer countries of the world to grow faster than richer ones.  So essentially this is an issue of differential growth rates which is taken up in detail in Chapter 4. 

33

At the same time, this approach to focus on inter-country inequality gives no impression about inequality between people of the world as tiny Mauritius is treated the same as populous China.  Type 2 inequality addresses this and weights the mean achievement of each country (e.g. per capita income) by population and then considers global inequality in this population-weighted terms.  This more reflects global inequality since now each person receives the same weight in the assessment of inequality.  But it still assumes that within each country there is no inequality and that higher growth of poorer countries is the only way to reduce this global inequality.  Type 3 addresses this last point and considers inequality between world citizens.  In that concept, actual achievements of all citizens are directly compared and the countries do not play an essential role here (other than that much of the data will come from country-level analyses and will have to be merged to create a global analysis).  This measures inequality among world citizens most directly.  This is probably the best way to capture inequality among world citizens and it closely aligns with cosmopolitan positions on global justice.

34

However, on its own, it is a statistic that is hard to interpret and to derive clear policy conclusions from as it reflects inequality within countries as well as between countries, and gives more weight to populous countries.  Of particular use can therefore also be ways that decompose global inequality into within and between country inequality; some of the measures discussed below allow such decompositions.  An important and robust empirical finding on measuring global inequality in this way has been that over the past 30 years the contribution of between-country inequality to global inequality, while still the largest component, has been falling while the contribution of within-country inequality has been rising (e.g. Bourguignon, 2015; Klasen et al. 2016).  See, for example, Figure 3.1 below. 

35

Figure 3.1: Within and Between Country Income Inequality of Global Inequality

36

Source: Global Income and Consumption Database.

37

Note: This chart is primarily for illustrative purposes on within and between country inequality. As discussed below, there are considerable uncertainties on exact global inequality trends, although most assessments agree that the within share of global inequality has been rising in the past 20 years.The data here use income as the concept and adjusts expenditure surveys through a regression-based adjustment to create comparable income inequality statistics.

38

2.3. Horizontal versus vertical inequality (between groups, include rural-urban)

39

Horizontal inequality refers to inequality among groups, usually cultural or identity groups, e.g. ethnic groups, races. This has to be contrasted with vertical inequality which concerns inequality among individuals or households, and which has largely focused on income or consumption (Stewart 2010). Horizontal inequality has been relatively under-researched and neglected. This is in line with traditions in welfare economics, and the policies that have been influenced by them, that have treated the individual as the domain of analysis and welfare (Stewart 2002). While not denying the importance of vertical inequalities, Stewart (2002) has persuasively demonstrated why it is important to understand and address inequalities among groups. Severe inequalities among groups threaten social stability, resulting in violent conflict and civil war. At a more mundane level, individuals belonging to disadvantaged groups may further their own well-being, and the well-being of the society, in a sub-optimal manner. Also, it may be easier to address deep social problems (e.g. unemployment, poverty) by focusing upon groups which carry a disproportionate burden of these problems. For example, it is commonly found to be the case that absolute poverty levels are substantially higher in rural than in urban areas in most developing countries, suggesting that it will be critical to poverty reduction efforts there (e.g. Ravallion et al. 2007).

40

While these are indirect reasons, Akerlof and Kranton (2000) and others have argued that group identity is important for individuals and directly affects them. Giving serious consideration to the welfare of underprivileged groups could, therefore, enhance the welfare of individuals belonging to them. While the coexistence of different groups could lead to problems, this is not always the case, and there are many examples of diverse, but peaceful societies. In fact, all societies are diverse in some dimensions and in many cases diversity is a source of enhanced well-being.  Also, people have many identities and can therefore belong to many different groups.  It is the combination of diversity and severe inequality among rigidly defined groups that can create the conditions for conflict.

41

Lastly, the study of horizontal inequalities has close linkages to inequality of opportunities (discussed above).  Often 'circumstances' that cause inequality of opportunities relate to group membership (e.g. ethnicity, race, gender) so that horizontal inequality is a form of inequality of opportunities if circumstances are defined in such a way.  

42

The measurement and conceptualization of horizontal inequality raises several complications since group membership could be fluid and inequality among groups (like among individuals and households) could exist on multiple dimensions. While the literature has acknowledged the former issue, it has taken the group definitions in a particular context as given. Stewart (2010) identifies economic, social, political and cultural status as the relevant broad dimensions. Each dimension in turn comprises of several elements, e.g. economic: asset ownership; social: education; political: control over government; and cultural status: recognition of cultural practices. For empirical purposes, it has been suggested (Stewart 2002) that it is useful to arrive at an index for each element/dimension, but not aggregate over the elements. This is because the differences across elements may be interesting and have consequences, e.g. a situation of consistency where one group is uniformly disadvantaged could be different from one where it is disadvantaged on only some dimensions. It is also worth pointing out that it is not just “objective” horizontal inequalities that matter – perceptions play an important role too.

43

The empirical literature examining the impact of horizontal inequalities has used two different methodologies: case studies (see Stewart (2008) for examples) and statistical/econometric analysis within and across countries (e.g. Østby 2008, Wimmer et al. 2009, Brown 2010, Mancini 2008, Motiram and Sarma 2014. See Stewart (2010) for some other references.). The latter studies suffer from severe data limitations. Stewart (2010) summarizes the broad conclusion that emerges from this agenda that horizontal inequalities: “increase the risk of violent conflict, especially when they are consistent in the economic, social, political and cultural status spheres.” She also argues that in societies with considerable horizontal inequalities, policies should be put in place to ameliorate them. She suggests three different approaches: direct (e.g. job quotas for specific groups), indirect (e.g. anti-discrimination) and integrationist (e.g. nation-building).

44

2.4. Individual versus household-based inequality

45

As stated above, much of the welfare economics and other social science literature on inequality studies vertical inequality using income as the focal variable.And inequality (and income poverty) is usually study at the level of households, i.e. inequality between households is measured.This choice is largely dictated by the available data, which measures total resources (income or consumption) a household has at its disposal and then only adjusts these total resources by household size or some equivalence scales which take into account different needs of members (e.g. children need fewer resources than adults) and economies of scale within the household (e.g. Buhmann et al. 1988).The construction of appropriate scales is controversial, can affect levels and trends of inequality, and there is no consensus on the right scales (Deaton and Zaidi, 2002), but some scales are quite commonly used.[5]  An important implicit assumption of these household-based analyses is that distribution within households is equal (or according to needs in the case of an equivalence scale).  To the extent that the distribution of resources within households is unequal, this will underestimate inequality, and also bias poverty assessments (Haddad and Kanbur, 1990).  In the income/consumption dimension, it is very difficult to move to an individual assessment as incomes are shared within households and a considerable share of consumption expenditures is used to finance household-specific public goods (e.g. housing, durable goods, utilities) which cannot be easily ascribed to individual members (Klasen and Lahoti, 2016).  As a result, it is very difficult to discern the role of gender inequality or inequality between age groups on aggregate poverty and inequality, as the intrahousehold dimension is neglected.  Using non-income measures, however, there is more scope for an individual assessment (Klasen and Lahoti, 2016).

46

Assessments of wage or earnings inequality, which focus on inequality in the labor market, are often based on individual data.The link between levels and trends in wage, earnings, and overall household income inequality is not as close as one might think since it strongly depends on how earners are distributed across households which, in turn, is affected by the distribution of unemployment or of female economic participation across the income distribution as well as marriage and household formation patterns (e.g. Gottschalk and Danziger, 2005).

47

2.5. Functional inequality

48

Another way to study inequality (esp. income inequality) is to link it to the distribution of production factors and their returns.Since capital (including land assets) and labor (including human capital) are the two factors of production, the share of incomes accruing to capital and labor can also be one way to describe inequality.Particularly examining trends in these shares can reveal patterns of inequality change and esp. link inequality changes to the ability of owners of capital and labor to increase their claim on levels and gains in income.The relationship between the capital and labor shares and vertical inequality is close but far from perfect.For example, Piketty (2013) shows that the capital share has been rising in many countries but also that the labor earnings of some individuals (e.g. in finance, senior management, and superstars) have also risen and both combined contributed to increased income inequality among households.Measuring these shares is not easy, particularly in developing countries where self-employment in agricultural and the informal sector employs a large share of the population and income from self-employment combines capital income (from land and assets) with labor income (Trapp, 2015).

49

2.6. Uni-dimensional versus multidimensional inequality

50

Most of the standard literature on welfare or inequality focuses on one dimension (often income or consumption, but sometimes also non-income achievements, e.g. Grimm et al. 2010).  Of course, income or consumption can be seen as a multidimensional well-being indicator with market (or shadow) prices being the weights (Ravallion, 2011).  But the reliability of these measures relies on highly unrealistic assumptions about the completeness and functioning of markets where prices reflect the accurate weights of all welfare-relevant goods and services, and additionally suffers from the conceptual problems of utilitarianism as the central welfare concept (e.g. Sen, 1998). 

51

In many conceptions of well-being, including for example Sen's capability approach (Sen, 1998), well-being itself is, however, seen as a multidimensional concept that cannot easily be reduced to a single dimension (or index).  For example, the capabilities to be educated, healthy, and socially integrated, may all be very valuable but cannot be reduced to one dimension (nor can prices or other weights be easily assigned to them).  If those capabilities are treated as the ultimate well-being outcomes, income will only be a highly imperfect proxy to capture these capabilities, related to the inherent heterogeneity of humans in their ability to translate incomes into capabilities, and the externality and public good aspects of health and education (where provision of quality health and education services depends more on public action than private incomes, Dreze and Sen, 1989).

52

 Also, the overlapping disadvantage of people suffering deprivations in several dimensions can be of particular interest as these might point to more structural and deep-seated inequalities and may also relate to horizontal inequality.  Studying multidimensional inequality (or multidimensional poverty) might be a way to uncover those overlapping disadvantages (and advantages, see Ferreira and Lugo, 2012).  

53

When it comes to the measurement of multidimensional inequality, however, a number of complex conceptual and empirical difficulties (e.g. Aaberge and Brandolini, 2014).  They relate, among others, whether different well-being dimensions can be considered as substitutes or complements, whether transfers of compensation between dimensions is possible, whether one should first measure inequality within a dimension across people and then aggregate across dimensions or aggregate dimensions within people and then aggregate across people, whether one has continuous or discrete variables at one's disposal.  These difficulties arise in addition to the 'usual' questions in multidimensional well-being measurement, such as the choice of dimensions and their relative weights (e.g. Aaberge and Brandolini, 2014; Bosmans et al. 2015, Mueller and Trannoy, 2012).  As a result, there is no consensus to date on appropriate summary measures of inequality. 

54

At the time, it is of course possible to study multidimensional inequality without such a summary measure.  One the one hand, one can rely on a dashboard approach and study inequality in each dimension separately using well-known uni-dimensional inequality measures (see Ravallion, 2011.  While such an assessment provides a more complete assessment of inequality than reliance on a single dimension, such an approach does not allow for a complete ordering or ranking of multidimensional inequality of the units considered (e.g. groups, countries); nor does it say anything about overlapping disadvantages as emphasized by Ferreira and Lugo (2012).  One approach to address this problem are dominance approaches where it is studied whether one multidimensional distribution dominates another one, i.e. is certainly more unequal based on the assumption of the dominance criteria.  To facilitate such comparisons, compensation and transfers can be conceptualized to help compare multidimensional distributions (e.g. Aaberge and Brandolini, 2014).  Another approach used standard inequality measures, dimension by dimension, and then aggregates them.  An example is the Inequality-Adjusted Human Development Index (UNDP, 2010), which adjusts achievements in the three human development dimensions (education, health, incomes) by a penalty for inequality within these dimensions.  The gap between the HDI, which measures average achievements, and the IHDI, can then be seen as a measure of multidimensional inequality.   While this nicely illustrates the extent of inequality within dimensions, it cannot say anything about overlapping disadvantages suffered by individuals in these dimensions (because the dimensions are considered separate).  An alternative proposal has been by Harttgen and Klasen (2012) to first create an HDI at the household level and then study inequality in that HDI across people (and across countries).  There, overlapping disadvantages are explicitly considered, but there are serious data issues and the aggregation of overlapping disadvantage across dimensions is based on a range of debatable normative assumptions.  

55

To conclude, the literature on multidimensional inequality is still in its infancy.  Clearly studying multidimensional inequality is important, also to consider overlapping disadvantages and advantages, but the best ways to do this are still being debated.

56

2.7. Similarity and differences in inequality versus poverty measurement

57

 There are important similarities and differences between a focus on poverty versus a focus on inequality.  Conceptually, the difference is strongest when poverty is seen as an absolute concept (with a poverty line invariant over time and space).  Then poverty is concerned with the absolutely worst off such as the international income poverty line promoted by the World Bank (currently at $1.90 in 2011PPPs, see Lustig and Silber, 2016; Ferreira et al. 2016, Klasen et al. 2016) or the poverty line in the USA (Citro and Michael, 1995), while inequality is examining the entire distribution of an outcome.  It is entirely possible for absolute poverty to fall while inequality is going up.  Sometimes, however, poverty is measured in relative terms (with a poverty line defined as the share of median income in a country and year, as done in the European Union).  Then increasing the income share of the poor is the only way to reduce poverty.  But even then, relative poverty is concerned with the situation of the poor relative to the middle of the distribution, but does not consider the entire distribution, in particular also not the situation of the well-off in its assessment. 

58

A concern for poverty over inequality derives, both in philosophical conceptions as well as practical applications, generally from a priority to be given to the worst-off in society.  This is how the World Bank, for example, justified its focus on absolute poverty in past decades (e.g. World Bank, 1990, 2000).  And many programs are 'targeted' to the poor to give them priority access to their benefits.  This targeting can have economic, political and administrative costs as discussed by Atkinson (2004 and Klasen and Lange (2016).  Sometimes, policies that are concerned with reducing overall inequality, also focus on the poorest and thus try to address poverty reduction as well.  And there are other linkages between inequality and poverty reduction which will be reviewed below. 

59

A concern for inequality, in contrast, is usually related to a concern for justice or fairness of an economic or social arrangement.  And unfairness can not only be related to the lower part but also to the upper part of the distribution.  Different arguments advanced in this area are reviewed below. 

60

A concern for inequality has received increasing attention in international policy debates.  While the Millennium Development Goals only included targets on absolute poverty reduction, the successor Sustainable Development Goals include a separate goal to reduce inequality between and within countries.  And the World Bank has recently added a shared prosperity goal (Basu, 2013) which tracks and seeks to increase the growth rate of incomes of the poorest 40%, which considers inequality (albeit in a somewhat limited form, see Klasen et al. 2016). 

61

2.8. Income versus wealth inequality

62

The vast majority of analyses focus on inequality in current incomes or current expenditures.  This is largely related to data availability and quality issues.  Wealth data are collected less frequently in household surveys and there are difficult measurement issues.  For example, wealth can include pension wealth in a public or private pension system which depends on complex pension rules (and may be uncertain), it can include wealth in capital accumulating life insurances, it can include assets that are difficult to value due to absence of liquid markets for them (e.g. shares in privately-held companies), among many other issues.  Lastly, since many households have 0 or even negative wealth, standard inequality measures that require positive incomes for all (such as the Theil measure or mean log deviation) cannot be used.  As a result, most empirical analyses, esp. those comparing inequality across space and time, focus on income inequality.

63

But this perspective is limiting in several dimensions.  First, it does not consider dynamic issues such as shocks or life-cycle movements of incomes.  Inequality in current incomes may be larger than in lifetime incomes if, for example, today's poor include poor students who will end up earning much more over their lifetime, or those who suffered a negative shock this year.  Longer-term inequality relates also, in important ways, to differences in wealth.  Wealth allows households to smooth over temporary shocks, households with wealth can use that wealth as collateral for credit for investments, and households can draw on income streams from their wealth as a long-term source of income.  In fact, income streams from wealth form a large part of incomes of the richest households in advanced countries (Alvaredo et al. 2013).  Wealth inequality is everywhere much more unequal than income inequality.  In fact, in most countries, the poorest 60% of the population own little wealth at all (and many may be in debt), and most wealth is concentrated in the top decile (and sometimes even the top percentile holds the majority of wealth, Alvaredo et al. 2013).  Also across countries, most of the world's wealth is concentrated among the very rich in rich countries while many poor countries have only a tiny share of the global wealth distribution at their disposal (Devies et al. 2008).  Thus it is important to consider wealth inequality as an important driver of income inequality, and policies that reduce wealth inequality (such as inheritance taxes or taxes on capital income, or direct wealth taxes) are likely to affect income inequality.  

64

2.9. Absolute versus relative inequality

65

Relative inequality is the most commonly used notion of inequality. In fact, in many academic and policy discussions, it is the only kind of inequality that has received attention. Many studies of income inequality use the Relative Gini, usually without the “relative” qualifier. Measures of relative inequality satisfy the property of “scale invariance” wherein inequality is unaffected if the distribution is scaled by the same non-zero factor. We will use the example of income inequality to illustrate this idea (and other ideas below) - if all the incomes are doubled (or multiplied by an exchange rate to express them in a different currency), then inequality does not change. This property makes measurement and comparison convenient since the measure is unitless.

66

In two seminal articles, Kolm (1976 a, b) argued that this convenience is obtained at a cost: “convenience could not be an alibi for endorsing justice”. Note, using the example above, when all the incomes double, the difference between the incomes of the rich and the poor increases. Kolm characterized relative measures as “rightist” and contrasted them with “leftist” measures which satisfy the property that inequality is unaffected if all the incomes increase by the same absolute amount.Kolm argued that in his experience, this property (which we can refer to as “translation invariance”) conforms to the notions that people have about inequality. An example of an absolute measure is the Absolute Gini, which is the Relative Gini multiplied by the mean. Is it possible to arrive at a trade-off between convenience and ethics, i.e. between relative and absolute measures? “Intermediate” or “Centrist” measures strive to do this by increasing when all the incomes are scaled up by the same factor and decreasing when all the incomes increase by the same absolute amount. An example of such a measure is the Intermediate Gini, which is the product of Relative and Absolute Ginis. Another desirable property that this measure satisfies is “unit consistency”, which ensures that two income distributions are ranked in the same manner irrespective of the units, i.e. the ranking of two distributions is unaffected if they are scaled by the same non-zero factor, e.g. if they are doubled. See Zheng (2006) for a description of this property and Subramanian (2013), Subramanian and Jayaraj (2013) for some more details.

67

As we mentioned above, studies that have used measures other than relative ones have been sparse. Subramanian and Jayaraj (2013) which focuses upon India, is a recent example. The Indian case is quite interesting because India has been one of the fastest growing economies in the world since the early 1990s, and there is considerable interest in the impact of growth on inequality. Subramanian and Jayaraj (2013) examine inequality in consumption and wealth in India using relative, intermediate and absolute measures. They find that using relative measures, increases in consumption inequality in India since the 1990s have been modest. However, both intermediate and absolute measures show sharp increases. The results for wealth inequality are similar, except that wealth inequality decreased slightly using relative measures. Jayadev et al. (2007) focus on interpersonal wealth inequality in India post-1990s, and use relative and absolute measures to find that it has increased.

68

Atkinson and Brandolini (2010) make a case for, and examine absolute inequality in the context of the world distribution of income. Bosmans et al. (2011) use absolute, relative and intermediate perspectives to analyze world inequality and argue that the choice of perspective matters. One confronts the relative versus absolute issue in debates on inclusion and the impact of growth on the poor (Klasen 2008, Grosse et al. 2008). It is obvious that the poor have to grow at a faster rate than the rich if they have to catch up with them. In general, the absolute perspective, which imposes restrictions on absolute increases, turns out to be stronger than the relative perspective, which focuses on relative growth rates, i.e. if the income growth rate of the poor exceeds that of the non-poor, relative inequality will fall while it may often be the case that, in this situation, the absolute increments of the poor are smaller than those of the non-poor and absolute inequality will rise. Examples of the application of these ideas are: Jayaraj and Subramanian (2012) and Motiram and Naraparaju (2008) from India; Grosse et al. (2008) from Bolivia, which also examines non-income dimensions.Grosse et al. (2008) and Klasen (2008) argue that considering absolute inequality is particularly appropriate for non-income dimensions such as education or health where absolute increments are a much more commonly accepted metric than proportionate changes.

69

2.10. Inequality versus polarization

70

Over the past two decades, a literature has emerged that has conceptualized the phenomenon of “polarization,” which is broadly: “the appearance (or disappearance) of groups in a distribution” (Chakravarty, 2009, p. 105). Several measures have been proposed and some scholars have argued that these are better connected to conflict than measures of inequality. An old idea that can be traced back to Aristotle (Motiram and Sarma 2013) is that societies with a strong middle tend to prosper, be stable, and are less prone to conflict. Measures of bipolarization (e.g. Foster and Wolfson 1992; Wolfson 1994; Wang and Tsui 2002) are motivated by this idea. They conceptualize the middle in terms of the median and divide the population into groups below and above the median. Many measures of inequality satisfy the Dalton-Pigou principle, which holds that regressive transfers (from a poorer to a richer person) increase inequality and progressive transfers decrease it. Measures of bipolarization replace this with two other principles: increasing spread and increasing bipolarity. The former demands that polarization increases if: a transformation makes a rich person richer or a poor person poorer without affecting the median, or a transfer occurs from a poor person to a rich person across the middle. The latter holds that polarization increases if a transfer occurs from a richer person to a poorer person on the same side of the median, resulting in their moving together. At an intuitive level, one can see the connection of this with the formation of poles on either side of the median. Also, while a progressive transfer decreases inequality, it could increase polarization.

71

As in the case of inequality, both relative and absolute versions of polarization indices are possible. Analogous to the ideas of Lorenz curve and Lorenz dominance, the Relative Bipolarization Curve and Relative Bipolarization dominance have been proposed (Chakravarty 2009, pp. 117-121).

72

Distinct from the idea of bipolarization, Esteban and Ray (1994) consider a situation where there are several identifiable, pre-existing income groups. They characterize polarization by arguing that individuals belonging to a particular group identify with one another and are alienated from those belonging to the other groups. Duclos et al. (2004) extend this analysis to a continuous income distribution where income groups are identified endogenously. While the above measures deal with one dimension (income), Zhang and Kanbur (2001) consider polarization in a context where various non-income groups coexist and individuals can be distinguished in two dimensions – income and membership of the group that they belong to. They use the single-parameter entropy family of inequality indices to decompose overall income inequality into two components: inequality between groups and inequality within groups, and argue that polarization can be measured by considering the ratio of the former to the latter. Such multidimensional polarization indices are also a way of measuring inequality among groups, or horizontal inequality (Stewart 2002).

73

Do the polarization measures provide different results and insights compared to inequality measures? The evidence on this question is mixed. Zhang and Kanbur (2001) use data from China during 1983-95 to conclude that they do not. Using their measure, they show that rural-urban polarization is quite high, whereas inland-coastal polarization is low, but rising. Motiram and Sarma (2013) use consumption data from India to find that at the all-India level, polarization and inequality measures show similar increasing trends since early 1990s. However, state-level comparisons and trends display some differences. They find that group-based polarization has increased on many fronts: state, geographic regions, and rural-urban. Duclos et al. (2004) use data from the Luxembourg Income Study for 21 European countries to find that inequality and polarization rankings differ. Ravallion and Chen (1997) consider data from developing and transition countries to find close correspondence.

74

2.11. Objective versus subjective measures (perceptions of inequality and mobility)

75

Most of the literature is focused on objective, quantifiable measures of inequality.  But there are, of course, other ways to consider inequality, including subjective measures.  Three different literatures can be subsumed under this heading.  The first considers subjective outcome indicators such as 'happiness' or 'life satisfaction' which has been used as an important alternative indicator of well-being.  While there is very large and mature literature concerned with measuring the levels and determinants of happiness and life satisfaction, inequality in happiness has been much less studied.  This is partly related to data issues as happiness and life satisfaction is usually measured on a 4, 5, or 10-step Likert scale from which it is no easy to derive valid, credible, and comparable inequality measures.  In it also not clear how to compare these inequalities with inequalities in an open-ended concept such as incomes.  There is a small literature that has considered inequality in happiness.  For example, Kalmijn and Veenhoven (2005) recommend using the standard deviation as an inequality measure. When using a 10-point scale for life satisfaction, the measured inequality in life satisfaction is quite small (as to be expected) and generally falls with high mean levels of life satisfaction (Ott, 2005).  This finding is similar to findings on inequality in non-income dimensions and might be related to the fact that these indicators are bounded (Grimm et al. 2008). Given the conceptual and measurement challenges, it is not clear what robust conclusions emerge from this rather small literature on inequality in subjective well-being.

76

A second literature uses subjective perceptions to derive objective metrics.  Examples of this are so-called subjective poverty lines which ask households what level of income they would require to not be poor (e.g. Pradhan and Ravallion, 2000), or surveys that ask households to subjectively place themselves in the income distribution (e.g. whether in which quintile they would place themselves).  Such measures can be a useful complement to objective inequality measures and yield interesting insights.  For example, as shown by Cruces et al. (2013), poor people tend to overestimate their rank in the income distribution while rich people underestimate it, which is related to their perceptions being based on comparisons with relatively homogeneous reference groups.  As a result, subjective inequality tends to be lower than objective inequality, but the difference is not uniform but depends on circumstances (see also Engelhardt and Wagener, 2014).

77

A third (closely related) literature asks about people's perception of the level of income or wealth inequality and mobility in a society (irrespective of their own position) and maybe one could use those subjective perceptions as measures of inequality.  As to be expected from the discussion above, income inequality is estimated to be much smaller than it objectively the case.  This biased perception seems to be even larger for estimates of the wealth distribution (Norton and Ariely, 2011).  Similarly, perceptions of social mobility tend to be higher than actual social mobility, and there are important international differences, with people in the US believing that mobility is substantially larger than in Europe (e.g. Alesina and Glaeser, 2004), even though it appears that mobility is lower in the US than in most European countries (Corak, 2013).  This literature has also shown, theoretically and empirically, that perceptions of inequality and of mobility affect preferences for redistribution (e.g. Cruces et al. 2013; Engelhardt and Wagner, 2014; Benabou and Ok, 2011).

78

All three literatures show that subjective assessment of inequality (and mobility) are generally a poor proxy of objective measures of inequality, and are affected by various systematic biases.  Thus as short-cut proxies for inequality or mobility, they are of limited use.  Instead, however, these subjective measures provide important complementary information that sometimes can be as important or more important for assessments of societies or preferences for political change.  

79

2.12. Static versus dynamic inequality: social mobility and stratification

80

Mobility is concerned with how individuals or groups fare over time. We might be interested in interpersonal mobility - how easy or difficult it is for individuals to move from one income or occupational group to another, or intergenerational mobility – how likely or unlikely it is for children to fall in the same group (income quantile, occupational class, educational level etc.) as their parents. This kind of dynamic analysis can be contrasted with static analysis, wherein we examine inequality (usually summarized by an index like the Gini) at a point in time, or compare inequalities at different points in time without tracking the same individuals or households. Social mobility and static inequality require different kinds of data. Panel data or data tracking different generations is needed for understanding mobility. Given the relative paucity of such data in developing countries, the empirical literature on mobility is sparse – two recent exceptions are Hnatovska et al. (2011) and Motiram and Singh (2012). In contrast, the literature from developed countries is richer and there are studies from several countries, e.g. Bjorklund and Jantti (2000), Bowles et al. (2005), Corak (2013), Blanden et al. (2013).

81

Understanding mobility also requires different approaches to measurement, and the axiomatic method is not as well developed as in the case of inequality (Fields and Ok 1999). We illustrate the basic ideas using the example of intergenerational income mobility. Scholars have relied upon transition matrices that give the relative frequency or probability of the next generation (usually the son) falling in a particular income quantile for each income quantile that the parent (usually the father) belongs to. Summary mobility measures have been constructed from transition matrices and these have used to rank in terms of mobility. One such measure is based upon persistence – the average probability that a child will fall in the same income quantile as a parent – the higher it is, the lower is the mobility. For a description of such measures, see e.g. Shorrocks (1978), Sommers and Conlisk (1979), Fields and Ok (1999), Van De Gaer (2000), and Formby et al. (2004).

82

Another approach is to perform an OLS regression of the permanent (long-term) income of the child on the permanent income of the parent and examine the coefficient (intergenerational earnings elasticity) – the higher the elasticity, the lesser is the mobility or the more responsive the child’s income is to the parent’s income (Corak 2013). There is enormous variation in intergenerational mobilities across countries. This is due to differences in families, labor market institutions and policies (Corak 2013). Even within a country, there are differences in mobilities across socioeconomic groups. Here, it is important to distinguish between upward and downward mobilities. In the intergenerational context, one way to assess the latter is to examine how children and parents fare compared to their respective peers, and see if children fare worse (e.g. if the father falls in the top quartile, whereas the son falls in the bottom quartile). See Acs (2010) for a discussion, (including the various nuances) and the finding that certain disadvantaged groups in the United States (e.g. black men) display more downward mobility. Motiram and Singh (2012) find that in India groups that have been historically discriminated against (Scheduled Castes and Tribes) are characterized by higher downward occupational mobility.

83

Mobility can be viewed from the perspective of equality of opportunity. An influential view due to Roemer (1998) links the degree of inequality of opportunity to the influence of factors beyond an individual's control (circumstances). Since parental background is a (usually particularly influential) circumstance variable, from this perspective, one could argue that societies with low intergenerational mobility are characterized by high inequality of opportunity. Roemer (2004) highlights various mechanisms (e.g. access to networks, health, family culture) through which children can benefit from parents.

84

The discussion so far was mostly concerned with long-term or intergenerational mobility.But in the short-term, there can also be a great deal of mobility, particularly related to positive or negative shocks.This issue has been discussed in the context of chronic versus transitory poverty (Ravallion and Jalan, 1998).Panel studies that study income dynamics tend to find that inequality in average multiyear incomes tends to be substantially smaller than inequality in annual incomes, suggesting that shocks and/or measurement error are responsible for the higher reported static inequality at any point in time (e.g. Grimm, 2007; Woolard and Klasen, 2007; Burger, Klasen, and Zoch 2015)

85

2.13. Discussion of summary measures of inequality polarization mobility and their policy relevance

86

This section will be added as an appendix.

87

2.14. How inequality is dealt with in the SDGs (and how inequality should inform all SDGs)

88

In contrast to the Millennium Development Goals, the recently concluded Sustainable Development Goals includes a goal relating to inequality.Goal states to 'reduce inequality between and within nations'(UN, 2015).Thus inequality has received greater recognition as a potential barrier to sustainable development.When it comes to targets and indicators, SDG10 faces the challenge that it seeks to address two rather separate issues: inequality between nations which, at least if measured in income terms, is largely related to differences in economic growth rates between poor and rich countries, and inequality within nations, which depends on the distribution of asset and returns to those assets as well as redistribution efforts of the state.The first four targets deal with inequality within nations and call for greater income growth of the poorest 40% (the World Bank 'shared prosperity' goal), promotion of social inclusion, reduction of inequalities of opportunities of outcomes, and of discrimination, and fiscal, wage and social protection policies for greater equality.While these seem useful targets, they touch only parts of a redistribution agenda which might also include asset redistribution or measures to improve the returns to assets for poorer population groups.The other targets deal largely with inequality between nations and include a rather selective list of items, including improving migration opportunities and lowering remittance costs, improving trade opportunities and aid for LDCs, better regulation of global financial markets, and greater representation of developing countries in international organizations.The most important measure to reduce (income) inequality between nations, higher growth for poorer countries, is not tackled here at all, but it mentioned as a target for goal 8 (promote sustained, inclusive, and sustainable economic growth...'), where a target calls for 7% annual growth in LDCs.The recently concluded indicators for Goal 10 (UN, 2016) follow this rather selective basket of targets but suffer from the additional problem that for some targets there are no readily available indicators.Clearly, the new inequality goal at present is extremely broad, covering within and between-country inequality, and rather selective in targets and indicators.More work will likely be required to develop a sound policy agenda from this.

89

2.15. Measurement challenges[6]

90

              When measuring inequality, many measurement challenges arise.  They (mostly) relate to the following issues:

91
  • Most inequality analyses inherently requires micro data at the level of households or individuals; such surveys must be done regularly, be consistent over time and (ideally) across countries, must be able fully capture the concept that is being measured (e.g. income, wealth, health), and must be based on reliable fieldwork and data gathering; in many countries of the world, particularly in developing countries, this remains very challenging;
  • Inequality statistics are inherently sensitive to the tails of the distribution (the different measures discussed above differ in their degree of sensitivity); thus measurement error in the tails of the distribution can be particularly damaging; but it is also the case that such measurement error is particularly common as both the very poor and worst off and the very rich and best off are either less likely to participate in surveys or their information is captured with less accuracy.  Both levels and trends in inequality can be affected by this.
  • Dynamic inequality assessments or assessments of mobility require panel data which are much rarer, particularly again in developing countries, and suffer from additional conceptual and empirical problems, including attrition and changes to household boundaries. 
92

    

93

              While these problems remain severe, we focus here on recent efforts to address them at least partly, particularly in the dimension of income inequality.  As a result of multiple efforts by academics, National Statistics Offices, and international organizations to improve and harmonize income inequality data and analysis, there has been an increase in the number of publicly available databases. All contain summary statistics that describe national-level estimates of inequality in incomes or consumption expenditures in multiple countries over multiple years.

94

              Depending on the source of the summary inequality statistics they report, there are four main types of databases.

95
  • First, there are those which calculate inequality measures directly from household surveys (in alphabetical order): CEPALSTAT, (UN Economic Commission for Latin America and the Caribbean); CEQ Standard Indicators (Commitment to Equity Institute, Tulane University); EUROMOD (University of Essex); IDD (Income Distribution Database/OECD); LIS (Luxembourg Income Study); SEDLAC (Socio-Economic Database for Latin America and the Caribbean/CEDLAS at Universidad Nacional de La Plata and World Bank); and, PovcalNet (World Development Indicators, World Bank).
96
  • Second, there are datasets that combine indicators from a variety of other sources which calculated inequality measures directly from household surveys (in alphabetical order): ): ATG (All the Ginis); the GINI Project ; and, WIID (World Income Inequality Database/UNU-WIDER). 
97
  • The third type of data sets generate inequality measures through a variety of imputation and statistical inference methods instead of relying directly on household surveys or unit-record datasets (in alphabetical order): GCIP (Global Consumption and Income Project); SWIID (Standardized World Income Inequality Database); and, UTIP (University of Texas Income Inequality Project).
98
  • One of the most important limitations in traditional measures of inequality is that they generally rely on household surveys (either directly or indirectly) which tend to significantly underreport the wealthiest so the vast majority of inequality measures underrepresent true levels of inequality.  The fourth type of datasets attempts to address this problem by relying on unit-record datasets such as tax records.  A salient example of this approach is the WID (World Wealth and Income Database).[7]
99

The above cross-national inequality databases are being used by researchers, with increasing frequency, to document global or regional trends. Yet, these different databases are often designed for different purposes, and are constructed in very different ways. In addition to reporting different inequality measures (the Gini coefficient being the most common metric but not the only one), the datasets differ on the nature of the individual welfare indicator for which inequality is calculated. To give a few examples, to the usual income versus consumption option, the databases can differ on whether the indicator of individual welfare is measured on a per capita basis or per equivalent adult (and, if the latter, on which equivalence scale is used); whether the welfare indicator is an estimate of monetary income/consumption only or it includes imputed rent for owner’s occupied housing and consumption of own production; and, whether adjustments (and which ones) are made to the microdata to correct for underreporting, to eliminate outliers, or to address missing responses. For the datasets which rely on secondary sources or use imputation methods, results are highly sensitive to the utilized methods and one often does not have the full information on the characteristics of the underlying data, even if the methods are described with care (which is also not always the case).

100

Depending on the database utilized, the analyses can therefore yield conflicting pictures of inequality, both in levels and in trends (Bourguignon, 2015; Ferreira, Lustig, and Teles, 2015; Gasparini and Tornarolli, 2015; Jenkins, 2015; Ravallion, 2015; Smeeding and Latner, 2015; Wittenberg, 2015).  Let’s illustrate with an example for Kenya:

101

Based on the World Development Indicators and All the Ginis, inequality rose steadily between 1994 and 2005.  In both cases, however, there is no data for the years in between 1997 and 2005.  SWIID estimates suggest that inequality fell over that period while WIID includes data from the Society for International Development that estimates inequality increased sharply by 1999. For a researcher studying inequality trends in Kenya in the 1990s and 2000s, this would be of great concern. If she began her analysis in 1992, every dataset would show a decline in inequality. If, however, she chose to begin the analysis in 1994, WDI would show increasing inequality, ATG would show an increase followed by a sharp decline, and SWIID would show a steady decline. In sum, inequality trends in Kenya – and a number of other countries – appear to be very sensitive to the dataset selected. (Ferreira, Lustig and Teles, 2015)

102

Furthermore, important questions such as whether there is or is not inequality convergence-- the finding that inequality has fallen in what had been highly unequal countries and risen in countries that had been more egalitarian (Benabou (1996), Bleaney and Nishiyama (2003), and Ravallion (2003))-- are affected by the choice of dataset.  As discussed by Lustig and Teles (2016), “…different datasets frequently produce different results [in terms of inequality convergence], even when the countries, the welfare concept, the inequality metric, and the time period are held constant.”

103

A remaining issue refers to the challenges of reconciling micro (i.e., household-survey-based) data with macro data (i.e., national accounts).  For most countries in the world, totals for household income and consumption from surveys do not match the equivalent totals from national accounts.  The difference in some countries can be huge (e.g., survey based total consumption can equal only fifty percent of consumption from national accounts).  Efforts undertaken by the OECD/Eurostat Expert Group in Integrating Disparities in National Accounts in Europe and the US Census and others in the United States, are designed to address this challenge. However, in most developing countries, analogous efforts do not exist. Furthermore, as argued by Deaton (2005), in developing countries admin data on disposable income (or private consumption) is often not very reliable so adjusting survey totals to be equal to national account totals may introduce more measurement errors of inequality than correct them. 

104

Given that inequality analyses are so sensitive to the choice of database, researchers should acquire a sound understanding of the assumptions and methodological choices embodied in the data they are about to use.  Researchers should also undertake systematic robustness checks: that is, are results sensitive to the dataset utilized in the analysis? In addition, “…database producers bear an important responsibility in documenting all their assumptions clearly and thoroughly, and making as much of their data, programs and results publicly available to allow for replicability whenever it applies. … Furthermore, dataset producers would be wise to compare their methods and results with one another and, eventually, perhaps even agree on conventions of best-practice in the calculation of inequality indicators from microdata-based, secondary, and imputation-based sources.” (Ferreira, Lustig, and Teles, op. cit.)

105

3. Why inequality matters

106

3.1. Introduction

107

A lot of the empirical literature on inequality, especially within economics, has focused on a purely economic perspective on inequality, that tends to examine the relationship between economic inequality (usually proxied by income inequality) and other economic variables such as economic growth, savings, labor market performance and the like.These relationships are important and will be discussed in detail below.When discussing the relevance of inequality, it is, however, critical to move beyond an economic perspective. Inequality is seen, both in the larger social science literature as well as in public discourses, as much as a moral issue relating to justice and human rights. It is seen as a political issue affecting the functioning of our political systems, and it is seen as a social issue affecting deep-seated social stratification.Moreover, inequalities in different spheres interact.For example, economic social inequality is likely to affect political inequality, which can in turn have repercussions for economic and social inequality.As a result, it is important to document levels and trends in these wider inequalities, examine their interactions, and consider how they matter.

108

3.1.1 Tolerable versus problematic inequalities

109

As a starting point, one might interpret the Kuznets ‘inverted-U’ hypothesis between income inequality and economic growth – inequality initially rises with economic growth and then falls eventually – as illustrating this phenomenon. In effect, inequality rises naturally as different individuals (even with the same initial allocation) behave differently toward investment projects due to differences in their rates of time preference. This initial level of inequality can be seen as ‘tolerable’. However, at a certain point inequality becomes ‘intolerable’ if it reduces economic performance or becomes intrinsically problematic.

110

In a more general sense, inequality may also be considered tolerable if well-informed individuals facing the same choice set make choices that lead to different and unequal outcomes. However, individual preferences themselves are often shaped by cultural practices, for instance, which lead the individual to make choices considered sub-optimal in a larger societal preference realm. In this regard, it is not even clear that the resulting inequality is actually tolerable, unless individuals can be considered as facing the same set of initial conditions ex-ante, including cultural practices.

111

UNDP (2013) provides some useful insights into the distinction between tolerable inequality and problematic inequality, based on interviews conducted to collate views from national policymakers on inequality. According to this report, inequality is acceptable if: (1) it is due to individual efforts and it originates from fair competition; (2) everyone is guaranteed a minimum standard of living; and poverty is declining. However, inequality is unacceptable if it undermines equality of opportunities (UNDP, 2013; p.207). While empirically appealing, the above hypothesis and views are difficult to assess empirically. Thus, one must resort to providing case studies for illustrative purposes.

112

In sum, inequalities might be tolerable if they are intrinsically defensible and/or lead to instrumental trade-offs with other desirable goals such as economic growth and aggregate improvements in well-being, or increases in mobility.  These are issues that will be discussed below. 

113

3.2. Intrinsic concerns

114

3.2.1 Inequality of opportunities, capabilities, primary goods

115

To what extent is a concern with inequality mandated by a concern for social justice?

116

Social justice has been conceived in diverse ways, as encompassing concerns for procedural rights and liberties, the absence of absolute deprivations and the maintenance of inequalities within tolerable limits if not the eradication altogether of specific kinds of inequalities deemed to be illegitimate.  The distinction between inequality in initial endowments and inequality in final outcomes has also figured prominently in recent debates (see e.g. the arguments of and with “luck egalitarians” who emphasize initial or “starting gate” equality in relevant endowments, such as G.A. Cohen, Ronald Dworkin, John Rawls, John Roemer and many others who have been influential since the 1970s).  A plausible case (see Sen, 1979) has also been made we should think about all theories of social justice as being egalitarian in the sense that they answer the question “Equality of what?” In some specific way, for example libertarian theories (such as that of Nozick (1974)) demand equality in the space of specific rights.  From such a perspective, a concern with inequality is always mandated by concern for social justice, even if the specific inequalities of concern depend on the particular theory of social justice that is adopted.   In addition to the kinds of inequality that are of concern, the extent to which inequalities are of concern may depend on specifics of the theory. For example, the difference principle advanced by John Rawls famously tolerates inequalities as long as they can be demonstrated to further a specific good, namely the level of advantage of the least advantaged members of society. Although perspectives on social justice that demand concern with equality of outcomes have been less influential in debates since the 1970s, they have been gaining in force. Anderson (1999) for instance, argues that social justice demands a felt solidaristic concern for others and not merely a reasoned commitment to correcting injustices, both for evaluative and empirical reasons.  This may in turn have as a consequence a concern with equality of outcome as well as of starting point. It would seem implausible to be indifferent to realized outcomes (as a matter of social justice assessment) while having felt concern for others as an integral demand of social justice. 

117

If we take “inequality” as referring to the concern with relative inequality only as distinguished from a concern with absolute levels of advantage or disadvantaged realized, then the case for disvaluing inequality as such within a theory of social justice becomes weaker. (Of course, the two kinds of fact are empirically deeply interlinked as shown below). Although inequalities among the rich and the super-rich may be realized due to the ‘brute luck’ of having been born in one position or another or of having been offered specific opportunities or not, the demand for corrective action from the standpoint of social justice may be thin.  The distinction between disvaluing inequality in itself and disvaluing it for some other reason (including its impact on absolute disadvantages) is of importance here (see e.g. Parfit (1997). However, in our world, relative inequalities coexist with low or even unacceptably low levels of absolute advantage, so it is of doubtful value to attempt to separate the discussion of the two concerns from each other.

118

3.2.2 Inequality of opportunities, capabilities, primary goods

119

As already noted, the question of whether to focus on ex ante inequality of endowments or ex post inequality of outcomes has been an important one in contemporary discussions. The appropriate form of corrective action to undertake (e.g. to attempt to equalize resource endowments or in educational opportunities vs. taxation and transfer of final incomes) would in part depend on what is thought to be the appropriate focus in this regard.  Another, distinct, concern has been that of what “space” to focus upon.  For instance, even if one is concerned with inequality of endowments, one can confine oneself to those resources possessed by persons and external to them or take a more capacious view that includes natural endowments that are constitutive of a person (such as athletic talents or intelligence).   Even if the latter cannot directly be redistributed the recognition of their inclusion in a perspective of what are the morally relevant considerations from the perspective of corrective action may influence those actions one undertakes in other dimensions in which such action is feasible. Similarly, one can be concerned with all outcomes realized by persons (e.g. health and wealth) or just some (e.g. wealth).  The choice to focus on specific outcomes may be based on considerations such as the special relevance of these considerations for human well-being or the role of social factors vs. individual decisions in determining individual outcomes. 

120

Rawls (1971) advanced the idea of “primary goods” as an index of the diverse resources, both social and individual in nature, relevant for ex ante assessment of the opportunities available to a person to realize a chosen life plan.  The concept of primary goods is inherently plural and extends beyond economic resources.  However, it is also squarely focused on means to achieve outcomes rather than outcomes realized. Sen (1992, and other writings) has emphasized the contrast between this Rawlsian conception and an alternate perspective, that of capabilities, which takes account of the freedoms to achieve specific outcomes.  An aspect of subtlety of the capability perspective rests in its being “freedom based” and thus concerned with opportunities rather than outcomes while at the same time taking note of the distinction between outcomes of different kinds by specially valuing the freedom to achieve those outcomes that one has “reason to value”.  In this way, it mutes the distinction between the ex ante and ex post approaches.  The capability perspective also takes note of interpersonal variations in the ability to translate resources into outcomes. Whether such variation is appropriate to take note of or not from a moral perspective can be debated (see e.g. Pogge (2002) who argues for a resourcist conception that takes note of a normal range of human variation in determining resource allowances but does not make interpersonal allowances beyond that).  Insofar as there are systematic variations in the mapping between resources and outcomes that are attributable to “ascriptive” aspects of identity (e.g. sex) or other unchosen factors (e.g. disabilities at birth) these may make problematic the conception of a normal range, and have special reason to be taken note of when inequality concerns are being addressed – as contrasted, for instance, with those variations due to acquired tastes. This is the perspective taken by those who argue for equality of opportunities (Roemer, 1998) who distinguishes between circumstances that are beyond the control of individuals and efforts which are. 

121

Clearly, all three approaches see particular forms of inequality as morally problematic and demanding of corrective action.  They differ in their approach and as a result their evaluation will differ by approach and the case being considered.  An element of evaluative judgment discriminating between such cases is inescapable in such exercises.

122

3.2.3 Relational aspects of inequality (power, status)

123

Perceptions of inequality and its impact on well-being

124

How do people feel about economic inequality? A large body of research in psychology, economics, and political science has shown that people are ‘inequality averse’ – they react negatively to unequal distribution or allocation of resources, regardless of whether it leaves them in an advantaged or a disadvantaged position (although, unsurprisingly, they feel more comfortable with the former than the latter; Fehr & Schmidt, 1999; Loewenstein, Thompson, & Bazerman, 1989). In fact, people’s aversion to inequality seems to be universal (Henrich et al., 2006). People are willing to incur personal costs to reduce a state of inequality (e.g., Camerer & Thaler, 1995; Johnson, Dawes, Fowler, McElreath, & Smirnov, 2009), and do so even as neutral third-party observers of inequality (Fehr & Fischbacher, 2004). In recent studies using functional magnetic resonance imaging (fMRI), researchers have documented neural activation of reward processing areas—associated with positive affect—in response to fair allocation of resources (Tabibnia, Satpute, & Lieberman, 2008; Tricomi, Rangel, Camerer, & O’Doherty, 2010), and activation of an area associated with specific negative emotions like anger and disgust in response to unfair allocations of resources (Sanfey, Rilling, Aronson, Nystrom, & Cohen, 2003). People’s aversion to inequality may be deep-rooted, emerging at a very young age. Children as young as 7 years old exhibit aversion to unequal allocation of resources (Fehr, Bernhard, & Rockenbach, 2008; Kogut, 2012; Sutter, 2007), and even 5 years olds apparently show a preference for egalitarian distribution of rewards (Gummerum, Hanoch, Keller, Parsons, & Hummel, 2010).

125

Given that people are overwhelmingly averse to even small and inconsequential economic inequalities in the lab, how can one explain people’s seemingly acquiescent acceptance of larger and more consequential inequalities in the real world? One possible solution is that people are simply not aware of the scope of economic inequality. Indeed, when Norton and Ariely (2011) asked people to estimate the wealth distribution in the United States, they found that Americans overwhelmingly underestimated the scope of inequality (believing, for example, that the bottom 40% of the population owns 10% – as opposed to the actual 0.3% – of total wealth.)Norton and colleagues (2014) found similar misperceptions of inequality among Australian respondents, who significantly underestimated the share of the wealth among the country’s richest and overestimated the share of the wealth among its poorest. Finally, Kiatpongsan & Norton (2014) found that across 16 countries, people dramatically underestimated the pay discrepancies between CEOs and unskilled workers, often by several degrees of magnitude. Thus, people may accept economic inequality partly because they simply don’t appreciate how bad it is.

126

Although misperception may shed light on why people accept vast economic inequalities, it cannot fully explain it. For example, although Americans and Australians underestimate the scope of wealth inequality in their countries, they still prefer their countries to be more equal than they perceive them to be (Norton & Ariely, 2011; Norton et al., 2014). And, although people across the world underestimate the pay discrepancy between CEOs and unskilled workers, they believe that even this underestimation is much higher than it should be (Kiatpongsan & Norton, 2014). Thus, while people significantly underestimate economic inequality, they still want to see the world becoming more equal than their overly rose-tinted views suggest.

127

Acceptance of inequality

128

Why, then, are people not ‘up in arms’ over economic inequality? One contributing factor may be the belief in upward social mobility (see also below). According to the “dominant ideology” in the United States (and, increasingly so, in many Western countries), people accept inequality as long as they believe that hard work pays off (Kluegel & Smith, 1986).  For example, in a cross-national survey of 25 countries, Shariff, Wiwad, & Aknin (2016) found a positive correlation between social mobility and acceptance of inequality: the more mobile a society, the more willing people are to tolerate large income discrepancies between rich and poor. However, as shown by Davidai and Gilovich (2015), people’s perceptions of economic mobility may be as distorted as their perception of economic inequality. People overestimate upward social mobility, believing that a person born into the poorest echelon of society is more likely to rise up the social ladder than they actually are (Davidai & Gilovich, 2015; see also Kraus & Tan, 2015; Kraus, 2015). A reason people may tolerate inequality is because they overestimate how easy it is to escape the bottom rungs of the income ladder.

129

The mis-calibrated belief in upward social mobility may be seen as part of people’s general tendency to see the institutional, judicial, and economic systems they depend on as fair and legitimate (Jost & Banaji, 1994). Because people have a need to perceive the prevailing economic system as one where people get what they deserve and deserve what they get, they glorify wealthy individuals as highly competent and industrious (Mandisodza, Jost, & Unzueta, 2006), and they vilify the poor as both incompetent and lazy (Fong, 2001). And, as they learn about “self-made” wealthy individuals, people become less supportive of social programs for helping the disadvantaged (Ho, Sanbonmastu, & Akimoto, 2002; Wakslak, Jost, Tyler, & Chen, 2007). Indeed, merely thinking about people’s ability to choose for themselves increases belief in personal responsibility, decreases concern with wealth inequality and, consequently, diminishes support for redistributive measures (Savani, Stephens, & Markus, 2011; Savani & Rattan, 2012).

130

Underlying the beliefs in social mobility and the attribution of positive traits to the wealthy (and negative traits to the poor) is the distinction between inequality and inequity. While generally averse to inequality, people are typically more concerned with procedural than with distributional justice (Tyler, 2011). People tend to be more egalitarian when they see others expend as much or more effort than they (Oxoby & Spraggon, 2008), but become less egalitarian when they have expended more effort than others (Cherry, Frykblom, Shogren, 2002; Hoffman, McCabe, Shachat, & Smith, 1994). Similarly, people are less concerned about inequality when it involves seemingly incompetent (rather than competent) others (Ruffle, 1998). To the extent that people perceive large wealth and income discrepancies as indicative of underlying differences in effort and competence, they are likely to accept these as relatively benign manifestations of inequity rather than unacceptable instances of inequality.

131

People are less concerned about economic inequality also because it tends to appear less personally salient. As a general rule, people pay attention to and care more about their local, relative standing than their absolute position (Frank, 1985; Norton, 2013). For example, a person’s income rank is a stronger predictor of life satisfaction (Boyce, Brown, & Moore, 2010) and physical health (Daly, Boyce, & Wood, 2015) than their absolute income.

132

Relatedly, people are more concerned about local, rather than global, inequality, and they tend to apply more rigorous standards of fairness and egalitarianism within small communities than large ones (Deutsch, 1975; Pfeffer & Langton, 1988). They are more likely to compare themselves to similar others, and to be disturbed by inequality compared with similar as opposed to dissimilar others (Baron & Pfeffer, 1994). In light of the stratification that occurs in contexts of inequality, people do not generally experience vast economic inequalities within their local context. While interactions between rich and poor happen on a daily basis, these often occur in a context that favors equity over equality (such as when a waiter serves on a wealthy patron). Only in those relatively rare cases where inequality is locally salient, and occurs in contexts that do not trigger equity considerations (e.g., when wealth inequality is clearly seen to lead to unequal health outcomes) might people become more resistant to ongoing economic inequality.

133

3.2.4 Impact of inequality on power relations

134

Inequality can be present in various independent dimensions.  It can be economic or it can take the form of social distinctions, differences in political power or influence, or cultural prestige.  The various forms of advantage, and associated inequalities, are distinct, and may each also be tied to deep-seated social distinctions based on gender, race, caste or other factors.   The day-to-day psychological and social experience of individual well-being or ill-being may be crucially shaped by inequalities in non-economic dimensions.  However, inequalities in these various dimensions are often closely causally and symbolically linked to economic inequalities too.  Although there are distinctions between them, the deep interlinkages that exist make it hard to analyze the historical evolution or the contemporary workings of any one dimension of advantage and disadvantage separately.

135

One of the most pernicious effects of economic inequality is that it can have an effect on power relations in a society.  These effects on power relations can take various forms, more obvious and more subtle.  For instance, the greater influence of the more privileged, both directly through the economic resources they command and indirectly through their social status, can cause the institutional arrangements and policy choices which emerge in a society to be those which favour the interests of the relatively more privileged.  In this way, institutions can become instruments for extending and cementing inequalities, cumulating in a vicious cycle.  A great deal of literature in political economy and in political sociology has focused on these connections.[8]  For example, economic resources can be used to buy votes of ordinary voters or of legislators, buy the time and influence of the media or of  ’opinion-makers’, coerce or threaten to coerce people[9] either through economic sanctions or through commanding outright ’muscle power’ and so forth.    Such mechanisms may have predictable effects, leading for instance to oligarchs protecting their wealth from taxation, or to generally lower levels of taxation than would otherwise be expected in a democratic context (as suggested, for example, by the median-voter theorem)[10].  On the other hand, there may be limits to such processes, especially where there are entrenched democratic institutions or active and politically mobilized groups (such as the middle classes) pursuing their own collective projects.  Inequality may give rise to an incline in politics, but nothing is fated.

136

Another more subtle possibility is that economic privilege becomes also overlaid over time with markers of social and cultural privilege, or becomes associated with particular traits such as supposed economic knowledge or worldliness, leading others to defer to those who are richer.   Such deference, more characteristic of traditional than contemporary societies, but also present to a degree in the latter, may be reflected in ideas as to who knows better how things really work, or in ideas of who can best command others’ respect.  It is not a surprise that very large numbers of legislators and leading politicians in many countries are people with personal wealth (although this may also reflect the independence that wealth provides). On the other hand, there is also in a democratic age skepticism of the role of the wealthy in political life, and an appropriate spotlight on inappropriate influence of the relatively prosperous. 

137

3.2.5 Inequality and its impact on well-being

138

This section will focus on direct impacts of inequality on well-being.  Indirect influences (e.g. via growth, social stability, etc.) will be covered under the instrumental role of inequality below.   Also, the focus will be largely on the impact of income inequality on well-being; other forms of inequality are being addressed in other parts of the chapter. 

139

The main theoretical arguments about why inequality matters for well-being are the following:

140

-in the utilitarian tradition, there is a strong presumption of declining marginal utility of income (essentially derived from declining marginal utility for each good), reflected in concave money-metric social welfare functions; sometimes (esp. in relation to Atkinson's formulation of inequality measurement), this is often also referred to as inequality aversion for which, as discussed above, there is substantial empirical support; 

141

-related arguments have been made by Sen (1973) and Dagum (1990) where well-being depends not only on one's own income, but on one's rank in the income distribution.  The key distinction to the declining marginal utility view is that the relational aspect of inequality is emphasized in the sense that the difference in well-being to a reference group is particularly important, rather than the absolute level of well-being.  For such a view of the impact of inequality there is also substantial empirical support from the happiness and experimental literature (e.g. Klasen, 2008). 

142

These two formulations have been incorporated in the Atkinson (the 'equally distributed equivalent income') and the Sen (welfare equals mean income multiplied by 1 mines the Gini coefficient) welfare measures, respectively.  Empirical applications include Jenkins (1997), Gruen and Klasen (2001, 2003, 2008, 2012, 2013). 

143

A first strand of the empirical literature has found, as already discussed above, substantial support from the experimental literature that inequality aversion is present and sizable and that fairness considerations are important drivers of assessments of well-being (as well as behavior) (Fehr & Schmidt, 1999; Loewenstein, Thompson, & Bazerman, 1989). In fact, people’s aversion to inequality seems to be universal (Henrich et al., 2006). People are willing to incur personal costs to reduce a state of inequality (e.g., Camerer & Thaler, 1995; Johnson, Dawes, Fowler, McElreath, & Smirnov, 2009), and do so even as neutral third-party observers of inequality (Fehr & Fischbacher, 2004).

144

A second strand of empirical literature has examined the impact of incomes and income inequality on subjective well-being.  At the micro level, this literature has found that income is associated with higher happiness but that this relationship is concave, suggesting that higher inequality reduces aggregate well-being (e.g. Deaton and Kahnemann, 2010).  The literature is somewhat less clear on whether income inequality in a society has an additional direct impact on well-being, apart from one's own position within the income distribution.  While a majority of studies find such a link ( Ferrer-i-Carbonell and Ramos, 2012; Gruen and Klasen, 2013; Blanchflower and Oswald, 2003), other do not (Sanfey and Teksöz, 2007).  Particularly in developed and transition countries, most studies find negative effects, while the effect for emerging and developing countries is less clear.

145

But clearly, the results of these two strands of literature combined confirm that inequality reduces well-being. Income inequality can also have an impact on non-income dimensions of well-being.  For example, it can affect average health and education outcomes.  For example, Pickett and Wilkinson (2015) find that income inequality causally leads to lower average health outcomes in a country.  (More studies)

146

3.2.6 Inequality and impact on intergenerational transmission of inequality

147

As argued above, inequality may be seen as tolerable if people believe that there is great mobility and thus the ability to rise up.  If there was a trade-off between the two, one would expect that inequality and mobility are inversely correlated.  The empirical evidence, however, points to the opposite conclusion: high inequality is associated with lower intergenerational mobility.  This relationship has been popularized in the so-called 'Great Gatsby Curve' (Corak, 2013, Krueger, 2012), a version of which is shown below in Figure 3.2.  It shows the relationship between intergenerational earnings mobility of fathers in the 1960s and sons in the late 1990s and inequality in disposable incomes for a range of OECD countries.  Clearly, more inequality is associated with less intergenerational mobility across this set of countries.  While there is less data on other countries, these findings have been replicated for a larger set of countries as well (Corak, 2013).  Among the mechanisms that can account for this relationship is the ability by the wealthy to invest more heavily in their off-spring, and conversely the difficulties the poor have to invest in their children, an association between a greater redistributive role of the state and more state investments to promote mobility, the direct role of inheritance, capital market imperfections that make it hard for the poor to invest, the ability of the wealthy to help secure better employment for their children, more generally different access to networks between rich and poor individuals, among many other possible mechanisms.  The relative importance of these various transmission channels has only partially been explored so far.

148

In sum, empirically it thus appears difficult to justify high inequality with the prospect of higher mobility.   

149

Figure 3.2: The Great Gatsby Curve

150

151

Source: Corak (2013)  

152

3.3. Instrumental roles of inequality 

153

3.3.1 Inequality and its impact on poverty/deprivation

154

Inequality has a direct impact on absolute poverty.  This can most easily be seen when examining the impact of income inequality on income poverty.  At a given mean income and a fixed absolute poverty line, higher inequality will invariably be associated with higher poverty (Bourguignon, 2003).  Less immediately obvious is the link between inequality and the poverty impact of growth, i.e. the impact of inequality on the poverty elasticity of growth.  Determining this relationship is useful for policy-makers as it determines the extent to which the poor benefit from overall economic growth. Empirical studies have generally found that the growth elasticity of poverty reduction has been quite low in Sub-Saharan Africa, and substantially larger in Asia (e.g. Adams, 2004; Fosu, 2009, 2011, 2015; Bresson, 2009; Kalwij and Verschoor, 2007). 

155

Bourguignon (2003) argues that instead of studying the growth elasticity of poverty reduction empirically, it can be shown that, under the assumption of a lognormal income distribution, there is a mathematical relationship linking growth, inequality levels and changes, and absolute poverty reduction.    In fact, he shows that this elasticity decreases with increasing initial inequality but increases with the income level relative to the poverty line.

156

While these results are all derived mathematically using the log-normality assumption, Bourguignon also tests these results empirically and find that they fit the data very well.  An important implication of his work is that one can essentially predict the impact growth will have on poverty depending on country characteristics.  For example, the impact of growth on absolute poverty reduction will be smaller in many Sub-Saharan African countries as they have high initial levels of inequality and the ratio of mean income to the poverty line is low.  So it is purely mathematics (rather than some kind of policy failure as often implied), when the low impact of growth on poverty reduction in Africa is lamented.  Conversely, fast poverty reduction in Asia in recent decades is linked to their lower inequality as well as a high ratio of mean income to the poverty line.   

157

A weakness of the growth elasticity of poverty reduction is that it considers percentage changes, rather than percentage point changes (e.g. a reduction of poverty from 50 to 40% is a 20% reduction, but a 10 percentage point reduction).  Policy-makers are generally interested in the percentage point reduction.  Klasen and Misselhorn (2008) therefore extend Bourguignon’s work by using absolute poverty measures, which is then called growth semi-elasticity of poverty reduction, i.e. “By how much percentage points does the poverty headcount change in response to a 1% increase in growth?” They find that again, higher growth, falling inequality, and lower initial inequality leads to higher percentage point poverty reduction.

158

3.2.2 Inequality and economic performance (outcomes)

159

3.3.2.1 Positive impacts of inequality on economic performance

160

One of the most controversially discussed effects of inequality is its impact on economic performance.  Some authors have argued that larger inequality can promote subsequent economic growth.  In particular, two arguments are made.

161

The first argument is that income inequality has a favorable effect on economic growth because the rich, in neoclassical as well as Keynesian models, are assumed to have a higher propensity to save than the poor. Hence, an increase in income inequality triggers higher aggregate savings, which in turn generate higher levels of investment and economic growth (e.g. Kaldor, 1955).[11]  Closely related is the view of the rich investing more for the benefit of all. According to this theory, larger investments of top income earners in businesses and equity markets will create more jobs for people at the lower end of the income distribution and stimulate growth. Trickle-down economics assumes that a larger share of income allocated to the top, relative to the rest of the income distribution, causes economic growth. Aghion and Bolton (1997) have developed a model of economic growth and inequality, which includes a trickle-down effect of capital accumulation in the presence of imperfect capital markets. Nevertheless, even in their theoretical model, redistribution brings greater equality of opportunity and accelerates the trickle-down process.

162

The second argument revolves around incentives.  Two aspects are worth noting here: one is that inequality itself provides incentives to work hard and invest as the returns to such investments tend to be higher than in a more equal setting.  These greater efforts and investments increase output generally.  A second aspect relates to redistribution by the state.  Redistribution can reduce inequality but can also negatively affect incentives, e.g. through high marginal taxation on labor incomes.

163

Li and Zou (1998) provide empirical support for the above hypothesis in a politico-economic setting, where a more equal income distribution triggers higher income taxation which in turn produces lower economic growth. This outcome is underpinned by the assumption that such taxation finances mainly public consumption rather than production services.[12] Forbes (2000) also empirically finds that increases in inequality are associated with greater economic growth in the short and medium term.

164

Drawing a distinction between inequality of outcomes and inequality of opportunities, it is expected that the former offers the needed inducement to invest in education and physical capital (capital accumulation) and to work hard, thus leading to an increase in growth (WDR, 2006). Furthermore, inequality of outcomes encourages innovation (Mirrlees, 1971) and agglomeration of economies (Fallah and Partridge, 2007; Castells-Quintana and Royuela, 2011), thus boosting economic growth. Relatedly, inequality that arises from market forces (market inequality), rather than that originating from socio-institutional factors such as class, ethnicity, gender, or location (structural inequality), could have beneficial impacts on growth (Easterly, 2007).

165

Debates on those mechanisms focus less on the potential plausibility of these mechanisms as their empirical relevance.  For example, few economists doubt that a marginal tax rate close to 100% will negatively affect effort, and that for example, one of the problems associated with the socialist system was that earnings were not closely related to effort (Kornai, 1992).  But it is less clear whether the marginal tax rates commonly observed today have that effect or whether those effects are outweighed by countervailing benefits of lower inequality (e.g. Diamond, 1998, Diamond and Saez, 2011).  Similarly, it is largely an empirical question whether inequality is promoting domestic savings as the rich may not save more than the middle classes and might take their monies out of the country (e.g Ray, 1998; Ndikumana, 2014). 

166

3.3.2.2 Negative impacts of inequality on economic performance

167

Following Voitchovsky (2009)[13], the theoretical channels of how inequality influences “the size of the pie”, that is, income levels and their growth rates, can be broadly grouped into four types of arguments relating to different parts and aspects of the income distribution: (1) circumstances of the poor, (2) the overall distance between individuals, (3) wealth concentration, and (4) size and circumstances of the middle class. The arguments can also be grouped by topic, which is the second dimension along which one can structure the arguments. For several of the identified channels, strong institutions are important in shielding the economy from adverse effects of inequality, and inequality itself can play a role in weakening those institutions. That inequality and the quality of institutions are intimately related has been suggested by a number of studies, including those by Engerman and Sokoloff (2002), and Easterly (2007).

168

Circumstances of the poor

169

Credit constraints

170

The poor are subject to credit constraints, which leads to foregone investment opportunities, and hence foregone economic growth (e.g. Birdsall 2006, Ghatak and Jiang 2002). That better financial systems lead to higher economic growth is well established (for a review of the relevant literature, see Levine 2004). Beck, Demirgüc-Kunt, and Levine (2004) find that financial development also leads to lower inequality and poverty alleviation. Credit constraints for the poor are particularly detrimental if they hinder investments in education, which may lead to long-term economic opportunities foregone, and to intergenerational poverty traps (e.g. Galor and Zeira 1993, Piketty 1997, Grossmann 2008). Perotti (1996) finds empirical evidence in favor of this claim, but suggests that other channels may be more important to explain the negative impact of inequality on growth.

171

Fertility

172

If earnings are low at the bottom end of the distribution, opportunity costs of having children are low. In the absence of a functioning social security pension system, parents have to rely on their family and children to take care of them at old ages. Additionally, if the expected labor market earnings of the child are low as well, it becomes more important to have several children to make sure there are enough means, later on, to provide for the elderly. Parents face a “quality-quantity trade-off” between the number of children and the amount of resources spent on each child.  The demographic transition, or fertility transition, has become an important concept in explaining economic growth and human development (e.g. Ahituv 2001). Its effect on growth is twofold: On the one hand, a decrease in population growth implies a higher per capita capital stock as well as per capita income. On the other hand, investment in children’s education also increases the human capital stock, which is crucial for economic growth. Finally, like most arguments pertaining to the circumstances of the poor, there is a self-reinforcing mechanism of a large unskilled labor force squeezing wages at the low end of the distribution, which lowers opportunity costs of children for the unskilled even more (e.g. Kremer and Chen 2002). Empirical analyses by Perotti (1996), Barro (2000), and de la Croix and Doepke (2003) support the fertility channel of inequality on growth.

173

Social cohesion, political stability, and civil unrest

174

Generally, higher socio-economic polarization has been shown to lead to higher income and wealth inequality (Mogues and Carter 2005). Ferroni, Maeto, and Payne (2007) explore the links between trust, inequality, and social cohesion in Latin America and find that social cohesion is positively linked to economic growth and growth-enhancing institutions in general, and negatively associated with inequality and low social capital in the form of trust. Generally, all studies on the topic agree that social capital and trust, social cohesion, political polarization, and inequality are simultaneously determined and reinforcing one another, making it a challenging task to pin down causalities of these broad concepts going in only one direction. More palpable aspects of social cohesion and political stability such as crime, political polarization, and lobbying, are easier to empirically capture and analyze.

175

From an economics of crime perspective, higher inequality leads to more crime. In particular, it makes (property) crime more attractive: on the one hand, there is a lower opportunity cost attached to it because the lower the wages, the less there is to gain at the bottom end of the distribution, and, on the other hand, there are higher expected gains because there is more to steal from the top end of the distribution (e.g. Chiu and Madden 1998, Josten 2003). Besides the foregone production of those committing crimes, one can think of further unfavorable consequences such as spending on unproductive protection mechanisms from the rich, and lower incentives for the overall population to engage in legal (as opposed to illegal) activities due to the higher risk of being robbed of the returns. This would again lead to higher crime rates, leading to a vicious cycle of higher crime and lower growth. A meta-analysis of existing time-series analyses on the effects of inequality on crime (Rufrancos et al. 2013) concludes that rising inequality, in particular at the bottom end of the distribution, is robustly associated with increasing property crime. That crime and social capital are interconnected has been widely recognized in the literature, and several authors have found that crime adversely affects social capital (e.g. Messner, Baumer, and Rosenfeld 1999, Liska and Warner 1991), with further implications on economic growth and development.

176

Distance between individuals

177

How far individuals or groups in a society are from each other in economic terms can have important repercussions on growth via the formation of social capital and trust. A large recent literature (most prominently by Nobel-prize winner Elinor Ostrom (1990)) has established that social capital and trust can help overcome prisoner’s dilemma-type situations and thereby increase overall cooperation within a society, which can have positive impacts on a wide range of outcomes, ranging from business transactions and technology adoption to improved health and education.

178

If very large, the distance between individuals can also have explicit negative consequences for growth via social unrest and the social political polarization of society (see e.g. Keefer and Knack 2002, Easterly 2001). A very polarized political landscape can have adverse macroeconomic consequences such as political stalemate for important policy reforms, political budget cycles, and increased uncertainty leading to a less favorable investment climate. Empirical evidence by Loaysa, Rigolini, and Llorente (2012) furthermore suggests that a large middle class has positive impacts on growth via improved quality of governance regarding democratic participation and official corruption.

179

Wealth concentration

180

An unequal distribution of income with high “top” inequality can be detrimental to growth by making it easier for the elite to capture institutions and bias the economy in their favor (see e.g. Glaeser Scheinkman and Shleifer 2002). An example can be the provision of exceptions for the very rich in contributing to public goods and services such as health care and education. It is also detrimental to the process of democratization, thereby perpetuating the weak institutions – inequality – low growth link.

181

Financial Crises

182

In the wake of the global financial crisis that started with the crisis in the US sub-prime mortgage finance sector, the question whether inequality promotes unsound financial practices that can promote financial crises (with negative growth implications) has received some attention.

183

This subject has received intense scrutiny recently with some observers arguing that the recent global financial crisis had rising inequality (particularly in the US) as one of its key causes.  Rajan (2010) and Kumhof and Ranciere (2011), Perugine et al. (2015), Stockhammer (2015) and Lysandrou (2011) argue that this is the case, while Bordo and Meissner (2007) dispute this link. 

184

Size & circumstances of the middle class

185

Domestic demand

186

Domestic demand is a crucial factor determining economic growth (at least in the short to medium term), and is typically associated with a strong middle class, implying a (relatively) equal income distribution with relatively few poor (who cover their basic needs and do not demand technologically sophisticated products) and relatively few rich (who primarily demand luxury goods) (e.g. Zweimüller 2000, Foellmi and Zweimüller 2006, Murphy, Sleifer and Vishny 1989). Note that in this context, redistribution from the rich to the poor would be growth-enhancing since it boosts demand. For a more detailed survey of the demand-side type of arguments, see Erhart (2009).

187

Political economy

188

Another well-known channel of how inequality and growth are linked is the median voter theorem, and related political economy arguments. Basically, the higher the inequality in a society, the lower is the income of the median voter and the higher his preference for redistribution, thereby lowering growth through high marginal taxation which reduces effort (e.g. Bertola 1993, Alesina and Rodrik 1994, Perotti 1993). Naturally, this channel would only work in democracies, which is also one reason for distinguishing between democratic and non-democratic countries in empirical analyses. It should be noted that the median voter theorem has been challenged not only on empirical grounds (e.g., Perotti 1996, Deininger and Squire, 1998; Milanovic 2000), but also theoretically. For example, Bénabou (2000, 2002) argues that if the rich have more political power than the poor, they may lobby against redistribution measures, even if these may be efficient. That the effect of redistribution on growth crucially hinges on the kind of taxation used for redistribution has been recognized by others as well. Proponents of the Scandinavian type welfare state argue that redistribution influences levels of social inclusion of the less privileged (for example, through education) and enables society as a whole to benefit from their talents. In particular, public investment in health and education (Easterly 2007) and taxes on activities involving negative externalities (such as excessive risk-taking in financial markets) (Stiglitz 2012) can be considered growth-enhancing (IMF 2014). One of the few studies examining inequality, redistribution, and growth simultaneously, a recent report by the IMF (2014) challenges this proposition using newer data and finds that higher inequality is indeed associated with more redistribution, but concludes that redistribution nevertheless is not harmful for growth. Quite to the contrary, it leads to both higher growth and a longer duration of growth spells. Notwithstanding, one should keep in mind the peculiarities of the underlying data and the fact that the overall literature is far from conclusive about the overall effects of the inequality-redistribution channel on economic growth.

189

The empirical impact of inequality on growth

190

In the mid-1990s, the empirical debate was significantly enhanced by the availability of a much broader set of data on inequality across the world. By now, there are many studies on the subject and we also draw here on existing reviews (e.g. Neves and Silva (2014).  Initially, the workhorse dataset was created by Deininger and Squire (DS1996) and used in a study by Deininger and Squire (1998) to show that, in a cross-section of countries, higher initial inequality (particularly of assets but, in some specifications, also of income) was associated with subsequent lower growth.  Generally, cross-sectional studies (Alesina and Rodrik 1994, Persson and Tabellini 1994, Clarke 1995, Perotti 1996, and Deininger and Squire 1998 tend to find a negative relationship between inequality and growth, whereas panel analyses yield mostly positive or insignificant results.

191

Ensuing debates focused on the one hand on weaknesses in the data, where Atkinson and Brandolini (2001) showed that the comparability and consistency of the (DS1996) data set were open to question. In addition, Knowles (2005) argues that most evidence on the growth and inequality relationship in cross-sectional studies is derived from inequality data which are not fully comparable. Once the heterogeneity in the underlying income concepts is accounted for, he concludes that there is no remaining relationship between income inequality and growth, but that inequality in expenditure is still negatively correlated with growth. Since then, the World Income and Inequality Database (WIID) was created which significantly enhanced not only the coverage but also the transparency of the inequality data used. Many studies on inequality have since relied on this dataset, where some authors used regression-based adjustment methods to address inconsistency issues (e.g., Gruen and Klasen 2008, 2012; Easterly 2007). Nevertheless, the dataset remains heterogeneous in terms of the underlying monetary concept (covering not only different types of income (net, gross, wage incomes) but also consumption and expenditure), the measurement unit (household vs. individual), and the type of equivalence scale used for adjusting household-level data, amongst other dimensions. As pointed out by Atkinson and Brandolini (2009), it is often not sufficient to account for these differences using dummy variables for each category underlying the data that are being used in a regression, as has been frequently done in the literature. Doing so implicitly assumes that the differences between the types of unit remain constant over time, which has been shown to not be generally true. More recently, Solt (2016) has, based on the latest version of WIID, used imputation techniques to also attempt to address data gaps and consistency issues in his Standardized World Income Inequality Database (SWIID). This approach has also been criticized (Jenkins 2015), but these data have been used in a number of subsequent studies (e.g. Scholl and Klasen, 2016, Ostry et al. 2014).

192

Endogeneity (i.e. caused by reversed causality or a spurious correlation related to an omitted third variable, so-called unobserved heterogeneity) also remains a problem in these cross-sectional studies.  The only cross-sectional study explicitly addressing the endogeneity problem is Easterly 2007, who instruments inequality with a country's wheat-sugar ratio, which is a function of the fraction of land suitable for growing wheat over the fraction of land suitable for growing sugar cane. The idea is based on the hypothesis by Engerman and Sokoloff 1997 that agricultural endowments predict a country's institutional environment. More specifically, growing sugar cane is more prone to large-scale farming involving slave labor, which leads to higher inequality and extractive institutions, whereas wheat production involves family farming and is associated with the emergence of a middle-class and less inequality. Instead of growth rates,(Easterly, 2007) then shows that higher inequality is associated with lower income levels, as well as worse institutions, and lower education. Most of the cross-sectional results should be viewed with caution because they may contain substantial omitted variable bias, given that any unmeasured factors which are associated with both inequality and growth can be wrongly attributed to an effect of inequality on growth.

193

Although panel data are not able to perfectly resolve this issue, the possibility of introducing fixed effects allows the removal of at least the time-invariant portion of the omitted variable bias, which is also the main explanation for the divergence in findings between cross-sectional and panel studies.

194

Forbes (2000) was the first to move towards a panel setting for two reasons. First, she wanted to address unobserved heterogeneity through fixed effects (and endogeneity through the use of GMM-type methods). Second, a fixed effects specification which exploits the within-variation is also the more policy-relevant question, as policy-makers are interested in whether changes in inequality in a country will promote or hurt subsequent growth. This approach came at the cost of using rather short panel periods of only five years. Essentially, this time span implies examining the short-term impact of changes in inequality on growth. While interesting, it is not so closely related to the theoretical literature which generally focused on longer-term impacts of inequality on growth (e.g., Galor and Zeira 1993; Alesina and Rodrik 1994. Forbes found that rising inequality is associated with higher subsequent growth, although the result is not significant when 10 year periods are used.

195

The paper by Forbes attracted a lot of debate and commentary. Apart from the above mentioned data issues (her analysis was based on the DS dataset), there was the concern that the use of fixed effects takes out most of the variation in the dataset and that the little within variation might be heavily affected by measurement error (Knowles 2005). Second, there was concern about the functional form. In particular, Banerjee and Duflo (2003) argued that the data are more consistent with the claim that any change in inequality (whether positive or negative) is associated with lower subsequent growth, which is, of course, a rather different interpretation. In addition, Roodman (2009) demonstrates that Forbes results become insignificant once the econometric issue of overidentification is being addressed, which is also confirmed in Scholl and Klasen (2016).

196

There have been further debates on this issue, but the question how inequality affects growth in a panel setting remains largely open. Another widely cited study, Barro (2000) finds, for a sample of 40 to 70 countries and 10-year time periods, that higher inequality leads to lower growth in poor countries and higher growth in rich countries, but there is little overall relationship between income inequality and growth. He refrains from using fixed effects in his preferred specification and points to the exacerbation of measurement error with this approach, but his results from a three-stage least-squares estimation do hold qualitatively in a fixed effects specification, although the latter is only able to capture the contemporaneous relationship between inequality and growth.

197

Deininger and Olinto (2000) focus on asset instead of income inequality in their panel of 60 countries, and find a negative and significant relationship with subsequent growth rates. In addition, they confirm the positive relationship with income inequality as found in previous studies, which continues to hold even when asset inequality is retained in the model. Ezcurra 2007 looks at annual regional growth across the European Union over the 1993-2002 period and concludes that higher inequality is associated with lower growth, thereby contradicting (Barro, 2000) who found that inequality is positively related to growth in rich countries although the differing results of the two studies could also be due to the different time frames they consider.

198

Ostry et al. (2014) from the IMF simultaneously examine the impact of initial inequality as well as redistribution on subsequent growth in a panel setting.  They find that initial inequality is associated with reduced growth in the next 5 years, that redistribution has no impact on growth, and that higher inequality shortens the duration of growth spells. Not enough information is provided on the details of their econometric methodology (using system GMM) to fully assess the robustness of the findings.  A study by Cingano (2014) from the OECD focusing only on OECD countries and using particularly high-quality inequality data also finds that initial inequality is associated with lower subsequent growth in this set of countries, with inequality at the bottom of the distribution having a particularly pronounced effect.

199

Scholl and Klasen (2016) re-examine the data and approach of Forbes (2000) using an expanded and updated panel data set, including many more and (arguably) improved data on inequality and using advanced econometric methods and tests.  They find that inequality has a small positive impact on growth but this is entirely driven by the experience of transition countries in the 1990s; when controlling for this, no significant effect emerges.  The most important difference to the studies from the IMF and OECD studies is the use of a lag of inequality rather than its value at the beginning of the period.  Thus a seemingly small decision about the time structure of the effect can affect the results substantially. 

200

Other approaches have used time series methods to assess the link.  For example, Herzer and Vollmer (2012) study the long-run relationship between inequality and growth in a panel of countries using panel cointegration techniques, which account for the fact that inequality is an endogenous variable. They conclude that inequality is harmful to growth and that the effect is quite sizeable.

201

In a recent empirical paper, Andrews, Jencks, & Leigh (2011) test the “trickle-down” hypothesis. They use unbalanced panel data for twelve developed countries and find that after 1960, a one-percentage-point rise in the top decile’s income share is associated with a 0.12-point rise in GDP growth in the following year—a small effect. Nevertheless, the result confirms the hypothesis that inequality in the form of larger shares for the top of the income distribution is good for growth. Their fixed-effects specification however does not take into account that top income shares are plausibly both exogenous and endogenous to economic growth. Thus, their estimates are quite likely to be biased. Herzer and Vollmer (2013) employ heterogeneous panel co-integration techniques, which are robust to omitted variables, slope heterogeneity, and endogenous regressors, to examine the long-run effect of top income shares on income per capita and thus long-run growth. They find for a panel of nine high-income countries that the effect of an increase in top income shares on economic growth is negative, and, moreover, that causality indeed runs in both directions, not only from top income shares to growth, but also from growth to top income shares. Hence, not only is there no evidence for a trickle-down effect, but in fact, top inequality seems to slow down overall growth.

202

In sum, results from reduced-form panel studies using different sets of methods are heterogeneous and despite the continuous improvement of the inequality data since DS1996, data issues as well as concerns about functional form and appropriate estimation techniques keep being raised in the literature. While there seem to be more panel studies in recent years that have found a negative effect of inequality on growth, those results are often driven by particular methodological choices, data, time periods, and lag structures.  At the same time, it appears clear that the initial finding of a positive effect of inequality on growth is an outlier and related to a particular empirical set-up.  Thus on average one should not expect a trade-off between inequality and growth.  At the same time, it is quite plausible that these average effects hide substantial variation.  In line with the theoretical literature, it can well be the case that certain forms of inequality change (e.g. investments in human assets of the poor) will promote growth while others might hurt it (e.g. arbitrary seizures of assets).   Unfortunately, at present, there is insufficient empirical literature on these different mechanisms.

203

3.3.2.3 The impact of inequality on ‘disaffection,’ social movements

204

In understanding the dynamics that produce or reproduce inequalities it is important to recognize that inequalities can help to cause social phenomena that in turn either diminish or reinforce inequalities.  There is no mechanical relationship between inequality and the presence or absence of social movements, and in particular ones that seek to contest or address inequality.  Nevertheless, it is possible to identify some plausible dynamics, as well as empirical cases. 

205

Consider first the impact of inequality on social movements.  Do stark inequalities accentuate the propensity to organize against them? Sociologists and political scientists have long considered the problem of whether absolute disadvantages or relative disadvantages may act as triggers for social protests[14].  This question, in turn, can be sub-divided into the question of whether it is the level of disadvantage (conceived in either of these ways) or the change in disadvantage that matters for social protest.  Unfortunately, the literature on these questions is inconclusive.  One important reason why is that unequal situations that persist over long periods of time may tend to generate cultural and ideological frameworks (e.g. ‘adaptive preferences’[15]) that rationalize inequalities and quell social discontent. Another important reason why is that in unequal situation those who benefit from inequality may ‘invest’ in strengthening the political forces (e.g. political parties, or non-party political formations including persuasive or coercive forces) that help to sustain or indeed widen inequalities, and by definition have the resources to do so to a greater extent than others.   Inequalities may also make for societies with more heterogeneous interests, across which collective action is more difficult.[16]  On the other hand, among those who share a common interest, the incentive for collective action may become stronger.  In most contemporary societies, however, the problem of large numbers of persons who often have no ready method of coming together to support shared interests makes such collective action difficult.[17]  The impact of inequality on the sense of ‘fellow feeling’ and of social anomie, with potential impact on collective action including in attacking inequalities, is also a potentially relevant factor that is little understood.

206

What about the impact of social movements on inequality? If the presence of social movements that contest inequality is in itself an important determinant of inequality, then to what extent is it so and through what mechanisms?  It would seem evident that social movements shape political outcomes that emerge and thereby the institutional and economic parameters (such as the role accorded to private property, the level of social investment in human capabilities, and the level of redistributive fiscal policies) that in turn influence inequalities.[18]  In recent years, the Occupy movement and other high-profile reactions to inequality have shaped mainstream political parties’ rhetoric and policy proposals, to varying extents (even if less than many advocates would have liked, notably in the UK and the US).  It is evident that the two-way relationship between social movements and inequalities can be important in determining what levels of inequality emerge and prevail over time.

207

3.3.2.4 Impact of inequality on behavior

208

Bandwidth

209

A central feature in the lives of people with low-income is that they do not have enough money to comfortably manage their basic needs. The poor must juggle and time their spending to coincide with sporadic income and expenses, they must continuously contemplate and weigh difficult, and often expensive, tradeoffs (Collins et al., 2010; Edin & Shaefer, 2015).  Financial services, from banks to informal lenders to rent-to-own stores, as well as landlords, and all kinds of bills, from utility to medical bills, impose high interest rates and steep late payment threats and penalties. A substantial percentage of the annual income of the poor is spent on just those juggling expenses, not to mention the hassle, the phone calls, and the long-term and costly penalties to your credit score. (Barr, 2014; Mullainathan & Shafir, 2009). It can be exceedingly expensive to be poor (which suggests actual inequality is often underestimated).

210

Not having sufficient funds means needing to devote substantial mental resources to preoccupations of everyday life in an attempt to make ends meet. Being poor means not only having less money – it means having to spend more of one's bandwidth managing that money. Our cognitive system, as it turns out, has severely limited capacity (Miller, 1956; Baddeley & Hitch, 1974; Neisser, 1976; Luck & Vogel, 1997).  A continuous preoccupation with budgetary concerns imposes demands on computation and attention that are bound to distract from other matters, resulting in diminished resources and in reduced performance (Mullainathan & Shafir, 2013).

211

Just as an air traffic controller who is focused on a potential collision course is prone to neglect other planes in the air, so do the poor, when attending to pressing monetary concerns, can neglect to give other problems their full consideration. The effects of cognitive load and diminished mental bandwidth can manifest themselves along several dimensions.  These include basic cognitive capacity -- the ability to solve problems, retain information, engage in logical reasoning, and so on -- and executive control, the ability to direct cognitive activities, overseeing the allocation of attention, planning, remembering to do things, initiating and inhibiting actions, self-monitoring and impulse control. (For a review and references, see Mani et al., 2013; Mullainathan & Shafir, 2013). 

212

Several other correlates of poverty may further contribute to diminished bandwidth, including dysfunctional institutions, dangerous neighborhoods, physical pain, noise, pollution, and sleep deprivation. Poverty also correlates with depression and anxiety disorders, negative affect, and stress -- defined as an organism’s reaction to environmental demands exceeding its regulatory capacity (WHO, World Health Report 2001; Haushofer & Fehr, 2014).

213

These constant and severe impositions on cognitive bandwidth can, in turn, lead to short-sighted and risk-averse decision-making, including the favoring of habitual behaviors at the expense of goal-directed ones (Haushofer & Fehr, 2014), and ranging from an increased proclivity to take short-term high-interest loans (Shah et al., 2013) to making more impulsive dietary choices (Shiv and Fedorikhin, 1999).

214

The challenges that come with poverty have long been observed to correlate with a plethora of counterproductive behaviors. (Mullainathan, & Shafir, 2013). Not only do the poor show low and often misguided participation in the financial mainstream (Bertrand, Mullainathan, & Shafir, 2006; Caskey, 2004), they also fail to take advantage of entitlement programs for which they are eligible (Currie, 2006), they engage in less preventive healthcare and fail to adhere to drug regimens (Katz & Hofer, 1994) they are tardier and are less likely to keep appointments, and they are less attentive parents (Karter et al., 2004; Lee & Bowen, 2006; Neal et al. 2001).

215

Poverty also correlates with low-literacy, which has its own consequences. Low literacy consumers tend to “satisfice” — settle for things that are "good enough" — based on limited information because of the challenges of digging deeper; consequently, they have a lower ability to avail themselves of various opportunities (Nielsen Norman Group). 

216

Stereotyping and affirmation

217

The stigma and stereotyping of poverty further impact perception and behavior. Research shows that the poor are scorned, perceived as incompetent, and disrespected (Fiske, 2011; Kerbo, 1976). That stigma carries with it the feeling of being viewed as a societal burden, lazy and unmotivated. This can lead to cognitive distancing (Reutter et al., 2009), and to “welfare stigma” (Horan & Austin, 1974; Rogers-Dillon, 1995), which purportedly causes the poor to forego important benefits offered in the public (Bissett & Coussins, 1982) and the nonprofit (Kissane, 2003) sectors.

218

When self-worth is threatened—when certain products or situations prove stigmatizing or intimidating, for example—defensive responding and concerns about being judged according to stereotypes, along with efforts to suppress negative thoughts and emotions, are presumed to consume executive resources (Schmader, Johns, & Forbes, 2008) and can disrupt performance (Spencer, Steele, & Quinn, 1999; Steele, 1997). In one study, for example, a test purportedly measuring intellectual ability was found to diminish the performance of students of lower socioeconomic status in a manner similar to that observed in other studies involving race or gender stereotypes (Croizet & Claire, 1998). When striving to provide products and services that might help the poor, poverty advocates thus face the additional hurdle of the stigma and stereotype threats associated with living in poverty.

219

A theoretical advance in the interpretation and manipulation of stigma and stereotype threat is self-affirmation theory, which posits that people are motivated to sustain a sense of self-worth and integrity (Steele, 1988). When aspects of the self, even those unrelated to the threat, are affirmed (e.g., when people are led to consider other, positive self-aspects), the need to sustain one’s sense of self-worth is met, and people respond less defensively to situations that otherwise would appear threatening (Aronson, Cohen, & Nail, 1999; Sherman & Cohen, 2006). Self-affirmation manipulations have been found to improve fluid intelligence and cognitive control performance among the poor, and to increase their tendency to resort to information pertinent to social benefits offerings (Hall, Zhao, & Shafir, 2014).

220

3.3.2.5 The impact of inequality on political participation

221

A mechanism through which economic inequality might have a substantial impact on society (in turn creating a ‘feedback loop’ relevant to its production and reproduction) is its possible impact on political participation.  The decision as to whether to participate in political processes (whether by voting, seeking to persuade or mobilize others in the society, to stand for office, etc.) may well be systematically shaped by economic inequality, especially if those with greater resources are able to deploy them in the political process, whether openly or covertly.  To the extent that those with greater command over resources are permitted to enter and shape the political process differentially, this may not only have an effect in determining one-off political outcomes but lead to processes of cumulative causation shaping the institutions that in turn govern day-to-day political decision-making.  Some of these institutions are ones open to public view (such as national electoral processes) and others are well hidden (such as administrative procedures governing the regulation of activities of specific industries, about which the general public may have limited awareness). The role of public vs. private financing of election campaigns of individual politicians and of political parties, the compensation of politicians affecting their ability to participate in politics without either being wealthy or being ‘paid off’ after their term in politics, and other institutional and policy factors may play a role in determining to what extent those with greater resources are able to shape political outcomes.[19]

222

The still low rates of political participation (e.g. voting) in many countries - including wealthy countries with established democracies - is illustrative.   While some voters may fail to vote because of the ‘rational’ calculation that their vote is unlikely to be decisive, others may not do so for other reasons, for instance because they believe that the major political alternatives are already substantially shaped by powerful economic or social interests - thus denuding their vote of consequential efficacy.  It is suggestive in this respect, although hardly definitive, that levels of trust in electoral democracy and in democratic institutions generally have been falling in certain countries in the same period that inequalities have generally been rising (most notably in the US, although there appears to be a somewhat different experience in European countries)[20].  Economic inequalities may also interact with gender and social inequalities to shape who is able to stand for public office, which messages are advanced more prominently, and what arguments are able to win the day in public policy making.

223

3.3.2.6 The impact of inequality on economic, social, and financial stability

224

A number of arguments can be subsumed under transmission channels related to political stability, social cohesion and civil unrest. These relate to almost all aspects of the income distribution: the circumstances of the poor, the overall distance between individuals, and the concentration of wealth.

225

Generally, higher socio-economic polarization has been shown to lead to higher income and wealth inequality (Mogues and Carter 2005). At the same time, inequality is a threat to social cohesion and political stability (Alesina and Rodrik, 2004).  For example, Ferroni, Maeto, and Payne (2007) explore the links between trust, inequality, and social cohesion in Latin America and find that social cohesion is positively linked to economic growth and growth-enhancing institutions in general, and negatively associated with inequality and low social capital in the form of trust. Generally, all studies on the topic agree that social capital and trust, social cohesion, political polarization, and inequality are simultaneously determined and reinforcing one another, making it a challenging task to pin down causalities of these broad concepts going in only one direction. More palpable aspects of social cohesion and political stability such as crime, political polarization, and lobbying, are easier to empirically capture and analyze.

226

From an economics of crime perspective, higher inequality is expected to lead to more crime. In particular, it makes (property) crime more attractive: on the one hand, there is a lower opportunity cost attached to it because the lower the wages, the less there is to gain at the bottom end of the distribution, and, on the other hand, there are higher expected gains because there is more to steal from the top end of the distribution (e.g. Chiu and Madden 1998, Josten 2003). Besides the foregone production of those committing crimes, one can think of further unfavorable consequences such as spending on unproductive protection mechanisms from the rich, and lower incentives for the overall population to engage in legal (as opposed to illegal) activities due to the higher risk of being robbed of the returns. This would again lead to higher crime rates, leading to a vicious cycle of higher crime and lower growth. Although there some cross-sectional (e.g. Kelly 2000) and longitudinal (e.g. Doyle et al. 1999) studies suggest that there is no association between inequality and property crime, a meta-analysis of existing time-series analyses on the effects of inequality on crime (Rufrancos et al. 2013) concludes that rising inequality, in particular at the bottom end of the distribution, is robustly associated with increasing property crime, thereby lending support to this theoretical channel (see also UNDP. 2013). According to the authors, the lack of findings in some studies can be attributed to missing control variables and flaws in the empirical specifications. That crime and social capital are interconnected has been widely recognized in the literature, and several authors have found that crime adversely affects social capital (e.g. Messner, Baumer, and Rosenfeld 1999, Liska and Warner 1991, UNDP, 2009), with further implications on economic growth and development. The same can be said about the overall distance between individuals.

227

The impact of income inequality on financial stability has received intense scrutiny recently with some observers arguing that the recent global financial crisis had rising inequality (particularly in the US) as one of its key causes.  Rajan (2010) and Kumhof and Ranciere (2011), Perugini et al. (2015), Stockhammer (2015) and Lysandrou (2011) argue that this is the case, while Bordo and Meissner (2012) dispute this link. 

228

3.3.2.7 Impact on conflict

229

There are two main theoretical lines of argument on the impact of inequality on conflicts: (1) the relative deprivation hypothesis (RDH), and (2) the resource mobilization hypothesis (RMH) (Muller, 1985). According to RDH, on the one hand, diverse forms of deprivation-induced discontent are positively associated with political violence.  Thus, it is not necessarily income inequality per se that triggers this discontent but rather the gap between an individual’s expected and achieved wellbeing. Testing this hypothesis requires qualitative data at the individual level, which is difficult to come by. However, most empirical research relies on polarization measures to proxy this relative deprivation (Thorbecke and Charumilind, 2002). Such measures tend to be positively related to the risk of conflict (Bueno de Mesquita, 1978; Keefer and Knack, 2002; Montalvo and Reynal-Querol, 2005). Thus, to the extent that it may lead to polarization, income inequality could indirectly result in conflict.

230

On the other hand, RMH draws attention to how the discontent generated by the inequality of resources, which could then lead to conflict. An offshoot of this hypothesis is the land maldistribution hypothesis, which posits that the discontent emanating from highly concentrated distribution of land and/or lack of land ownership in agrarian societies is a significant determinant of mass political violence (Thorbecke and Charumilind, 2002).

231

The empirics regarding the inequality-conflict nexus attempt to distinguish between land inequality and income inequality, and between horizontal inequality and vertical inequality. With reference to the former dichotomy, there is evidence in support of the view that land inequality tends to raise the incidence of conflicts (Russet, 1964; Huntington, 1968; Midlarsky and Roberts, 1985; and Binswanger et al, 1995). However, Muller and Seligson (1987) find that it is the inequality in the distribution of income rather than in land that primarily explains the rates of political violence across countries. And, André and Platteau (1998) fault both land inequality and income inequality for provoking violence (particularly in the 1994 Rwandan genocide).

232

Finally, even though horizontal inequality may not be associated directly with conflict, it stimulates grievances which intensify the likelihood of a rebellion (Gurr and Moore, 1997); in certain cases it actually facilitates the outbreak of conflict (Østby, 2008).  However, empirical research is yet to establish a significant relationship between vertical inequality and the risk of war onset (Collier and Hoeffler, 2004; Fearon and Laitin, 2003).  Instead, the literature tends to find a stronger relationship between polarization of groups and conflict, suggesting that the type of inequality matters significantly for conflict (e.g. Esteban and Ray, 2011).

233

3.3.2.8 The role of middle classes and top earners in affecting economic performance

234

Apart from the debate about the impact of inequality of economic growth, a sizable literature has developed on the effects of particular types of inequality.  Particularly prominent in these debates has been the role of middle classes in affecting economic development.  Starting with Easterly (2001), it has been argued that a sizable middle-class with a significant income share can promote economic development through several mechanisms.  Middle classes may have particularly large savings rates (and are more likely to save domestically rather than take their money out of the country as richer groups might do), they are advocates of political and social stability, transparency, the rule of law, and meritocracy, as all of these aspects will promote upward mobility, a key aim of middles classes.  At the same time, these features are also growth-enhancing.  While there is a large sociological literature on the role of middle-classes in advanced economies (e.g. Wright, 1979; Goldthorpe, 1987), in the context of developing countries, a literature has developed on how to measure the size and fortune of the middle class, and assess its consequences for economic development.  On the measurement front, researchers have either borrowed from the literature on particular traits of households to define them as middle-class which usually then implies that they belong to the upper parts of the income distribution in many poorer developing countries. Others have focused on those in the middle of the distribution, which may be very poor by Western definitions; others have applied absolute cut-offs which.  Overall, the literature on the definition, size, and role of middle classes in developing countries is still quite inconclusive (e.g. Banerjee and Duflo, 2008; Giesbert and Schotte, 2016; Lopez-Calva and Ortiz-Juarez, 2014).

235

3.3.2.9 The impact of gender inequality on growth

236

There has been a growing literature examining both the determinants as well as the consequences of gender inequality in education and employment on economic growth.  In particular, a number of theoretical papers have emphasized how gender inequality in education and employment can reduce economic growth (e.g. Lagerlöf, 2003; Rees and Riezman, 2012; Galor and Weil, 2000, Teignier and Cuberes, 2015).  On the empirical side, a number of studies have argued that gender inequality in education reduces economic growth (e.g. Knowles et al. 2002; Yamarik and Ghosh, 2003; Klasen, 2002; Klasen and Lamanna, 2009).  Here the evidence appears to be quite robust, although most is drawn from cross-country econometric studies.  There is similarly also some evidence that gender inequality in employment can reduce economic growth (e.g. Esteva-Volart, 2009; Taignier and Cuberes, 2015; Klasen and Lamanna, 2009); the literature here is smaller and less conclusive.  The quantitative importance of these gender gaps for economic performance, as well as the precise transmission channels on how gender gaps affect economic growth, are less clear, with some arguing that the growth costs of gender gaps are very large (e.g. Taignier and Cuberes, 2015; King, Klasen, and Porter, 2009) and others suggesting these gaps do not have such large consequences (Duflo, 2012) or might not even have clear impacts (Bandiera and Natraj, 2013) or could even be beneficial for growth (Tertilt and Doepke, 2014).  Lastly, there is a literature that has shown that gender gaps in access to land and other resources for agricultural production reduce agricultural productivity in some settings (Udry, 1996; Goldstein and Udry, 2008).

237

3.3.3 Interlinkages of inequalities in incomes, education, and health

238

Since there are well-documented causal relationships in both directions between income, education and health, inequality in income will likely promote inequality in education and health, unless specific measures (often public action) are adopted to break that link.  Thus it essentially boils down to investigating the importance of these linkages from an empirical point of view.  The literature generally addresses this issue using two different approaches: one is to study the actual correlation between for which there is a large literature.  It finds that there is a substantial correlation between income inequality, educational inequality, and health inequality (e.g. Checchi, 2001; Deaton, 2003).  While this correlation is hardly disputed and there are good theories to explain causal links in both directions, less obvious is the impact of income inequality on average health or education outcomes.  But here, most evidence points to higher income inequality leading to worse average health and education outcomes.  For example, a review by Pickett and Wilkinson (2015) conclude that the literature points to a strong causal impact of income inequality on average health outcomes.

239

The other common approach to this topic of interlinkages is to actually study the joint distribution of income, health, and education.  In the cross-country context, this was, for example, done by Grimm et al. 2008, Grimm et al. 2010, and Harttgen and Klasen (2012).  They find a strong correlation between income inequality, educational inequality, and health inequality.  They also find that inequalities in non-income dimensions are smaller the higher the overall levels of achievements in those dimensions. The conceptual and empirical literature on multidimensional inequality measures also discusses ways to capture the joint distribution of income, education, and health inequality (e.g. Aaberge and Brandolini, 2014).

240

4. Global inequality: Trends and drivers

241

 Estimation of global income inequality between people requires two basic types of data, namely data on income levels for all countries and how incomes are distributed within each country. Data on levels can be obtained either using GDP per capita data from national accounts or household income (expenditure) surveys; and distribution data can be obtained from social tables or household surveys or assuming some functional form for the country’s distribution or by proxying the income distribution using distributions of some other variable like human height. Data on levels and the distribution of incomes have only recently been available for sufficient countries to make an estimation of global inequality a possibility. To calculate global inequality in the long-run going back 60 or more years is far more difficult due to lack of data and, consequently, relies on making many assumptions with unreliable data[21]. Figure 3.3 presents estimates of global inequality and the between country contribution to total inequality as calculated by various studies from 1820 to 2013. These estimates are not always comparable as data and assumptions made vary across studies, but they are the best available estimates of global inequality over the long run. We first discuss in brief the data and methods used by different studies and then present the trends and drivers of global interpersonal inequality.

242

The estimates from 1820 to 1992 are from seminal work done by Bourguignon and Morrisson (2002). For information on income levels of countries, they use data from Maddission (2001) and distribution data is based on 33 income distributions of varying quality. In cases where distribution data was missing for countries they were either assumed to have the same distribution as “similar” countries or 20th-century distributions were used to predict prior distributions.  Van Zanden, Baten, Foldvari and van Leeuwen (2013) improved on Bourguignon and Morrisson’s method for estimating distributions in cases where they were missing. In cases where distribution data was missing, they used either the evolution of unskilled wage/GDP ratio or data on individual heights to estimate income distributions rather than relying on interpolation or grouping of countries to predict distributions. Milanovic (2011) on the other hand uses social tables from thirteen 18th and 19th century countries to estimate global inequality for the early 19th-century.  All these studies use data on GDP per capita from Maddisson (2001) for income levels information for late 19th and 20th century. A major drawback of the analysis based on historic data, in addition to being based on sparse and uneven quality of data, is that it uses 1990s purchasing power parity[22] (PPPs) to convert local currency into one international numeric currency. PPPs depend on the consumption pattern and size of consumption of various countries in the year when they are estimated and cannot reliably be used for temporal comparisons and do not represent the purchasing power of currencies over time (Reddy and Pogge 2002).

243

For the more recent periods from 1960 to 2013 we present global inequality data from the Global Consumption and Income Project (GCIP) and also compare estimates for recent periods from other studies. Household income/consumption surveys that have become frequent across countries only in the recent period have been used to estimate global interpersonal inequality since 1980s. These estimates differ in levels from the ones using national accounts discussed so far, but the broad trends are similar across the two. Milanovic (2002, 2005, 2012), Lakner and Milanovic (2013), Anand and Segal (2015), Jayadev, Lahoti and Reddy (2015) are some studies that use levels and distribution information directly from household surveys to estimate global interpersonal inequality. Some other studies use survey data to obtain distribution information but “scale” the within-country distributions to national accounts estimates of mean income or consumption (Chotikapanich et al. 2012; Dowrick and Akmal 2005; Sala-i-Martin 2006; and Schultz 1998). The levels and the rate of change of global inequality and its components can vary substantially between using household survey means vs. scaling up to national accounts means (Deaton 2005). Anand and Segal (2008) and Atkinson and Brandolini (2001, 2010) provide a detailed discussion of the various choices involved and their implications on measurement of global inequality using household surveys. Anand and Segal (2008) analyze a range of studies estimating global inequality for their methodology and the type and quality of data used to construct the global distribution. They conclude that the confidence intervals for global inequality measures would be large due to various sources of uncertainty in the estimates arising from both measurement and estimation problems making it difficult to conclude anything about the direction of global inequality for the period between 1970 and 2000. Apart from the standard survey sampling based errors global estimates also have to contend with errors arising from measurement of PPP’s which are not quantified. In Table 3.4, we present global estimates from several different studies from 1988 to 2008 to give a sense of the range and trend of global inequality as estimated by different studies.

244

In our discussion below, we restrict our discussion to survey-based estimates by Global Consumption and Income Project (GCIP) from 1960 to 2013, Lakner and Milanovic (2013) from 1988 to 2008 and Anand and Segal (2015) from 1988 to 2005. GCIP provides the longest estimates of global inequality based on survey data and standardizes welfare concepts measured in surveys to provide pure income based estimates of global inequality (Lahoti, Jayadev and Reddy 2014). Lakner and Milanovic (2013) provide estimates for benchmark years between 1988 and 2008, and mix consumption and income surveys based on availability of surveys in the country. Anand and Segal (2015) use the same data as Lakner and Milanov (2013), but augment data with top incomes information for various countries to account for the underrepresentation and underestimation of the top end of the distribution in surveys. Though the level of global inequality varies across the three studies, the trends and patterns are qualitatively similar.

245

4.1 Trends in between country and within-country inequality in incomes

246

Looking at the shape of the overall global income distribution, the world income distribution has been found to be unimodal and basically log normal between 1820 and 1913. By 1950, this unimodal distribution switches to a bimodal distribution with the emergence of twin peaks that get pronounced up to 1980s. The distribution gradually again shifts to a unimodal distribution by 2010 (van Zanden, Baten, Foldvari and van Leeuwen 2013, Jayadev, Lahoti and Reddy 2015).

247

Global interpersonal inequality estimates using historical data shows that inequality was high (a Gini coefficient of about 0.5) even in early 19th century when the industrial revolution was in its early stages in the western world[23]. Inequality increased substantially until World War I – 1913 – with the spread of industrialization and rapid growth in Western Europe, Americas and western offshoots (US, Canada, Australia and New Zealand). An important feature of this period was a sharp increase in the between-country component of global inequality (inequality that would exist if each individual had the mean income of his/her country). Using Maddison data, Milanovic (2011) estimates that between-country inequality doubled from 15 to 32 Gini points between 1820 and 1870. Bourguignon and Morrisson (2002) find that between-country component’s contribution to overall global inequality increased from only 12 percent in 1820 to 40 percent in 1910.  Pritchett (1997) shows that in the period between 1870 to 1990 divergence between developed and developing countries widened while there was convergence between developed countries. The ratio of per capita incomes between the richest and poorest countries increased by about a factor of five due to different growth experiences of the two set of countries.

248

Figure 3.3: Long-term trend in global inequality

249

250

Source: Bourguignon and Morrisson (2002) and Jayadev, Lahoti and Reddy (2015)

251

Global inequality then stabilized until the Second World War and then increased but at a slower pace and peaked around the level of 0.65 in the 1980’s.The between component of global inequality shows an increase until 1955 when it peaked and accounted for 60 percent of total inequality which remained stable until the late 1980’s according to national accounts data analysis by Bourguignon and Morrisson (2002). Overall, there was a large increase in inequality in the 19th century, followed by steady but slow increases until the 1950s and relative stabilization after that period until the 1990s. This was accompanied by a complete reversal in the roles played by within and between components of total inequality. The between component of global inequality contributed only a small portion of total inequality in 1820’s but this has increased rapidly to play a bigger role than within country inequality in the mid 20th century. The period after the 1950s to the 1990s saw a relative stabilization in the contribution of each of these components.

252

GCIP estimates that consumption inequality was around 0.69 Gini points in 1960 and stayed steady until 2005 when it was 0.68, but has seen decline since to 0.63 in 2013 based on household survey data. Lakner and Milanovic (2013) also estimate similar levels and trends in inequality during the period from 1988 to 2008 (0.72 to 0.70), with only a slight decline in overall inequality levels[24]. The within-country component of inequality and the total inequality increases when household surveys are augmented with information for top incomes by 2-4 gini points, but trends remain similar (Anand and Segal 2015). The changes in the recent period seem small and insubstantial as compared to the increase in global inequality during the 19th and early 20th century.

253

At the same time, the within country inequality has seen rapid increases in the recent period with the population-weighted average of within-country income inequality increasing from 0.42 in 1990 to 0.47 in 2010 according to GCIP. In contrast, the between-country component of total inequality witnessed declines in the same period due to the faster growth rates of China since the 1980s and more recently of most countries of the developing world in the 2000 decade. Estimating global inequality and the between country inequality excluding China from the world eliminates most of the decline in both of these measures.

254

Other global inequality measures like the Mean Log Deviation (MLD) and Theil show more substantial declines than the Gini during the recent period. This is likely because of different sensitivities of these measures to changes in different parts of the income distribution[25]. Though all relative inequality measures show a small decline or no substantial movement in the last two decades, absolute inequality measures have substantially increased for most of this period. Absolute inequality measures register increase as long as the absolute income increase going to a poorer person is lower than that going to a richer person. Absolute Gini has increased throughout the period, while the absolute Theil and the absolute mean log deviation have seen relative stability since 2005 after large increases in the prior period (Jayadev, Lahoti and Reddy 2015, Anand and Segal 2015).

255

One other way to measure differences in the world is by estimating differences between geographical regions. Maddison (2001) defines a regional dispersion measure as the ratio of income per head (GDP per capita) of the richest to the poorest region of the world to look at regional divergence for regions as defined in Table 3.1. This ratio increased from 3:1 in 1820 to 9:1 in 1913, to 13:1 in 1950 and 19:1 in 1998. It experienced a small period of decline between 1950 and 1973 from 13:1 to 12:1. In 2010 this ratio saw a decline to 15:1 from 19:1 in 1998 due to faster growth in the poorer developing regions in this period (Table 3.1).

256

4.1.1 Proximate Drivers of Global Inequality

257

Here we will only talk about proximate determinants of changes in global inequality.  What drives these proximate determinants, esp. the role of growth differences between countries, will be examined in detail in Chapter 4. 

258

Differential income growth rates across regions of the world, within region differences among countries and within country inequality are the major drivers of global inequality. They bring about divergence or convergence depending on their movement and their importance also varies by time and region. Population growth, another possible reason for changes in global inequality does not empirically play a major role in affecting changes in global inequality in the past although that could well change in future since the divergence in population growth rates between many poor African countries and richer developed and industrialized countries has been rising substantially in the last two decades.

259

The divergence between countries in the 19th century and first half of 20th century was driven partly by the relatively slow growth rate in Asian countries. According to Maddison (2001) between 1820 and 1950 income per capita in Asia grew at an annual rate of 0.2 percent that is six times slower than the growth rate of Western Europe and its offshoots (Table 3.3). The two most populous countries China and India were major contributors to this slow growth of Asia in this period.  The share of Asia in world GDP declined from 59.2 percent in 1820 to only 18.5 percent in 1950, while the share of western offshoots increased from 1.9 percent to 30.6 percent in the same period (Table 3.2).  There has a been sharp role reversal in the recent period from 1998 to 2010 with Asia witnessing a per capita growth rate of about 5 percent with China and India driving the growth and Western offshoots and Western Europe (the richest regions) growing the slowest. This drove convergence in world economies and global inequality, and especially the between country component experienced some decline. Even before this since 1950’s Asia has been growing faster than world average with different countries (Japan, Korea, Taiwan, Malaysia, Indonesia, Thailand, Vietnam, China and India) leading growth in various periods.

260

Western European countries saw strong and sustained growth for more than a century (growing at about 1.2 percent per annum from 1820 to 1913, Table 3.1) after starting off the industrial revolution, thereby acting as an important source of divergence in this period. The share of world GDP coming from these countries increased from 23.6 percent in 1820 to 33.5 percent in 1913, when it peaked. The world wars and the slower relative growth of the region in the inter-war period led to decline in contribution of the region to world GDP from 33.5 percent in 1913 to 20.6 percent in 1998 that declined further to 17.1 percent in 2008. Inequality across countries in Western European and their offshoots increased initially from 1820 to 1870 but was a substantial equalizing force from 1910 to 1992. This was partly driven by the strong recovery and catch up by Western European economies with the United States after the Second World War.

261

Latin America grew at a steady pace close to the rate of world average growth until the 1970’s and saw its share in world GDP increase from 2 percent in 1820 to 8.7 percent in 1973. The region has experienced lower growth rates since then as compared to other regions of the world and its share in world GDP has dropped to 7.9 percent. It has not played a major role in either as a driver of convergence or divergence (Bourguignon and Morrisson 2002).

262

Africa, the region with lowest per capita GDP, has been an underperformer relative to world average and other regions for most of the period from 1820-1998. The region’s share in world GDP has been around 3 to 3.5 percent after 1820. Africa saw no growth during the 1973-1998 period, counter-balancing some of the convergence impacts in global inequality of fast growth in China. The recent period after 1998 saw an increase in growth rates for the region driven by the commodity boom, helping reduce the divergence with other regions to a small extent.

263

Eastern Europe and former USSR contributed about 13 percent to world GDP in 1973 but the collapse of the USSR led to a sudden decline and negative growth in the region resulting in its share declining to close to 5.4 percent in 1998. Faster growth in 1998-2010 period has resulted in some catching up by the region with its share in world GDP increasing to 6.4 percent in 2008.

264

In addition to differential growth rates across regions of the world, divergence within countries and within regions also contributes to global inequality. The decline in inequality in Western European countries due to redistribution policies in the first half of 20th century acted as an equalizing force in this period. The communist revolution in Soviet Union, East European countries, and China in the same period also acted to decrease global inequality. Both of these factors reversed in the recent period as inequality within countries of Western Europe and former USSR has seen increases in the recent two decades (Bourguignon and Morrisson 2002). This is in line with the U-shaped pattern in inequality observed by van Zanden, et al. (2013) and also presented as so-called Kuznets wave by Milanovic (2016). Between 1990 and 2010 East Asia and Pacific saw substantial declines in inequality within the region (Jayadev, Lahoti and Reddy 2015). This was led by catching up of China with other East Asian tigers (Japan, Korea and Taiwan).

265

4.2 International inequality in education and health

266

Human wellbeing is a multi-dimensional phenomenon that is often correlated with incomes, with the correlation being far from perfect (Dreze and Sen, 1991). Several dimensions such as health, education, income have been combined to create different composite indices of wellbeing, with UNDP’s Human Development Index being among the most popular such index. Prados de la Escosura (2014) uses historical data on health (life expectancy), education (adult literacy and enrollment rates) and income (per capita income) to create a historical human development index (HHID) spanning 1870 to 2007. The index has increased six-fold over the period, but with differing movements over the years. From 1870-1913 saw steady but moderate progress which accelerated during 1913-1970 period followed by sustained deceleration till 1990 and an expansion after that. Analysis by Prados de la Escosura (2014) shows that the gap between OECD countries and the rest of the world was stable until 1913 and the rest of the world caught up at a stronger pace until 1970. After 1970 the pace of catch up by rest of the world has declined. Different regions vary in the timing and pace of catch up with OECD countries in human development. Latin America converged until the 1980’s and continued converging later in 2000’s, Africa caught up until 1970’s but since then Sub-Saharan Africa has not caught up further. Asia – driven by China and India – has seen rapid catch up since 1990 and before witnessed a period of convergence until the 1970s, dominated by improvements in education and health in China. The Soviet Union saw substantial gains in human development between the 1920s and 1960s, leading to substantial catch up of Central and East European countries with the OECD. But since late 1960’s the gaps have increased again between OECD and Eastern European countries, and are now stagnating since the mid-2000s.

267

To take account of inequality within dimensions UNDP introduced the Inequality-adjusted Human Development Index (IHDI). IHDI adjusts the Human Development Index (HDI) for inequality in the distribution of each dimension across the population and is based on indices proposed by Foster, Lopez-Calva and Szekely (2005) which are based on the Atkinson family of indices. IHDI discounts each dimension’s average value according to its level of inequality. The difference between HDI and IHDI increases as inequality rises. Though not a perfect measure, IHDI can be useful to denote the loss to human development due to inequality. In 2015[26], IHDI was measured for 151 countries and globally indicated a loss of 22.8% in human development to inequality. The loss was largest for Sub-Saharan Africa (33.3%) and smallest for Europe and Central Asia (13%). 

268

4.2.1 Education Inequality

269

The Human Development Report (HDR) estimated inequality in the mean of years of schooling using Atkinson index (ε=1) in 2015 to be 26.8% globally, with South Asia being the most unequal region (41.5%). Using Math and Science scores of 8th graders in 49 mostly rich countries, Sahn and Younger (2007) estimate within and between country inequalities in learning achievements. They find that for this group of countries slightly more than half of all inequality is due to within country differences in achievements. Ferreira and Gignoux (2014) calculate inequality in educational achievement using 2006 round of PISA surveys in 57 countries. They show that standard inequality measures like the Gini and Theil are not appropriate to measure inequality in standardized test scores, but a simple variance measure is better suited. Just over a third of total inequality in education achievement in these countries is due to inequality of opportunity.

270

4.2.2 Health Inequality

271

Different health indicators (child height, adult height, life expectancy, infant or child mortality, morbidity etc.) have been used to measure health inequality with differing conclusions. Pradhan et. al (2003) use the height of pre-school age children (which is supposed to be comparable across the world) to analyze global health inequality. They find considerable variation in intra-household inequality across countries. Decomposing health inequality within and between countries the authors find that more than two-thirds of health inequality can be attributed to within-country variation, as opposed to income inequality where most of the contribution to global inequality comes from the between component.

272

Joerg et. al (2014) use data on adult human heights for 25 countries spread across regions of the world to study between country inequality. They find that coefficient of variation for heights between countries increased in the period 1870 to 1970, matching the pattern observed in dispersion of GDP per capita.

273

Studying life expectancy across world regions Joerg et. al (2014) find that though life expectancy increased by about 30 years in all regions of the world the trend shows large variations. At the global level, they find that life expectancy diverged in the late 19th and early 20th century, and then converged in the late 20th century due to rapid gains in non-OECD regions of the world. But in the last two decades up until about 2000, levels of life expectancy are diverging between countries in several regions of the world and between country inequality is rising (Goesling and Firebaugh, 2004); this was heavily affected by drastically rising AIDS mortality as well as high overall mortality in crisis-prone Sub-Saharan African countries.Recent years have seen large gains in life expectancy and fall in child mortality in those regions, thereby contributing to a return to convergence trends in life expectancy globally.

274

Table 3.1: World GDP per capita by region (1990 international $)

275
 

1820

1870

1913

1950

1973

1998

2010

Western Europe

1232

1974

3473

4594

11534

17921

20889

Eastern Europe

636

871

1527

2120

4985

5461

8678

Former USSR

689

943

1488

2834

6058

3893

7733

Western Offshoots

 1201

2431

5257

9288

16172

26146

29564

Latin America

665

698

1511

2554

4531

5795

6767

Asia

591

548

691

716

1717

3529

6307

Africa

418

444

585

852

1365

1368

2034

World

667

867

1510

2114

4104

5709

7814

Ratio of highest to lowest GDP per capita

3

5

9

13

12

19

15

276

Source: Data for 1820 to 1998 is from Maddison (2001) and for 2010 is from The Maddison-Project, http://www.ggdc.net/maddison/maddison-project/home.htm, 2013 version.

277

Table 3.2: Share of World GDP (percent of world total)

278

1820

1870

1913

1950

1973

1998

2008

Western Europe

23.6

33.6

33.5

26.3

25.7

20.6

17.1

Eastern Europe

3.3

4.1

4.5

3.5

3.4

2.0

2.0

Former USSR

5.4

7.6

8.6

9.6

9.4

3.4

4.4

Western Offshoots

1.9

10.2

21.7

30.6

25.3

25.1

21.5

Latin America

2.0

2.5

4.5

7.9

8.7

8.7

7.9

Asia

59.2

38.3

24.5

18.5

24.1

37.2

43.7

Africa

4.5

3.6

2.7

3.6

3.3

3.1

3.4

279

Source: Data for 1820 to 1998 is from Maddison (2001) and for 2008 is from The Maddison-Project, http://www.ggdc.net/.

280

Table 3.3: Rates of growth of World GDP per capita (annual average compound growth rates)

281

1820-1870

1870-1913

1913-1950

1950-1973

1973-1998

1998-2010

Western Europe

0.95

1.32

0.76

4.08

1.78

1.29

Eastern Europe

0.63

1.31

0.89

3.79

0.37

3.94

Former USSR

0.63

1.07

1.76

3.36

-1.75

5.89

Western Offshoots

1.42

1.81

1.55

2.44

1.94

1.03

Latin America

0.10

1.81

1.43

2.52

0.99

1.30

Asia

-0.15

0.54

0.10

3.88

2.92

4.96

Africa

0.12

0.64

1.02

2.07

0.01

3.36

World

0.53

1.30

0.91

2.93

1.33

2.65

282

Source: Data for 1820 to 1998 is from Maddison (2001) and for 2010 is from The Maddison-Project, http://www.ggdc.net/.

283

Table 3.4: Comparison of Global Inequality Estimates from Various Studies

284

1988

1993

1998

2003

2005

2008

Global Consumption and Income Database (Income)

69.6

69.6

69.2

70.6

68.4

66.6

Anand and Segal (2014) (with top incomes)

72.6

72.7

72.2

73.5

72.7

Anand and Segal (2014) (without top incomes, survey means)

70.5

70.7

69.8

71.1

70.1

Anand and Segal (2014) (without top incomes, Household Consumption from National Account )

73.9

72.1

71.1

70.6

69.8

Lakner and Milanovic (2013) (National Account Means + top heavy Pareto imputation)

76.3

76.1

77.2

78.1

75.9

Lakner and Milanovic (2013) (National Account Means)

71.5

70.5

70.6

70.7

67.6

Lakner and Milanovic (2013) (only survey means)

72.2

71.9

71.5

71.9

70.5

Milanovic (2012)

67.8

69.3

68.8

70.1

Milanovic (2005)

61.9

65.2

64.2

Milanovic (2002)

62.5

65.9

Bhalla (2002) (Income)

67

65

Bhalla (2002) (Consumption)

66

63

Bourguignon & Morrisson (2002)

66

Chotikapanich et. Al (1997)

65

Dikhanov & Ward (2002)

69

68

Dowrick & Akmal (2005) (GK)

64

Dowrick & Akmal (2005) (Afriat)

71

Sala-i-Martín (2006)

65

64

64

Bourguignon (2012)

71

69

66

Notes: Milanovic (2012): Table 4, p. 14: Gini from row 5 (2005 PPP, sep. rural-urban prices for China, India & Indonesia); Theil from row 3 (2005 PPP, sep. rural-urban prices for China only); 2002 figures for 2003 benchmark. Milanovic (2002): Table 16, p. 72: Using full sample; Table 19, p. 78 (decomposition): Only for common sample. Milanovic (2005): Table 9.4, p. 108: Using full sample; Table 9.5, p. 112 (decomposition): Only for common sample. Bourguignon (2012): Figure 1, only approximate, because read-off from figure; 1988 refers to 1989, 1998 refers to 1997, 2008 refers to 2006. Otherwise: Anand and Segal (2008), Table 1: Survey estimates allocated to benchmark according to rules with micro data: 1988: Bhalla (2002), CVR (1997), and DW (2002) all refer to 1990; 1993: BM (2002) refer to 1992; 1998: Bhalla (2002) refers to 2000, and DW (2002) refers to 1999. GCIP's 1988 estimate refers to 1990.

285

Source: Based on Lakner and Milanovic (2013) and additions done by Klasen, Scholl, Lahoti, Ochmann and Vollmer (2016)

286

5. Trends of within-country inequality 

287

5.1 Empirical Trends by Outcomes 

288

5.1.1 Policy Instruments

289

As discussed above (3.2), income inequality data often suffers from problems of incomparability across countries and across time, from infrequent surveys, and from missing data for some countries.  Although the situation has improved in recent years, data problems remain severe, and different data sets provide different findings on levels and trends in inequality across countries and regions. Of particular note is that trends can often be sensitive to the particular start and end-point used, which is often dictated by data availability, which may differ across countries. Below we provide data from studies that have carefully combined data from different regions. To give a sense of whether the reported trends are robust or not, we also provide estimates in the appendix to this section based on the Global Consumption and Income Project (GCIP), which is trying to be particularly transparent about data sources and adjusts for differences between income and consumption based surveys.  Of course it has its own problems but the idea is simply to provide some robustness check (see further discussion in 3.2).  

290

For long, within-country income inequality was considered a stable feature of each economy, as its economic structure was thought to evolve slowly (Deininger and Squire, 1996). Yet, already in the late 1970s there were signs of rising inequality in many advanced economies (Piketty, 2014). In the 1980s the Gini coefficients started moving upward also in the developing and transitional countries, and this tendency continued in the 1990s. During this period, these economies were affected by the debt crisis of the 1980s, the liberalization of domestic and external transactions, technological shocks that pushed upward the skill premium, rising migration and – for the former communist countries after 1989 – radical policy changes that often triggered acute transformational recessions, accompanied by sharply rising inequality from previously low levels (Gruen and Klasen, 2001).[27] As a result, over 1980-2000, 69 percent of the 109 countries with available data (Table 3.5, top panel) recorded a systematic increase in income inequality, and only 23 percent a decline. The greatest number of increases was recorded in the economies in transition, Latin America, the OECD and South East Asia. As shown in the appendix, the GCIP data largely concur with this observation on a trend towards rising inequality, although it finds it somewhat less pronounced in Latin America and South-East Asia.[28]

291

Table 3.5. Number of countries for each region where the Gini coefficient rose, stagnated or declined over 1980-2000 (top panel) and 2000-2010 (bottom panel)   

292

OECD

European

Transition

Economies

Asian Transition

Economies

Latin

America

 MENA

South

East

Asia

South

Asia

Sub-

Saharan

Africa

World

1980s (or earlier available year) and 1990s

Specific period for

each region/3

1980-2001

1990-

1998

1980-

2000

1980-

2002

1980-

2000

1980-1995

1980-

2000

1980-1995

Rising inequality

14            24

2

14

2

5

3

9

73    (69%)

 No change

1               0

0

1

1

3

0

0

2

8       (8%)

Falling inequality

6

0

0

3

3

2

2

8

24    (23%)

Total

21             24

3

18

8

7

5

9

105 (100%)

2000-2010 (or similar period)

Specific period for

each region /3

2000-

2010

1998-

2010

 2000 –

2009

2002-2010

2000-2007

1995-2009

2000-

2010 

1995-2007

    

Rising inequality

9

3

2

2

4

3

4

7

44    (41%)

No change

4

5

1

1

0

0

1

1

13 (   12%)

Falling inequality

8

6

0

15

4

4

0

13

50     (47%)

Total

21

4

3

18

8

7

5

21

 107   (100%)

                             
293

Source: Cornia and Martorano (2012). Note: /1 Countries were assigned to the rising, no change or falling inequality groups on the basis of an analysis of time trends and of the difference between the initial and final Gini coefficients for each of the two sub-periods considered. 

294

Since the early 2000s (a period characterized by the rise of the Asian economies, a stable business cycle in the USA till 2008, a rapid increase in the global demand and prices of primary commodities, and an increase in global financial flows and remittances), there was a slowdown in the frequency and intensity of inequality rises. In fact, as shown in the bottom panel of Table 3.5, over 2000-2010 there was a divergence in Gini trends. Income inequality fell in practically all of Latin America. In turn, consumption inequality fell in 13 of the 21 Sub-Saharan African countries with consistent data. Income inequality fell also in South Korea, Thailand, the Philippines and Malaysia in the aftermath of the 1997 bath crisis (Cornia and Martorano 2012). In contrast, inequality continued its upward trend – if at a slower pace than during the prior two decades - in most OECD countries and several European and Asian transition economies, while it accelerated in most of South Asia and China. Overall, during this period, 47 percent of the countries recorded an inequality decline and 41 percent an increase. The GCIP data (see appendix) largely confirm these trends, although they find more evidence of (slightly) falling inequality in South Asia (after previous substantial increases).  Over 2012-14 the regional Gini coefficients became more unstable and it is not clear yet what has been the distributive impact of the financial crisis of 2008 and of the sluggish growth of the 2010s.

295

How about aggregate inequality trends by region and income level? In this regard, Table 3.6 presents the unweighted average Gini coefficient by region and income category for the period 2000-2010, a period during which – as noted above – regional inequality trends diverged and inequality rose in most cases more slowly than during the two prior decades.[29] There are several observations to be made in this regard.  In 2010, average inequality across the world can be classified as moderate:  the average Gini is just below 0.4, the median of the difference between the regions with the lowest and the highest Gini coefficient. The level of inequality in the Advanced Economies, Eastern Europe and Central Asia and South Asia is below the world average; it is above the world average in Sub-Saharan Africa and Latin America and the Caribbean. East Asia’s inequality is roughly the same as the world’s average. Latin America and the Caribbean is the region with the highest extent of inequality: its average Gini is twelve points above the world’s average. As shown in the appendix, using GCIP shows exactly the same trends but finds higher levels of inequality in all developing regions except Latin America and in fact suggests that Sub-Saharan Africa is the region with the highest inequality. This is due to the fact that consumption surveys (prevalent in all developing regions except Latin America) tend to show lower inequality than income surveys and GCIP adjusts for that. Using the World Bank’s classification of countries by income level, low income countries’ inequality is well below the world’s average.  The most unequal group is comprised of upper middle income countries, a reflection of the influence of unequal Latin America since a significant number of middle income countries are from that region.  Trends in inequality confirm much of the discussion above: slight further increases in inequality in advanced economies and East Asia, large reductions in Latin America, small reductions in South Asia, and little changes in other regions, leading to the overall impression of very slightly falling average within-in country inequality in the world between 2000-2010.[30]

296

Table 3.6: Average Inequality by Region and Income Level 2000-2010 (5 Year Averages)

297

Gini Coefficienta

Regionb

2000

2005

2010

World…………………………. 

0.390

0.385

0.380

Advanced Economies………… 

0.298

0.302

0.304

East Asia and the Pacific……… 

0.380

0.391

0.389

Eastern Europe and Central Asia

0.331

0.329

0.333

Latin America and the Caribbean

0.551

0.532

0.502

Middle East and North Africa

Not Enough Data

South Asia……………………  

0.354

0.351

0.328

Sub-Saharan Africa…………..    

0.445

0.434

0.440

Income Categoryc

Low Income Countries………  

0.316

0.32

0.323

   Lower Middle Income Countries…   

0.421

0.412

0.399

   Upper Middle Income Countries…   

0.442

0.436

0.428

Total Middle Income Countries 

0.431

0.423

0.413

High Income Countries……….

0.397

0.386

0.386

298

Source: Lustig (forthcoming), Table 1.

299

Notes: the above data have been computed  on the basis of the OECD Income Distribution Database http://www.oecd.org/social/income-distribution-database.htmPovcalNet http://iresearch.worldbank.org/PovcalNet/index.htm?0,0., CEDLAS Accessed July 22, 2013. http://sedlac.econo.unlp.edu.ar/eng/statistics-detalle.php?idE=35. The income concepts used for the computation of the Gini coefficients vary from region to region. Precise details on the data sources and on how the above data were calculated can be found in Lustig (forthcoming)

300

As for the overall inequality trends over 2000-2010, there is evidence of a slight decline in world inequality[31] and inequality convergence.[32]In particular, Latin America and the Caribbean—which was and still is the most unequal region—has experienced as noted a significant decline.Inequality in low-income countries has experienced a slight increase, while inequality in middle and high income countries has fallen a bit. Convergence is graphically apparent in Figure 3.4 below that also shows that declining inequality has been more frequent in the 2000s.Of the 78 countries included in the graph, 45 experienced a decline, 30 an increase, and three no change.Data from CGIP (see appendix) largely confirm the trend of slight inequality convergence.

301

Figure 3.4: Gini Coefficient: Level and Change by Country, 2000-2010

302

303

Source: Lustig (forthcoming), Figure 3.4. Note: for data sources and income concepts used see the note to Table 1.  Note that green lines and dots refer to observations where inequality increased between 1998-2002 and 2008-2012, while red lines and dots refer to observations where inequality fell. 

304

5.1.2 Trends in Education and Health Inequality

305

5.1.3 Perceptions of inequality and inequality change

306

5.1.4 Inequality in access to environmental resources and distribution of damages

307

Environmental resources include both classic public goods that have a low degree of excludability and rivalry (e.g. air, open seas) and private goods (e.g. agricultural land). Access to these resources is typically unequally distributed within countries and literature has explored differences across various axes (e.g. poverty status, race, class). A related issue is one of environmental hazards (e.g. pollution) and unequal exposure of different groups to these hazards.

308

By now, a considerable literature has accumulated from various contexts that argues that both environmental “goods” and “bads” are unequally distributed across the population. For example, a report of the World Health Organization (WHO 2012) focusing on Europe finds that perceived access to green space varies across different socioeconomic groups – based upon income, education etc. In the United States, considerable evidence has been presented to argue for the existence of “environmental racism” – the idea that the poor and particularly African Americans share a disproportionate burden of polluted air, polluted water and toxic waste (e.g. Bullard (1993), Eligon 2016).

309

One of the important environmental resources is land. In urban areas, both in developed and developing countries, ownership of land and inequality in its distribution is usually construed and understood with respect to its residential use, i.e. housing. Almost by definition, the poor and disadvantaged groups live in inferior quality housing. For a description of the relevant issues and theories, particularly for developed countries, see Bruckner (2011). For developing countries, one particular phenomenon that has received some attention in recent times is the growth of slums, which are characterized by poor housing and sanitary conditions (see e.g. Davis 1999).  

310

Compared to land in urban areas, a relatively larger literature has focused on agricultural land, its distribution and the functional implications of land inequality. Agricultural land is typically unequally distributed, although the extent of inequality varies across the regions of the world ( Griffin et al., 2002). In the developing world, countries in Latin America and South-Eastern Africas have typically been characterized by higher levels of land inequality. In certain contexts, historically disadvantaged groups (e.g. the Scheduled Castes in India) are distinguished by lower ownership of land. Inequality of land can be linked to the existence of various agricultural institutions (sharecropping, fixed rent etc., see e.g. Ray (1999)). Land inequality has also been linked to lower growth (Alesina and Rodrik 1994) and conflict, even revolutionary conflict. A negative relationship has also been documented between the size of farms and land productivity. An argument can therefore be made for land redistribution on grounds of both equity and efficiency. However, ambitious land redistribution schemes (“land reforms”) have generally met with limited success (for various reasons, e.g. landlords acting as an interest group) except in East Asia (see Ray (1999)). Despite this, Griffin et al. (2002) has argued land redistribution remained a development priority and, based upon successful cases of land reforms, made an argument for land confiscation and against market-based schemes. However, with growing urbanization, the spread of income transfers schemes and public work programs that alleviate extreme rural poverty and the undiminished power of the agrarian elites, during the last two decade even partial land reform programmes planned by progressive centre-left governments (as Bolivia, Brazil, and Paraguay in Latin America) were abandoned (Cornia 2014). Meanwhile, since the early 2000s access to the land by poor farmers has been increasingly threatened by over one thousand foreign land deals in which customary land users were often evicted (Nolte et al 2016). At the same time, improvements in the certainty of use of the land were gradually achieved in Sub-Saharan Africa and possibly elsewhere since the mid-1990s by means tenancy reforms that allowed for the formal or informal registration of customary land rights of small farmers (Cotula et al., 2004).

311

5.1.5 Levels and trends of inequality of opportunities

312

Theoretical Framework

313

Starting in the second half of the twentieth century the concept of inequality of opportunity (IOp) has gathered considerable attention from scholars and policymakers alike (for recent overview articles see Roemer and Trannoy 2015; Ramos and Van de Gaer 2016). Roemer (1993, 1998) proposed a concept of IOp which takes the source of the unequal distributions of certain outcomes (e.g. health, income or well-being) into account. According to this notion, individual outcomes can be influenced by circumstances and efforts. Circumstances are defined as all factors which are beyond individual control - such as parental education, gender, or ethnic origin - and for which individuals should not be held responsible. Effort, however, describes all actions and choices which are within individual control - such as schooling choices or career decisions - and for which individuals should be held (partially) responsible. Though, some scholars even argue that before a certain age of consent is reached, all individual behaviors are due to circumstances; meaning that children below that age of consent should not be held responsible for their achievements (e.g. Roemer and Trannoy 2015; Hufe et. al 2015). Most importantly, taking account of its source inequality can be decomposed into morally acceptable and unacceptable inequality. Hence, income differences due to efforts are considered as acceptable, while differences due to circumstances are not. Equality of opportunity (EOp) is given when the chances faced by individuals to achieve an outcome in question are influenced only by individual effort, irrespective of individual circumstances. Thus, an EOp policy aims at equalizing opportunities by compensating individuals for their disadvantaged circumstances while leaving differences due to effort unchanged; guaranteeing that those who put in equal degrees of effort achieve equal outcomes. The concept of EOp, therefore, departs from the traditional notion of equality of outcomes (EO), where economic outcomes (e.g. income, consumption or well-being) are equally distributed across the population.

314

Overview of Empirical Results

315

Empirical studies have measured the extent of IOp for various outcomes (e.g. income, wages, health) and in a variety of countries (see Table 3.7 for an overview). A study by Aaberge et al. (2011) for example, estimates both IOp for annual and permanent incomes in Norway. They find that IOp accounts for 23% to 26% of income inequality in annual income and about 28% in permanent income. For Italy, estimates show that IOp accounts for about 20% of overall income inequality when using only a rather broad set of circumstances (Checchi and Peragine 2010). In other European countries, this value ranges from 5% in Slovenia to 24% in Luxemburg (Checchi et al. 2010). In a recent follow-up study these authors furthermore show that these estimates are rather stable between 2005 and 2011. For the US Pistolesi (2009) finds that between 1968 and 2001 between 20% and 43% of earnings inequality is due to IOp, however he also finds that overall IOp is on a decreasing trend. Results for less developed countries show that in Brazil and Guatemala roughly 33% of income inequality is explained by circumstances (Ferreira and Gignoux 2011), while this is the case for 21% in Madagascar and 13% in Ghana (Cogneau and Mesple-Somps 2008). However, the empirical assessment of IOp is complicated by the fact that datasets rarely allow for all individual circumstances to be accounted for. Since the effect of omitted circumstances appears as the effect of effort, these measurements of IOp are downward biased. Using demographic surveys from 27 countries Balcázar (2015) suggests that between 73% and 93% of IOp are left unexplained by the standard lower bound estimator. This fact has led critics to question the relevance of IOp for policymakers (Kanbur and Wagstaff 2014) altogether. In particular, it is argued that these studies could be used to play down the problematic nature of inequality, as most of it appears ethically acceptable from an equal-opportunity perspective. One strategy to address this downward bias is the use of detailed datasets that contain broader information on circumstance variables. Björklund et al. (2012) use intelligence tests from military records to obtain a better measure for individual ability, which indeed evolves as one of the strongest contributors to IOp in Sweden. Hufe and Peichl (2016) use data with genetic information to move the lower-bound estimator in the direction of its true value. Another strategy is the use of panel data to estimate individual fixed effects with respect to the outcome of interest. As such scholars measure the effect of all time-invariant circumstances. On the one hand, this estimator only is an upper bound to the extent that circumstances do not change over time. On the other hand, this measurement of IOp might be biased upward if some components of individual effort are also constant in time. Such studies show that IOp accounts for 33% to 36% of income inequality in the USA and for 47% to 62% in Germany (Niehues and Peichl 2014). Both strategies are rather data intensive. While the first strategy requires detailed information on circumstances and large sample sizes to maintain an appropriate number of degrees of freedom, the second strategy affords a certain number of repeated individual observations to yield a reasonably stable estimate of the person fixed effect. In summary, the empirical literature on IOp provides evidence that circumstances matter strongly for the determination of desirable life outcomes. Fruitful avenues for further research include the narrowing of the range between upper and lower bound estimators, increasing the comparability of IOp estimates across countries and time (e.g. Brunori et al. 2016) and the reconciliation of IOp with other normative values (e.g. Foguel and Veloso 2015)

316

Table 3.7: (Overview of studies measuring IOp)

317

Author

Outcome

Data

Circumstances

Country

IEO

Aaberge et al. (2011)

permanent income

Statistics Norway, 1967-2006

birth cohort, parental educational, urbanity of birth place, family size

Norway

28%

period-specific income

Statistics Norway, 1967-2006

birth cohort, parental educational, urbanity of birth place, family size

Norway

23-26%

Björklund et al. (2012)

permanent income

Statistics Sweden, (1955-1967)

parental income, parental education, family type, number of siblings, IQ, BMI

Sweden

13 41%

Brunori et al. (2016)

hh consumption (per capita)

EIM (2004)

birthplace, parental education, parental occupation

Comoros

30-42%

ECM, 2010

birthplace, parental education, parental occupation

Congo D.R:

47%

GLSS, 2013

birthplace, parental education

Ghana

35-56%

EIBEP, 2003

birthplace, parental education, parental occupation

Guinea

35-37%

EPM, 2005

birthplace, parental education, ethnicity

Madagascar

28-56%

IHS3, 2010

birthplace, parental education, mother tongue

Malawi

43-56%

ECVM, 2011-12

birthplace, ethnicity

Niger

33-40%

GHS, 2010-11

birthplace, parental education, parental occupation, ethnicity

Nigeria

38-42%

GHS, 2012-13

parental education, parental occupation

Nigeria

37-43%

EICV, 2000

birthplace, parental education, parental occupation

Rwanda

26-42%

NPS, 2009-10

birthplace, parental education, parental occupation

Tanzania

43-49%

NPS, 2010-11

birthplace, parental education, parental occupation

Tanzania

41-46%

UNPS, 2009-10

birthplace, ethnicity

Uganda

39-48%

UNPS, 2010-11

birthplace, ethnicity

Uganda

40-53%

Checchi and Peragine (2010)

gross earnings

SHIW,1993, 1995, 1998 and 2000

parental education, sex, region

Italy

20%

Pistolesi (2009)

annual earnings

PSID, 1968-2001

age, race, parental education, region, father’s occupation

USA

20 - 43%

Ferreira and Gignoux (2011)

hh income (per capita)

ECV, 2003

sex, race, parental education, region

Colombia

23%

Checchi and Peragine (2010)

gross earnings

ENAHO, 2001

sex, race, parental education, region

Peru

28%

Pistolesi (2009)

annual earnings

ENCOVI, 2000

Same as above plus father’s occupation

Guatemala

33%

Ferreira and Gignoux (2011)

hh income (per capita)

PNAD, 1996

Same as above plus father’s occupation

Brazil

32%

ECV, 2006

Same as above plus father’s occupation

Ecuador

26%

ENV, 2003

Same as above plus father’s occupation

Panama

30%

Niehues and Peichl (2014)

annual income

SOEP, 1984-2009

sex, country national, father’s education and occupation, urbanity of birth place, height, birth year, born in East or West Germany

Germany

47 - 62%

PSID, 1981-2007

sex, country national, father’s education and occupation, urbanity of birth place, height, birth year, born in the South of the US, race

USA

33 - 36%

Peichl and Ungerer (2014)

total net income

SOEP, 1992-2012

sex, country national, father’s education and occupation, urbanity of birth place, height, birth year, born in East or West Germany

Germany

35 (1991)-

24%

(2011)

Cogneau and Mesple-Somps (2008)

hh consumption (per capita)

EPAMCI, 1985-88

GLSS, 1998

EICVM, 1994

EPAM, 1993

NHIS, 1992

father’s education and occupation, region (Colombia, Peru without father’s occupation)

Ivory Coast

Ghana

Guinea

Madagascar

Uganda

13%

11%

13%

21%

9%

Checchi et al. (2010)

net individual earnings

EU-Silc 2005

parental education and occupation, sex, nationality, region

Austria

Belgium

Cyprus

Czech Rep.

Denmark

Estonia

Finland

France

Poland

Greece

Hungary

Ireland

Portugal

Latvia

Lithuania

LuxemburgNetherlands

Slovakia

Slovenia

Spain

Sweden

UK

22%

17%

30%

11%

14%

11%

10%

13%

9%

17%

10%

22%

12%

12%

15%

24%

19%

14%

5%

19%

11%

21%

318

5.1.6 Levels and trends in mobility and intergenerational inequality

319

5.1.7 Inequality in political participation and power 

320

Broadly put, there are two distinct kinds of societies that exist today. In the first, traditionally construed as “dictatorships”, a small group controls political power and the institutions of the state, with limited accountability to the people. In the second, usually considered as “democracies”, there is more accountability to the people and stronger checks and balances (e.g. through elections) on the actions and power of the state and those who control it. However, in both these kinds of societies, disproportionate power could be wielded by certain individuals and groups. A vast and burgeoning literature, from different perspectives (e.g. New Institutional, Marxist etc.) and spreading across various social sciences, exists on the state and the nature of democracies and dictatorships (e.g. Moore (1993), Poulantzas (1974), Therborn (2006), Acemoglu and Robinson (2009), and the references therein). Within economics, there is a literature that examines the conceptualization of differential political participation and access to power across groups, and the implications of this. In the interests of space, in this section, we focus on this literature.

321

One useful distinction that has been made in the above literature is between inequality among individuals and inequality among groups, usually identity groups (e.g. ethnic groups, races etc.) (Stewart 2001). Studies on the former (“vertical”) inequality have largely examined consumption and income and virtually ignored political participation. This is notwithstanding the fact that some scholars have argued that certain individuals (the rich) have more power and are also more likely to vote, so that it is not the median voter who is decisive, and this has implications for redistribution (e.g. Benabou (2000)). On the contrary, studies on the latter (“horizontal”) inequality have explicitly focused on politics and power. Stewart (2001) argues that inequality among groups is multidimensional and the political dimension is crucial, along with economic and social dimensions. Differences in political participation across groups could exist on several fronts, e.g. government ministers, parliament, civil service etc. Several studies have documented inequalities among groups in various contexts, e.g. Østby 2008, Wimmer et al. 2009, Brown 2010, Mancini 2008, Motiram and Sarma 2014. This literature has argued that such inequalities are unfair and also lead to severe instability and conflict.

322

Gender is an important dimension on which differences in political participation and power exist, and has therefore been examined. In the spirit of the above discussion, if one examines women in parliament, this is only about 23% of parliamentarians in 2016, although there is an increase of 12 percentage points from 1993 (UN Women 2016). Wide variations exist across geographical regions, with the Nordic countries at the top and Pacific countries at the bottom (UN Women 2016). How do we address the issue of differential political participation and power? The literature on horizontal inequalities argues that different kinds of policies can be put in place to address them: affirmative action, anti-discrimination, and nation-building (Stewart 2001). To explicitly address the issue of power and political participation, one option is to “reserve” positions for disadvantaged groups. This policy has been experimented with in certain contexts, and evidence suggests that it has produced desirable outcomes, e.g. Chattopadhyay and Duflo (2004) and UN Women (2016) for women; Pande (2003) for Scheduled Castes and Tribes in India.

323

5.2 Group-based inequalities in outcomes: Gender, race, ethnicity, spatial

324

6. Accounting for within-country inequality trends

325

6.1 Trends and drivers of inequality in OECD countries

326

6.1.1 Inequality Trends

327

During the period between the mid-1980s and 2013, the majority of the OECD countries experienced moderate increases in the level of disposable household income inequality. In the mid-1980s the average Gini coefficient for OECD countries stood at 0.29. By 2013, this average had increased by 10% to 0.32, rising in 17 of the 22 OECD countries for which long-time series are available (OECD 2015). A more significant increase of inequality was averted by government spending on social policies – if taxes and benefits were less targeted to the poor, the growth in inequality would have been much more rapid. It is noteworthy that there is strong heterogeneity within the set of OECD countries that is masked when focusing on average measures. While the majority of countries experienced increases in inequality, the income dispersion in Belgium, Netherlands, France and Greece remained stable. Turkey even experienced a decrease in the Gini from 0.434 to 0.412 (Förster, 2016, see Figure 3.5 below).

328

Figure 3.5: Gini coefficients of income inequality, mid-1980s and 2013

329

330

     Source: Förster (2016)

331

The rise in income dispersion between the mid-1980s and the mid-1990s was stronger compared to later decades. A dominant pattern in this time period was a marked increase of inequality in market income. On average, the income distribution widened by 0.018 points (6%) and by slightly less (0.014 points, 5%) when excluding Mexico and Turkey (OECD 2008).

332

During the mid-1990s to the mid-2000s, the average Gini within the OECD rose by another 0.002 points. However, the average conceals marked differences within the OECD group. The income distribution narrowed in 10 countries, with large declines in Mexico and Turkey. Without the equalization of incomes in Mexico and Turkey the mean increase of the Gini in the OECD would have been 0.005 points higher in this period (OECD 2008).  In contrast, several other countries, among others the US, experienced strong increases in income inequality. While the widening of income inequality observed in the first decade mainly reflected greater inequality in the distribution of market income that was partly offset by tax-benefit systems, the growth of market-income inequality decelerated from the mid-1990s up to around 2000.

333

Since 2000 income inequality increased strongly in Canada, Germany, Norway and the US and, to a lesser degree, in Italy and Finland. It fell in the UK, Mexico, Greece, Australia and, to a smaller extent, in Sweden and the Netherlands. In the economic recovery following the great recession, income inequality before taxes and benefits continued to rise, while the cushioning effect of taxes and benefits has become weaker. As a result we observe a continued upwards trend in disposable income inequality in the most recent years (OECD 2015).

334

There is a very large literature that has analyzed drivers of inequality change in OECD countries.  We will provide a brief summary of the drivers that have been analyzed in greatest detail.

335

Wage inequality

336

Gross labor earnings make up the largest share of total household incomes and are an important driver of income inequality. Between the mid-1980s and today, wage disparities in the OECD among full-time workers have increased by between 20 and 25% (OECD 2011). Common explanations for the increasing polarization of wages are changes in the supply and demand for skills and tasks as well as changing labor market institutions. For the case of the US, Autor et al. (2008) conclude that the lower tail of the wage distributions was affected by declines in the real minimum wage. The development in the upper tail, however, was best explained by differential returns to skill as exemplified by increases in the college-wage premium. An underlying force of this tendency is the skill bias inherent in technical change, which has led to relative increases in the demand for high skilled workers (Acemoglu and Autor 2011). These considerations highlight that the drivers of increasing wage inequality are multifaceted and particular to different segments of the wage distribution. Atkinson (2008) confirms this conclusion for a subset of OECD countries in recent decades. He finds substantial increases in upper tail inequality, whereas he observes a more heterogeneous cross-country pattern at the lower end of the wage distribution.

337

Changes in demography and living arrangements

338

Changes in demographic structures, household sizes and household compositions are important co-determinants of the observed inequality patterns in OECD countries. While demographic factors only account for a part of the observed change in the income distribution in most countries, the impact of changes in living arrangements is much more important. All OECD countries have experienced a gradual movement away from the typical family structure that was most prevalent in the past. The increase in the share of single-parent households and of people living solo translated into a decrease in average household sizes with the corresponding loss of economies of scale and the need for a higher monetary income to assure the same level of well-being. For instance, Peichl et al. (2012) quantify how the trend towards smaller households has influenced the change in the income distribution in Germany. The results show that the income gap would have increased regardless of the demographic trend, but on a yet lower level. Furthermore, total disposable household income depends on the characteristics of individuals forming households. Within the OECD the increasing female labor force participation has exerted a moderating influence on the observed upwards trend in inequality (OECD 2015). This is consistent with findings from Cancian and Reed (1998) who study the role of female earnings on inequality. To the contrary, the correlation between male and female earnings within households, commonly termed “assortative matching”, has increased in recent decades. This tendency magnifies existing inequalities across households. For example, Burtless (2009) and Schwartz (2010) find the increasing correlation between husbands’ and wives’ earnings to be among the main drivers of increasing inequality.

339

Intergenerational mobility

340

Countries with lower intergenerational mobility tend to exhibit wider income inequalities. Following the work of Corak (2013), this relation has gained traction as the “Great Gatsby Curve”. In general, the countries with the most equal distribution of income at a given point in time feature the highest income mobility across generations. Those countries that are characterised by high intergenerational earnings mobility are the Nordic countries, Australia, and Canada, while the opposite applies to Italy, the US and the UK. Education systems are an important determinant for the transmission of disadvantages from generation to generation. In addition to education policies, universal health care provision, as well as family policies interact with existing inequalities in the parent generation to generate inequality patterns at the level of children (Corak et al. 2011). Taking a rather normative perspective, the literature suggests that a large portion of the income differences transmitted from one generation to the next relates to factors that are largely beyond the control of the child or of the parents (OECD, 2008). This of course raises ethical concern from an equal opportunity perspective (Roemer and Trannoy 2015).

341

Tax-Benefit Systems and Public Policies

342

Tax-benefit systems are an important mean to cushion the effects from increasing market inequalities. As shown in Immervoll and Richardson (2011), the redistributive effect of tax-benefit systems, measured as the difference in inequality between market and disposable income, has increased in OECD countries within the last decades. However, this rise was outpaced by according increases in market inequality. Evidently, tax-benefit systems comprise a multitude of instruments with differential distributional implications. For example, Fuest et al. (2009) show for a set of European countries that personal income taxes and social insurance contributions have a strong redistributive impact, whereas social benefits are less effective means for redistribution as they are largely untied from the market income of the recipient. As a consequence of the economic crisis, many OECD countries have been forced to cut-back redistributive programs for the purpose of fiscal consolidation (OECD 2015). Therefore, we can expect dispersions in market income to prevail to a stronger extent at the level of disposable income in these countries in the future.

343

Recent cross-country work by the OECD has tried to provide a comprehensive overview of the potential causal drivers of inequality change in OECD countries.  This work is summarized in Figure 3.6 and tested.  It confirms the role of skilled biased technological change, rising unemployment, eroding minimum wages, falling unionization, assortative matching, and reduction in the redistribution by the state as significant drivers.  But there remains non consensus on the respective size of these influences, in addition to inconclusive results in many areas that have been examined (see Förster, 2016a).  

344

Figure 3.6: Types of drivers of inequality change in OECD countries

345

346

Source: Förster (2016a)

347

The role of top incomes

348

Most analyses of inequality in OECD countries are based on standard household surveys which, as discussed above, systematically underreport top incomes.  Using tax records and other information, recent work has documented levels and evolution of top incomes (e.g. Atkinson and Picketty, 2007; Picketty, 2013).  This work has shown that the share of top incomes in total incomes has been rising substantially in most OECD countries since the 1970s, thereby leading to a further widening of inequality not captured in standard surveys.  The rising market income share of top income earners is mostly related to disproportionately rising compensation for top managers as well as top employees in finance, as well as growing wealth inequality and rapidly rising returns from wealth among top income earners.  Inherited wealth plays an important role in the transmission of top incomes. Falling taxation on wealth and very high incomes further exacerbated this trend towards rising income shares of top earners (Picketty, 2013).      

349

6.1.2 Changing inequality trends and drivers in three Asian sub-regions

350

South Asia and Indonesia: rising income inequality amid rapid economic growth

351

During the last two decades, the region recorded an unprecedented growth acceleration that contributed to a rapid reduction in the incidence of poverty. Yet, in South Asia, income inequality rose in most cases (Table 3.8). 

352

Table 3.8. Trend in the Gini coefficient of selected South Asian and South-East Asian countries 

353

Country (and years of reference )

First year

Second year

Bangladesh (1991-2010)

27.6

32.1

India (1993-2010)

32.5

37.0

Nepal (1995-2010)

35.2

32.8

Pakistan (1990 2011)

33.2

30.6

Sri Lanka (1990- 2006)

32.5

40.3

China P. R. (1990-2008)*

32.4

43.4

Cambodia (1994-2008)

38.3

37.9

Indonesia (1990-2011)

29.2

38.9

Malaysia (1992-2009)

47.7

46.2

Philippines (1991-2009)**

43.8

43.0

South Korea (1998 -2011)*

37.5

31.1

Thailand (1990-2009)

45.3

40.0

Vietnam (1992-2008)

35.7

35.0

354

Source: excerpted from Kanbur et al. (2014), Table 2.2 Notes: /* data from WIID; ** data from Li (2015)

355

In India, the Gini rose over 1993-4/2008-9 from 25.8 to 28.3 in the rural sector, and from 31.8 to 38.2 in the urban sector, the gap between urban and rural areas widened, while interstate inequality nearly doubled (Himansu and Lanjouw, 2015). As a result, the overall Gini rose by 4.5 points (Table 3.8). A key factor behind this surge was the growing wage gap between the organized and unorganized sectors, as well as between urban and rural areas (Datt and Ravallion, 2002, 2009; Ghosh 2014). At the same time, the remunerations of managers and capital owners grew rapidly especially in ‘rent sectors’ dependent on government licenses (mining, metals, construction, land, real estate, telecoms), or in knowledge-intensive sectors such IT and pharmaceuticals (Gosh, 2014). The wages of unskilled workers belonging to minority groups were further penalized by social norms that have segmented for centuries the labor market and discriminated workers belonging to scheduled castes and tribes and religious minorities who, because of low education and ‘pure segregation’, can offer their work only in poorly paid sectors (Ghosh 2015).

356

As for the other South Asian countries and Indonesia, Kanbur et al. (2014) argue that inequality rose because of the Skill Biased Technical Change recorded in the 1990s and 2000s. The importation of capital- and skill-intensive technologies raised the demand for workers with secondary and higher education, and so pushed upward the skill premium, due also to the lagging supply of such workers. Decompositions presented in Kanbur et al (2014) suggest that growing educational inequality explained up to 45 per cent of the increase in income inequality. Other factors that contributed to the rise of earnings inequality include weak labor institutions (minimum wages and collective bargaining), a commodity boom which favored only parts of the country, large and highly regressive fuel subsidies, and limited public expenditure on health and education. In addition, as the new imported technologies are capital intensive, there was an increase in the ‘capital share’. Its rise was due also to the fall of the bargaining power of organized labor and to job informalization (Kanbur et al. 2014; Yusuf et al. 2015; Miranta et al. 2013).

357

A second explanation of rising inequality in South Asia focuses on the increase in migration. Such literature emphasizes the rapid growth of the effective world labor supply (IMF 2007), global integration of labor markets, increased offshoring of production, and growing migration of semi-skilled workers that broaden the labor pool in countries of destination, but reduce it in those of origin. This has depressed the unskilled wage rate in both countries of origin and destination.

358

The unfettered adoption of market-oriented reforms was also a source of inequality. Trade liberalization was un-equalizing in most cases, as it increased the demand for skilled workers and for the reallocation of labor across regions and sectors. Such reallocation was however hampered by low labor mobility (Koujanou-Goldberg and Pavcnik 2007). Similar effects were generated by FDI allocated to capital-and skill-intensive industries (Te Velde and Morrissey 2002), consisting in Mergers and Acquisitions, or replacing the output of labor-intensive local firms. An even stronger impact was caused by the capital account opening (Prasad et al. 2003). 

359

Kanbur et al (2014) argue also that spatial and regional inequality increased, as new firms flocked to areas already specialized in the production of manufactured goods and modern services. This permitted to benefit from economies of scope and agglomeration, but entailed leaving behind the remote regions, causing in this way an increase in spatial and total inequality. Finally, Claus et al (2014) show that fiscal policy hardly reduced market income inequality. In the region, the corporate income tax, social security payments, sale taxes, excises and custom duties were all – on average – regressive. Only the personal income tax was progressive, though in many countries high exemption thresholds and generous deductions reduced its redistributive potential. In turn, public expenditure on social protection and housing was regressive (while it is progressive in other regions). Only that on health and education was progressive. Thus, in the absence of explicit social policies targeting the poor, the inequality of market income was hardly affected by the tax-and-transfer system.

360

The moderate inequality decline recorded in South East and East Asia after the 1997 crisis

361

After the Asian crisis of 1997, South Korea, Thailand, the Philippines and Malaysia experienced a moderate decline of income inequality (Table 1). The impact of improvements in terms of trade appears to have been limited, and cannot therefore account for the observed drop in inequality (Cornia and Martorano, 2012). Instead, inequality declined thanks to a pro-growth pragmatic and prudent macroeconomic policy characterized by small budget deficits that assured stability and boosted growth while government expenditure remained moderate and was exclusively financed with domestic revenue, much of which originated from direct taxation. Trade policy remained open and export growth was sustained by means of a stable and competitive real effective exchange rate. Thailand adopted de facto a managed float, while Malaysia shifted to a managed float in 2006. In both cases, after the large devaluations of 1997-8, the REER remained broadly stable including after the 2008-9 crisis. Malaysia and Thailand also introduced controls on portfolio flows to avoid an appreciation of the REER (Cornia and Martorano, 2012).

362

These countries also invested massively in public education. This raised the average number of years of schooling of the workforce, improved the distribution of human capital and reduced the skill premium. The governments also strengthened traditionally weak labor policies by introducing after the 1997 crisis unemployment insurance, and developing social insurance and assistance policies and institutions. These focused on social protection (in South Korea), the reduction of the rural-urban gap (in Thailand), and the narrowing of income gap by ethnic groups (in Malaysia). As a result, on average, income inequality fell in the 2000s by 2-3 points. These measures strengthened the redistributive capacity of fiscal policies and in 2007 the difference between the Gini of market income and disposable income was close to 4 points.

363

China: A sharp increase in inequality until 2008, and its modest decline since then

364

Despite its marked regional diversity, in 1978 communist China had a low Gini, i.e. 0.32 economy wide and 0.21 for rural area. The distributive impact of the post-1978 reforms varied markedly over time. The egalitarian agricultural reforms of 1978–84 generated only a modest upsurge in both rural and urban inequality. In contrast, income concentration grew rapidly over 1985-2000 as the second wave of reforms focused on the urban industrial sector, with the result that the national Gini coefficient reached 0.43 in 2000. This large surge in income disparity was due to the following factors: a rise in the urban–rural income gap; mounting inter-provincial inequality; a widening of rural inequality due to a rise in earnings inequality in township and village enterprises; and a surge of urban inequality due to the mass exploitation of rural workers without hukou (Selden and Wu, 2011). By repressing the wage of migrant workers, earnings inequality and corporate profits rose in line with a surge of the skill premium in the modern sector.

365

Public policy contributed to this Gini rise. The fiscal decentralization of 1978 reduced the redistributive ability of the central government to minimize regional inequality by means of transfers. The 1994 tax reform re-centralized revenue collection and allocation but failed to reduce provincial disparities, as tax/GDP ratio remained moderate and transfers continued to favor formal sector workers. Industrial policy played an even greater un-equalizing role as it explicitly favored the coastal provinces through the granting of administrative and tax privileges that facilitated the development of export industries and attracted FDI. Finally, with the dismantling of rural communes and closure of inefficient state factories, there was a de facto privatization of health and education, and a reduced coverage of the pension system (benefitting only state employees and workers of large companies and multinationals). Delays in building alternative solutions contributed to the inequality rise.

366

Despite mounting concern among central authorities and the launch of programs such as “Go West” and ‘The Harmonious Society’, during the third reform phase (that focused on export-led growth during the 2000s) the Gini coefficient grew further from 0.43 at the turn of the century to 0.49 in 2007 (Li 2015). The success of the export–lead model depended in fact on labor policies that repressed wages, and raised private, corporate and public savings to finance a rapid capital accumulation. Much of the escalation in inequality prior to and during the last decade pivoted around the hukou system mentioned above (Selden and Wu, 2011).

367

It is unclear whether 2008 marks the beginning of a fourth phase in the Chinese inequality trend. The data show that the overall Gini index declined between 2007 and 2013 from 0.49 to 0.45, following a narrowing of the urban-rural gap of income per capita from 3.3 to 3 due to the rapid rise in the wages of rural migrants and in their remittances to rural areas. The fall in the urban-rural income gap was facilitated by the introduction of pro-rural policies such as tax exemptions, the abolition of school fees, agricultural subsidies, and new health, pension and anti-poverty schemes (ibid.) made possible also by the rise of the tax/GDP from 12.2 in 1996 to 22.5 in 2010. However, both urban and rural inequality continued rising (Li et al 2015).

368

6.1.3 The bifurcation of inequality trends in Sub-Saharan Africa

369

Consumption inequality trends

370

There are very few analyses of inequality changes in Sub-Saharan Africa, not least because of limited data availability. Fosu (2015) presented evidence of changes in inequality for 39 countries based on the Povcalnet database. Table 2 of his study shows that inequality declined since the mid-1990s in 21 of these 39 countries. Inequality fell on average by half percent per year. Interestingly, the author finds that – on average - higher inequality went hand in hand with faster GDP growth. An analysis by Anyanwu et al (2016) based on a panel of 17 West African countries for 1970-2011 identified a non-monotonic inverted U-shaped inequality trend. In turn, Cornia and Martorano (2016) developed an Integrated Inequality Database (IID-SSA) that selects according to a standard protocol the best Gini coefficients of the distribution of consumption per capita taken from all existing datasets. IID-SSA covers 29 countries that account for 81 percent of the continent’s population and an even greater share of its GDP. Over 1993-2011, the average unweighted Gini coefficient of the 29 countries fell by 3.4 points, or two points for the population-weighted Gini. Yet, a country-by-country analysis shows that such average decline conceals more than it reveals. Indeed, this result is the sum of diverging falling, rising, ∩-shaped and U-shaped inequality trends, as shown in Figure 1. By restricting the analysis to the 2000s, one obtains a steadily declining trend in 17 countries (the two left panels in Figure 1) and a steadily rising one for 12 countries (the two right panels). In West Africa, inequality fell steadily in nine mostly agricultural economies out of 12, while a modest decline was recorded also in Eastern Africa. In contrast, Southern Africa and Central Africa recorded a rise since around 2003, in line with the increase in the world prices of oil and minerals. These trends point also to growing intra-regional divergence of inequality levels, as many low-inequality nations experienced a Gini fall and the high-inequality ones a rise or stagnation.

371

Factors explaining the bifurcation of consumption inequality trends 

372

What were the main drivers of these diverging inequality trends? Given the heterogeneity of growth patterns in the region, the rate of growth of GDP/c is not statistically significant. What mattered was instead the pattern of growth. Indeed, inequality rose in countries that experienced a value added shift towards sectors – such as the resource sector (Anyanwu et al 2016) - characterized by high asset concentration and capital- and skilled-labor intensity - such as mining, oil extraction, finance-insurance-real estate, and the public sector - or towards unequal informal services. In contrast, inequality fell or remained stable where growth occurred in agriculture, manufacturing, construction and a number of service subsectors (Cornia 2016). It is important to note that the Gini fell where there was a surge in the value added share of agriculture driven by increases in land yields and agriculture’s total factors productivity (Block 2010). Summing up, despite a regional growth of GDP/c of 4.1 percent over 1990-2011, a number of SSA countries followed a suboptimal pattern of growth characterized by re-primarization, de-industrialization, informal tertiarization and rising inequality. In addition, even where it declined, inequality in the region remained high and reduced the poverty alleviation elasticity of growth well below that of other regions (Beegle et al., 2016).

373

Figure 3.7. Trend of the unweighted Gini coefficient of the distribution of consumption expenditure per capita for four groups of countries, 1993-2011

374

375

Within each sector, consumption inequality varied in line with the intra-household distribution of production factors and their rates of return. In Ethiopia, where land distribution is egalitarian, rapid agricultural growth did not push up the rural Gini of 0.26. Nevertheless, hardly any new land redistribution was carried out during this period in the coninent. Yet, tenancy reforms and land titling programs improved the security of tillers, and raised investment, benefitting in particularly the poor and women (Cheong 2014). At the same time, the Land Matrix database lists 375 ‘land grabs’ in 27 Sub-Saharan African countries, including some with very low land/man ratios. The distributive impact of these transactions is controversial (Deininger and Beyerlee,2011). While land grabs may modernise agiculture, it is unclear whether large capital-intensive farms can generate enough rural jobs, promote broad development and not infringe on the rights of traditional users.

376

Educational policy affected inequality. Yet, while primary enrolments rose on average by over 20 points between 1998 and 2012, secondary enrolments rose only by half that amount. Cogneau et al. (2012 ) show that, especially in urban areas, the skill premium rose because of a rapid increase in the demand for skilled labor. An econometric analysis confirms that the number of workers with secondary or higher education over that of workers with lower education affects inequality (Cornia, 2016). And Anyanwu et al. (2016) find that increased access to secondary education and lower age dependency ratios equalize income in West Africa.

377

A persistently high population growth and ensuing population density raised inequality (ibid). Between 1995-2000 and 2010-2015, the population grew on average by 2.7 percent a year, as the regional TFR fell only from 5.91 to 5.10 (United Nations Population Division 2015). Rapid population growth increases inequality because of its impact on the land/man ratio, forest cover, distress migration to the urban informal sector, growing differentials in dependency ratio between rich and poor households, falling unskilled wages, and reduced social spending per capita. Differences have however emerged. While in Niger the TFR stagnated over time at 7.6-7.7, in Ethiopia and Rwanda it fell to 4.5 and 4.0 thanks to proactive policies (ibid). 

378

As for its underlying causes, inequality was reduced where a stable and competitive effective exchange rate shifted production towards the labor-intensive tradable sector that also offers protection to the import-competing domestic production. The opposite was also true. Finally, with trade liberalization average tariffs fell from about 15 to 8 percent. This lead to deindustrialization and a rise in inequality, and confirms the findings of Koujianou Goldberg and Pavcnik (2007) about the increase in inequality over several years after trade liberalization. Anyanwu et al (2016) found similar results on a panel of 17 West African countries.

379

Since the early 2000s the region recorded an average increase in the tax/GDP ratio and, in some countries, in the share of direct taxes in the total, that regression analysis shows was equalizing (Cornia 2016). In several countries, growing tax revenue and the cancellation of the foreign debt thanks to the HIPC initiative allowed to increase public social spending as a share of a rising GDP. Where this occurred, the effect was equalizing. For instance, in Southern Africa, public expenditure on social transfers and non-contributory pensions rose perceptibly. Where social spending stagnated (in spite of a growing fiscal space) the Gini index tended to rise (ibid).

380

Changes in global economic conditions affected inequality in a variety of ways.  Overall, gains in terms of trade in extractive industries had a dis-equalizing effect. Due to its specific features (i.e. the comparatively low-cost and a high share of migration to neighboring countries), rising remittances in Sub-Saharan Africa were equalizing. Gains in international terms of trade were also equalizing, except for the mineral rich countries. Inward FDI were unequalizing (Anyanwu et al, 2016, Cornia 2016). Finally, foreign aid rose from 15 billion over 1990-2001 to about 40 billion by 2006-7. Despite the doubts of several authors about its impact, an examination of ODA allocations since 2000 shows that they were distributed according to MDG-sensitive criteria (Hailu and Tsukada, 2012).

381

Since the mid 2000s, the incidence of HIV/AIDS declined slowly, and regression analysis shows it exerted a modest equalizing impact in those countries affected by (primarily in East and Southern Africa). The 2000s also witnessed the endogenous diffusion of low-cost and highly-divisible technologies such as cell phones, internet and solar panels, that might have integrated into the market marginalized producers and consumers. While the growth effect of such shock was favorable, that on inequality was likely to be concave, as such new technologies were initially acquired by the middle class. Meanwhile, between 1993 and 2010 the number of conflicts in the region fell from 25 in 1993 to 10 in 2010. Such decline affected favorably growth and inequality (Cornia 2016, Anyanwu, 2016) .

382

6.1.4 Declining Inequality in Latin America

383

Inequality trends

384

With an average Gini coefficient hovering between .50 and .55, high inequality is a long-standing feature of Latin America. After rising in the 1980s and 1990s, however, income inequality in the region declined rapidly until 2012 (Figure 1). [33]  Lustig et al. (2016) show that the Gini coefficient of household income per capita fell from a weighted (unweighted) average of 0.550 (0.532) in the early 2000s to 0.496 (0.483) circa 2012. During the same period, the incidence of total poverty –defined as those earning below US$4.00 per day in 2005ppp– fell from 42 to 25.3 percent, a reduction of roughly 57 million people. Applying the Datt-Ravallion decomposition reveals that, on average, 39 percent of the reduction in poverty was due to the decline in inequality.[34]

385

Remarkably, for the period 2002-2014, all seventeen countries in Latin America and the Dominican Republic showed a decline in the Gini coefficient. The rate of decline ranged from an annual average of -1.57 percent in Bolivia to -0.26 percent in Costa Rica.[35] Gasparini et al. (2016) and Cornia (2017), however, show that in the more recent period, in a number of countries the decline slowed down significantly or stopped falling altogether.

386

Figure 3.8. Trend in the average un-weighted regional Gini coefficient  of the distribution of household income per capita, early 1980s- 2014

387

 

388

Source: Cornia (2014a) and SEDLAC http://sedlac.econo.unlp.edu.ar/esp/estadisticas.php (accessed on 10 June 2016) for 2013 and 2014.

389

The data shown above has a severe limitation: the rich are undersampled or, the rich households included in the surveys, tend to under-report their income, as shown by Alvaredo and Piketty (2010). Such estimation bias may affect not only the level of inequality but also its trend. However, the literature reviewed by Cornia (2014b) indicates that, at least for Argentina, Colombia and Uruguay, while the level of the Gini coefficient corrected on the basis of the tax returns of the top one percent is consistently higher than the uncorrected Gini, the inequality trend of the corrected Gini was similar to that of the uncorrected one.

390

Determinants of Declining Inequality in Latin America

391

Interestingly, there is no clear link between the decline in inequality and economic growth. As discussed in Lustig et al. (2016), inequality declined in countries that experienced rapid economic growth, such as Chile, Panama and Peru, and in countries with low-growth spells, such as Brazil and Mexico. Nor is there a link between falling inequality and the orientation of political regimes: inequality declined in countries governed by left regimes, such as Argentina, Bolivia, Brazil, Chile and Venezuela, and countries governed by centrist and center-right parties, such as Mexico and Peru. However, while inequality fell in countries of different political orientations, the fastest decline was recorded in the social-democratic regimes.[36] Most of the existing studies point to two main explanations for the decline in inequality: a reduction in hourly labor income inequality, and more robust and progressive government transfers, with the former contributing the lion’s share as shown in Figure 2.[37] On average, 54 percent of the decline in the Gini coefficient can be attributed to changes in the distribution of hourly labor income.[38]

392

Figure 3.9 Contribution of proximate determinants to the decline in inequality

393

Percentages of contribution; Latin America, circa 2000-2010

394

395

Source: Nonparametric results are from Azevedo et al. (2013a) and parametric results were provided by CEDLAS.(Center for Distributional, Labor, and Social Studies, Universidad de La Plata, Argentina). The positive (negative) sign indicates an equalizing (unequalizing) effect of each determinant. The results shown are averages for 14 countries in the case of the nonparametric decomposition (see the authors paper) and 12 in the case of the parametric decomposition. The sum of the contributions of each determinant is, as expected, 100 percent.

396

What explains the reduction in hourly labor income inequality? The available evidence suggests that a common factor present in practically all countries was a fall in the returns to human capital –or, more precisely, in the relative returns to secondary and tertiary education—the so-called skill premium.[39] Several authors underscore supply factors: that is, an increase in the relative supply of workers with completed secondary and tertiary education, a result of the significant educational upgrading that took place in the region during the 1990s.[40]  Other authors have given more emphasis to demand factors or a combination of both.[41] Studies for Brazil and Mexico indicate that the expansion of education spending that underlies the change in labor composition by skill, in turn, seems to be associated with higher public spending per student on primary and secondary education and a rise in education coverage in rural areas. These factors eased supply-side constraints. In addition, the conditional cash transfer programs Bolsa Família (Brazil) and Progresa/Oportunidades (Mexico) reduced demand-side constraints by compensating poor households for schooling costs and for the opportunity cost of children’s labor.

397

It is not just changes in the returns to education, however, that lie behind declining inequality in hourly earnings. For example, Ferreira et al. (2014) conclude that in the case of Brazil rising minimum wages and a substantial reduction in the gender, race and spatial wage gaps explain the lion’s share of the decline in earnings inequality.  Cornia (2014a) finds that the macroeconomic conditions and rising minimum wages played a role in a number of countries.

398

The determinants of the decline in non-labor income inequality include: returns to capital (interests, profits and rents), private transfers (for example, remittances), and public transfers (for example, CCTs and noncontributory pensions). As shown by Azevedo et al. (2013a), the contribution of changes in returns to capital in Argentina, Brazil and Mexico, for example, tended to be small and unequalizing. However, as said above, a well-known fact is that household surveys under-estimate income from capital so the unequalizing effect may have been larger than current estimates indicate. Esquivel et al. (2010) show that, in Mexico, remittances proved to be equalizing and became even more so in the 2000s because they closed the gap between rural and urban household per capita incomes. Cornia (2014a) also shows that the increase in migrant remittances in total household income appears to have had an equalizing effect in El Salvador and Mexico; however, in Honduras their effect was unequalizing.

399

As for public transfers, Azevedo et al. (2013a) find that, on average, government transfers account for 21 percent of the overall inequality decline.[42] The role of noncontributory pensions, however, cannot be disentangled because they are included not in the government transfers but in the total pensions category (which account, on average, for 9 percent of the decline in overall per capita income inequality). Their analysis, therefore, may underestimate the role of government transfers in explaining the decline in inequality. For example, Lustig and Pessino (2014) show that for Argentina the large expansion in noncontributory pensions was fundamental in accounting for the reduction in inequality during 2006-2009. In the case of Brazil, Barros et al. (2010) find that for the period 2001-2007 changes in the size, coverage and distribution of public transfers account for 49 percent of the decline in inequality while in the case of Mexico, Esquivel et al. (2010) find that these factors account for 18 percent of the decline in inequality over 1996-2006. 

400

6.2 Cross-cutting issues emanating from all or most regional trends

401

Inequality is on the rise in many countries. As noted in sub-section 6.1, this is especially true for developed countries and some fast rising Asian economies while inequality trends have been more mixed in the other developing countries and in the emerging economies, with some countries, especially in Latin America, experiencing declining inequality. However, even in these countries inequalities in access to finance, education and health care remain important, while institutions in the labor market and social protection remain underdeveloped. These main cross-cutting issues are discussed in what follows.  

402

6.2.1 Access to credit and financial markets

403

The literature on financial development and inequality finds a link between better financial development and decreasing or lower inequality. For example, Burgess and Pande (2005) found that a better access to formal credit and saving opportunities led to a faster decline in poverty rates in financial less developed areas in India. Similarly Clarke et al (2006) and Beck et al (2007) show by conducting cross-country analyses that better financial development decreases inequality. Suggesting that financial development benefits also the poor. Lo Prete (2013) suggests that a key variable on whether the poor can benefit from improved access to financial markets is literacy. In turn, Agnello et al (2012) shows that a way to broaden access to credit and lower inequality consisted in the removal of directed credits, lowering of high reserve requirements and the development of a more efficient security market.

404

While the literature on financial development mainly focuses on developing countries, the literature on financial deregulation focuses more on developed countries. For example, Tandal and Waldenström (2016) found that Big Bang financial deregulation in Japan and the UK lead to an increase in top income shares. Beck et al (2010) analyze an intrastate branch banking reform in the United States which increased competition in the banking sector. They show that the reform led to less income inequality by increasing the incomes of low wage workers. So whether financial deregulation increases or decreases inequality seems to depend on the specific policy change which is implemented.

405

From a theoretical perspective the relation between financial markets and inequality is unclear (see for example Demirguc-Kunt and Levine (2009) for an overview). For example, financial development could provide finance to poor people which did not have access to finance before and therefore increase their incomes. On the other hand, financial improvement could benefit the rich by enhancing their financial service. For example, Greenwood and Jovanovic (1990) show that in the transition from a little financially developed and slow growing economy to a developed economy, a country passes through phases where the distribution of wealth across citizens becomes more unequal.

406

Another strand of the literature investigates the effect of asset price changes on inequality. For example, Adam and Tzamourani (2015) show that equity price increases give rise to increases in wealth inequality in the Euro Area, while bond price increases leave wealth inequality largely unchanged.

407

6.2.2 Access to education and health care, and human capital formation

408

Income inequality depends in an important way on the distribution of human capital across households. In this regard, inequality in education has declined significantly over the last 50 years or more in most countries. Especially in developing countries, this is mostly driven by improvements in access to human capital at the bottom of the income distribution (Castello-Climent and Domenech 2014), though large differences exist between regions, with most of Latin American countries having substantially reduced the secondary enrolment gap of the poor (Cruces at al 2014).  Despite this improvement, education outcomes remain much worse for disadvantaged groups (Dabla-Norris and Gradstein 2004). In advanced economies, in some countries rising university costs have contributed to lower access to education by the poor. In the United States, for instance, college costs grew must faster than most households’ income since 2001 (Federal Reserve 2014).

409

Health inequality is a key determinant of educational and income inequality. Over 2000-2015 average health status improved universally, including in the poorest developing countries. According to the 2015 “World Health Statistics: Monitoring Health for the SDGs”, life expectancy at birth increased on average by 5 years over this period, the fastest increase since the 1960s. Yet, inequality in health outcomes remains widespread in developing economies. For example, in developing countries the infant mortality rate is twice as high among poor than rich households. Similarly, female mortality rates tend to be disproportionately higher for lower-income groups. Inequality in health care access is even more pronounced in developing countries, as shown by Gwatkin et al. (2007) based on  Demographic and Health Surveys spanning 20 years. However, even in advanced economies, income inequality is increasingly being reflected in lower life expectancy. This is particularly striking in the United States, where income today is a stronger predictor of life expectancy than it was a generation ago (Murray, Lopez, and Alvarado 2013).

410

6.2.3  Demographic change and inequality

411

Other determinants of inequality include demographic changes (e.g. in young age and old age dependency rates) and patterns of household formation and composition. For instance, using decomposition methods for measures of inequality, poverty and richness, Peichl et al. (2012) quantified how the trend towards smaller households has influenced the change in the income distribution in Germany. The results show that the income gap would also have increased without the above demographic changes. But its level would be lower than it actually is. Similar studies have been conducted for the United Kingdom (Jenkins, 1995) and the United States (Martin, 2006). In the comparative work by OECD (2008), these findings are confirmed for almost all OECD countries. While the trends towards smaller household sizes have worsened income distributions, they only account for a small part of total inequality increases.

412

Cancian and Reed (1998, 1999) study the role of female earnings on overall inequality and find an equalizing effect due to increasing female labor force participation. In turn, Burtless (1999, 2009) and Schwartz (2010) find that the increasing correlation between husbands’ and wives’ earnings as well as the increasing share of single-person households has contributed to more inequality. Hyslop and Mare (2005) also find that increasing inequality in New Zealand is to a large extent driven by changes in household structures and attributes. Daly and Valletta (2006) and Martin (2006) find similar results and trends for the US.

413

In low and middle income countries, the increase in life expectancy and old-age dependency ratios tend to generate unequalizing effects. Indeed, the ILO (2014) has shown that roughly 70 percent of the elderly (falling to 40 per cent in Latin America) are not covered by formal old-age income-protection schemes. Lack of pension coverage is often a source of poverty and of an upward pressure on overall inequality, especially if the elderly live alone. Means tested and universal non-contributory pension schemes are however expanding – including in Southern Africa and Latin America - and have been shown to have a strong poverty alleviation and equalizing effect (Nino Zarazua et al 2011). Rapid aging may represent also a drag on economic growth (and indirectly on inequality) to the large amount of private and public resources to be allocated to pensions, health care and the care of the elderly (Vos et al, 2008).  

414

Also, a faster decline in fertility among the rich (the most common case) has unequalizing effects. Some regions (such as East Asia and Latin America) have already come close to the replacement ratio, with positive effect on income inequality, as the poor household’s fertility rate converge to the average fertility rate. As noted above, Baradacco (2015) has shown that the fastest relative decline in fertility rate and in young–age dependency ratio among poor households over 1990-2012 reduced the Gini by 0.7 and 2.0 points respectively in Chile and Peru. The opposite is true in most of Sub-Saharan Africa. In polygamous Niger and Nigeria, TFR stagnated at a high 7-7.7. In other countries, including Rwanda and Ethiopia, fertility declined steadily starting around 1995, while in South Africa TFR fell rapidly throughout the last two decades (Canning et al 2015). Persistently high TFR and young-age dependency rates raise inequality due to mounting stress on global commons that affects most the weakest, rising food prices, a decline in soil availability and fertility and forest cover. It also causes a rise in the skill premium, dis-equalizing changes in young age dependency rates of the poor in relation to those of the better off, and an unequal access to fresh water, decent jobs, and social services.

415

6.2.4 Taxation, social protection and inequality

416

The progressivity of tax systems has declined in some advanced economies over the past few decades though it as improved in some developing regions, most notably Latin America (Cornia, Gomez Sabaini and Martorano, 2014; Lustig, forthcoming). Rising pre-tax income concentration at the top of the distribution in many advanced economies has hence coincided with declining top marginal tax rates. Conditional and non-conditional cash transfers, as well as non-contributory social pensions have become an important policy tool for directing resources towards the lower end of the distribution in developing countries (IMF 2014a), but their redistributive impact varies widely across countries, reflecting both differences in the size and progressivity of these transfers.

417

6.2.5 Labor markets

418

Skill biased technological change and the resulting increasing skill premium together with the decline of labor market institutions have contributed to increasing inequality in both advanced and developing economies. This is because technological changes can disproportionately raise the demand for capital and skilled labor over low-skilled and unskilled labor by eliminating many jobs through automation or upgrading the skill level required to attain or keep those jobs´(Card and Dinardo 2002; Acemoglu 1998). Indeed, technological advances have been found to have contributed the most to rising income inequality in OECD countries, accounting for nearly a third of´the widening gap between the 90th and the 10th percentile earners over the last 25 years (OECD 2011). Evidence from developing countries economies also shows a similar trend of a growing earnings gap between high- and low-skilled workers despite a large rise in the supply of highly educated labor. Changes in the skill premium were found to explain the rise or fall in earnings and income inequality also in Asia (Kanbur et al, 2014), Latin America (Keifman and Maurizio, 2014) and several African countries (Cogneau et al., 2007) 1990s.

419

Globalization has played a smaller but reinforcing role. This has led to a reduction in the middle class in many advanced economies and some large emerging market economies (Autor, Katz, and Kearney 2006). In developed countries, the largest driver has been the declining share of middle-skilled occupations relative to low- and high-skilled occupations (Autor, Kerr, and Kugler 2007; Goos, Manning, and Salomons 2009). In developing countries, this is more due to income polarization (Duclos, Esteban, and Ray 2004; Fan,   Kanbur, and Zhang 2011).

420

While evidence suggests that labor market regulations (such as minimum wages, unionization, and social security contributions), on average, tend to improve the income distribution (Calderón and Chong 2009; OECD 2011), these institutions have been on the decline in many countries. This has led to further increases of inequality.

421

To summarize, less-regulated labor markets, financial deepening, and technological progress have contributed to the rise in income inequality in many countries. Improvements in health and education outcomes at the low end of the distribution, as well as the rise of non-contributory cash transfer programs have  mitigated some of the increases or led to declines. Demographic shifts as well as changes in taxation also affected inequality trends.  The relative importance of these factors varies across (groups of) countries.

422

7. Deep drivers of inequality

423

7.1 History and path dependence of inequality        

424

Especially in developing countries that did not undergo a transition to democracy, carried out some assets redistribution, introduced a progressive tax reform, and broadened the access to land, education and credit, the current level of income inequality is influenced by the high asset inequality that prevailed in the past. Back in time, a high asset concentration was the result of the feudal nature of some countries (e.g. India), the dispossession by colonial rulers of the land and natural resources of indigenous populations (as in Latin America and Eastern and Southern Africa), and local cultural values discriminating against women and specific groups such lower castes and religious and ethnic minorities.

425

For instance, the high inequality experienced at Independence in Latin America depended on the high concentration of land and political power inherited from colonial times (Engerman and Sokoloff 2005). This led to the development of institutions that perpetuated well into the post-WWII period the privileges of an agrarian and commercial oligarchy, by facilitating the diversification of their land and mining assets into industry and financial assets (Torche and Spilerman 2006). Control of the political system and the army ensured that such dominance continued until the spread of democracy and mass education.

426

Also, the evidence from Sub-Saharan Africa supports the hypothesis of the path dependence of current inequality, and of its relation with past property rights regimes over the land. Around Independence, in the then land-abundant West-Central Africa, land ownership was communal, and land allocations were made at the village level in a very egalitarian way. These countries were characterized by a near absence of large-scale properties, landed gentry and low consumption inequality (Moyo and Yeros, 2007). In contrast, in the white settlers economies of Eastern and Southern Africa, large estates and plantations were owned by the former colonial rulers or their successors. These countries were characterized by high land concentration, depressed rural wages and high income inequality (Table 3.9). The current level of income inequality relates closely with these differences in land property rights and the control of mining resources.

427

Table 3.9: Gini coefficients of land concentration by type of land tenure system

428

Countries with a dominant traditional

Communal land tenure systems

Land Gini

Countries with a dominant ‘white settlers’ dualistic land tenure systems

Land Gini

         Burkina F aso    (1993)

39,1

Liberia             (1971)

68.1

                  Mali                   (1960)

Uganda            (1991)

57.4

Niger                  (1980)

31.2

Tanzania          (1960)

70.0

Senegal              (1960)

46.7

Zambia             (1971)

69.9

Guinea               (1989)

45,2

South Africa     (1960)

64.3

          Sierra Leone      (1970)

42.4

Swaziland         (1971)

83.5

         Cote d' Ivoire      (1974)

41.5

Madagascar      (1970)

80.4

                  Ghana                (1970)

53.0

Mauritius          (1930)

74.2

Togo                  (1961)

45.2

Cameroon          (1972)

40,7

Gabon                (1974)

40.2

Congo (Zaire)   (1970)

53.2

Ethiopia             (1977)

42,4

Mozambique     (1999)

36.8

Average

43.0

Average

70.9

429

Source: author’s compilation on Frankema (2005) who relied mainly on FAO agricultural censuses.

430

A key issue concerns the persistence over time of path-dependence. Indeed, the relationship between initial asset concentration and subsequent income inequality may be eroded by several factors, starting from the structural transformation of the economy. In fact, the path-dependence may not survive the withering away of the agricultural share of GDP. Its survival requires imperfect financial markets, such that only households with strong initial wealth are able to borrow and invest in new sectors. It requires also that the initial assets inequality maps into a low and unequally distributed human capital accumulation. And that political democratization is postponed.

431

Yet, by itself, democracy (or revolutions) is no guarantee of a decline of inequality, as shown in the case of the recent political liberalization of the former communist countries of Europe and of the authoritarian regimes of Latin America in the 1990s. In the latter region, inequality declined only since 2002 in countries run by center-left regimes that had placed social justice at the core of their electoral and government programs (Table 3.10).

432

Table 3.10:Inequality Trends from the Early until the Late 2000s (depending on the latest available data) by the Ideological Profile of Governing Parties.

433

Total change in Gini Index during each regime

Average change

Per year of political regime

Radical Left

-3.59

-0.42

Social democratic 

-3.21

-0.54

Centrist

-1.79

-0.20

Centre-Right

-1.01

-0.08

434
  •  
435

The path dependency hypothesis may break down also because of growing conflicts between the old and new elites. History provides counterexamples to the thesis that a unified oligarchic interest creates a smooth pathway between agrarian and industrial asset inequality. The most notable of these counterexamples is the 19th century UK debate over the Corn Laws that pitted the old agrarian elite against the new industrial interests, as cheap food was a major determinant of industrial wages. In general, newly created industrial elites may sooner or later be pinned against the rural elite. In Chile, for instance, a peasant- capitalist alliance overcame agrarian elite opposition to land reform. In general terms, an elite transformation over time may weaken path dependence.

436

External factors may also affect the duration of path-dependence. For instance, Prado de la Escosura (2007) emphasizes that the gains in international terms of trade experienced during the globalization of 1870-1914 by Latin America (that had become a major world supplier of agricultural commodities) exacerbated the colonial-inherited asset inequality. Indeed, globalization raised the returns to the land that benefitted a tiny class of large landowners. Yet, the effects of gains in terms of trade are endogenous to the political orientation of governments. Indeed, while over 1870-1914, gains in terms of trade exacerbated inequality, econometric evidence for the period 2002-2012 comes to opposite conclusions, as the distribution of the benefits of recent gains in terms of trade was influenced by the region’s increased capacity to tax land and mining rents and the establishment of new redistributive institutions (Cornia 2015).

437

Demography may also change the inequality’s path dependence. For instance, in the originally land-abundant communitarian nations of West-Central Africa rapid population growth altered farming systems and asset distribution. Initially the increasing demand for food was satisfied by extending the surface of good land under cultivation. At a later stage, and due to countries’ inability to modernize their farming techniques, it raised the migration to marginal lands with low yields, the frequency of food crises, and the sale of small farmers’ land to an emerging class of medium size farmers. This raised land concentration and contributed to the proletarization of part of rural labor.

438

7.2 Demography, migration and inequality

439

7.2.1 Pace and patterns of the demographic transition

440

Overview of the demographic transition model; position of regions along the demographic transition (far advanced in OECD countries, transition countries, and much of East Asia, and parts of Latin America; in progress in South Asia, most of Middle East and North Africa, parts of Latin America, Sub-Saharan Africa; largely stalled in parts of sub-Saharan Africa); demographic implications of demographic transition; emphasis on within-region differences (e.g. in OECD countries between very low and low fertility regimes, in Africa between stalled and transition regimes (Bloom and Williamson, 1998)

441

7.2.2 Fertility, reproductive rights and early pregnancy  

442

High fertility linked to low reproductive rights, poor access to reproductive health and family planning, and early pregnancy; often also linked to low women's autonomy and rights, and also links to poor education and low labor market performance of women (and worse outcome for children); all perpetuate inequality, generally: link fertility-labor market performance

443

7.2.3 Drivers of the demographic transition; demography and inequality

444

Theoretical linkages between demographic change and within-country inequality; empirical literature on impact of demographic change on inequality; Include inequality impacts of drivers of the demographic transition, inequality effects of demographic burdens and demographic gifts (Coale-Hoover, Bloom-Williamson, Kremer-Chen, Mankiw paper on U-shape of fertility by income, papers by Tertilt on growth impacts of polygamy, De la Croix, D. and M. Doepke. 2003. Inequality and Growth: Why differential fertility matters.  American Economic Review 93: 1091-1113, Kremer, M. and D. Chen. 2002. Income Distribution Dynamics with Endogenous Fertility.  Journal of Economic Growth 7: 227-258)

445

7.2.4 Demography and health  

446

The health burden of fertility; the role of mortality reduction for demographic change (by affecting surviving children and by indirectly affecting fertility) (2016 JAMA paper on effect of income on health, Chetty et al. paper; papers by Strulik)

447

7.2.5 Migration flows and inequality trends  

448

Importance of type of migration (by skill and ability to integrate); impact of migration on inequality in receiving countries (depending on migration type); impact of migration and remittances on economic performance and inequality in sending countries (role of who migrates and sends remittances); short-term versus long-term impact; Impact of internal migration on inequality (including migration controls and barriers to migration)?(e.g. look at papers by Rappaport, Yang, McKenzie, etc.)

449

7.3 Link economic-political inequality                             

450

7.3.1 Influence on national and international policies

451

7.4 Link of social stratification  

452

7.4.1 Inequality of access to policy  

453

7.4.2 Discrimination  

454

7.4.3 Inequality of aspirations  

455

7.5 Role of social movements in affecting inequality  

456

7.5.1 Unions, labor movements, civil society alliances, NGOs

457

 

458

7.6 Preferences and norms regarding inequality  

459

7.6.1 Experimental evidence

460

    (e.g. Falk et al.)  

461

7.6.2 Regional differences

462

               World Values Survey

463

 7.6.3 Gender and religious norms 

464

7.7 Globalization and inequality 

465

The last two decades have seen the emergence, consolidation and diffusion of an economic paradigm that emphasizes domestic financial liberalization, the removal of barriers to international trade and financial flows, and the transfer of technology according to the rules of the WTO’s TRIPS agreement. While the free circulation of labor across borders is not part of this paradigm, the demographic imbalances existing between different parts of the world (and the spread of information made possible by the ITC revolution) have increased substantially migrant flows, whether formal or informal.

466

This now dominant paradigm aims at the creation of a global market in which competition among economic agents operates with limited government interference. Its proponents claim that these measures reduce rent-seeking, increase competition and efficiency, offer opportunities for export and growth to developing countries with narrow domestic markets, and promote the convergence of the incomes and living standards of poorer countries with those of the advanced ones. They claim also that the distributive impact of these policies is likely to be on the whole favorable in labor-abundant and low-savings countries.

467

The diffusion of this influential paradigm and the extraordinary development of the ITC technology lead to a rapid economic integration. As shown in Table 3.11, over the 1980-2002 average import tariffs fell by 70-80 percent of their initial level. Likewise, the capital account was liberalized substantially except in South Asia, Sub-Saharan Africa and the Asian economies in transition. And domestic financial liberalization – a precondition for the success of free capital flows – also recorded a rapid progress. In turn, the world stock of migrants rose from 154.2 million in 1990 to 244 in 2015 (UN DESA 2016). Of these, 138 reside in advanced economies (mainly in Europe) and 75 million in Asia (especially the Middle East). Over 2000-2015 there was an acceleration of the South-North migration.

468

What has been the impact of these policy and exogenous changes on income inequality? Did we observe the changes predicted by the proponents of the globalization paradigm? And what were the effects of the increase in migration observed during broadly the same period? Obviously, the impact may change according to space and time, but there are some common findings that are reviewed hereafter.

469

7.7.1 Trade liberalization

470

Its rationale was – as suggested by the neoclassical trade theory and in particular by the Hecksher-Ohlin theorem and its Stolper Samuelson corollary – that free trade leads to greater specialization in sectors that use more intensely the production factors of which each country is endowed, with mutual growth benefits for all trading partners. In developing countries with a strong supply of unskilled labor, trade liberalization was to shift production from the non-tradables and capital-intensive import-substitutes towards unskilled labor-intensive exportables, generating in this way favorable distributive effects. In contrast, in the advanced countries specialized in the production of goods requiring lots of capital and skilled labor, inequality was expected to rise.

471

Table 3.11: Changes in policy stance about domestic and external liberalization 

472

Regions

1982-90

1991-97

1998-2002

2002-10

Average import tariff*

South America

40.0

19.0

12.2

10.6

Central America and Mexico

46.6

18.1

8.8

7.2

Sub-Saharan Africa

26.7

24.9

14.5

13.2

MENA

29.7

21.9

17.3

16.2

South Asia

62.9

52.9

20.8

14.9

East and South East Asia

20.3

16.7

7.6

6.9

Asian economies in transition/1

44.5

38.9

15.5

12.6

EE-FSU

…..

11.0

9.0

6.0

Advanced economies

8.5

7.1

3.3

4.2

Kaopen index of capital account openness (-2.5 in case of closure and +2.5 for complete openness)

South America

-0.78

-0.17

0.76

1.00

Central America and Mexico

-0.84

0.29

1.18

1.67

Sub-Saharan Africa

-0.91

-0.82

-0.59

-0.56

MENA

-0.64

-0.35

0.02

0.36

South Asia

-1.29

-0.74

-0.93

-0.90

East and South East Asia

0.85

0.96

0.50

0.57

Asian economies in transition/1

-1.73

-1.31

-1.07

-1.00

EE-FSU

-1.84

-0.53

0.01

0.65

Advanced economies

0.83

1.89

2.28

2.32

Frazer Index of domestic financial liberalization (ranges btw 0 and 10, in case of total liberalization)

South America

5.1

6.8

6.9

7.7

Central America and Mexico

6.7

7.3

7.5

8.4

Sub-Saharan Africa

4.5

5.1

6.6

7.4

MENA

3.6

4.6

5.8

6.5

South Asia

4.7

5.6

6.4

7.4

East and South East Asia

5.9

6.9

6.6

8.2

Asian economies in transition/1

0.0

2.9

4.6

8.0

EE-FSU

0.5

3.2

7.4

8.7

Advanced economies

7.6

8.2

8.6

8.8

473

Source: Cornia (2014)

474

These predictions were validated in Europe and the USA during the trade liberalization of 1870-1914 and on occasion of the export drive of the Asian Tigers (Wood 1994). A study by Bourguignon and Morisson (1989) found similar results for 35 small and medium-sized developing countries. Yet, an equally important literature comes to opposite conclusions. For instance, a meta-analysis of the empirical evidence (Koujianou-Goldberg and Pavcnik 2007) concluded that trade liberalization generated adverse distributive effects. And, in examining recent Latin American data, Szekely and Mendoza (2016) found that trade liberalization was un-equalizing during the 1980s and 1990s but – as expected - such effect does not intensify once import tariffs were substantially reduced. 

475

The counter-evidence produced by these studies traces the un-equalizing effect of trade liberalization to the immobility of production factors (that could not be easily be reallocated to the export sector for lack of sector-specific skills and human capital) and the informalization of employment following the simultaneous liberalization of capital account and the subsequent appreciation of the real exchange rate, that in turn shifted resources towards the high-inequality non-traded and informal sectors. In this regard an analysis of 21 liberalization episodes over the 1980s and 1990s showed that inequality rose in 13 cases, remained constant in six, and fell only in two (Taylor 2004). Indeed, as shown also by the case of Latin America (Ocampo 2012) and Sub-Saharan Africa (Cornia 2016), trade-liberalization has led to deindustrialization and the re-primarization of production. 

476

What factors, besides the two just mentioned, explain this discrepancy between the conclusions of the Hecksher-Ohlin model and the empirical evidence? A first explanation is that developing countries often export goods that incorporate not so much unskilled labor but natural resources and skilled workers (as in mining and oil extraction). In turn, land-intensive agricultural exports have equalizing effects only in case of low land concentration.  Second, the advantages of trade liberalization are less evident in countries exporting raw materials subject to large price variations (Erten and Ocampo 2012). When commodity prices fall, these countries’ ability to import goods for the leading sector falls while employment and incomes decline. Third, for middle-income countries, free trade may turn out to be un-equalizing when exports are liberalized also in other countries with more favorable factors endowments and production structures. The best example of this situation is offered by the decision of China and other low-wage East Asian economies to assign to labor-intensive exports the role of key growth driver. Their decision undoubtedly lowered the international price of labor-intensive goods but also eroded the comparative advantages of the middle-income countries of Latin America and South East Asia vis-à-vis the developed countries. In addition, and contrary to the findings of Wood (1994) on the first wave of Asian Tigers, the phenomenal growth and export performance of China, Indonesia, Bangladesh and so on was accompanied by a rapid rise of domestic inequality due to a variety of factors including wage suppression, imports of capital-intensive investment goods complementary to skilled labor, and rising capital share. And, finally, inequality may rise when trade liberalization does not take place simultaneously for all trading partners. For instance, the developing countries exporting agricultural goods opened up to foreign imports (and so recorded job losses in the formerly protected sector) but achieved an unsatisfactory export growth not only because of their internal problems but also because of persistent protectionism in the advanced countries.

477

7.7.2 Capital account liberalization

478

Mainstream theory has until recently maintained that the liberalization of portfolio investments – i.e. purchases of bonds, shares and securities by non-residents in local stock markets, lending by foreign to domestic banks, borrowing abroad by domestic firms, families and the state, and derivatives - raise investment, growth, employment, productivity and equity in countries with low savings but high rates of return on capital and an abundant supply of cheap labor. Other supposedly positive effects include a decline in domestic interest rates, a faster accumulation of currency reserves, and a ‘disciplining effect’ on domestic macro policy. In addition, capital account liberalization was to allow the portfolio diversification of national investors, generate a disciplining effect on the developing countries’ macroeconomic policies, and improve the global allocation of world savings. Subsequent analyses have however shown that such gains were illusory, as growth did not accelerate while instability rose (Prasad et al 2003), and there is now consensus about the usefulness of capital controls.

479

Indeed, contrary to the above predictions, the evidence points to a consistent deterioration of income inequality and growth prospects associated with the liberalization of portfolio inflows and outflow, particularly in countries with weak labour institutions and social safety nets. Explanations of the discrepancy between mainstream theory and evidence suggest that: (a) left to themselves, deregulated financial systems do not perform well owing to problems of incomplete information, markets and contracts, herd behavior, pure panics, weak supervision and assets price speculation. Much of the recent instability and recession (including that of 2008-2013) derives from the deregulation of domestic and external financial transactions; (b) large portfolio inflows can cause an appreciation of the real exchange rate, reduce employment and growth in the export sector, shift resources from the tradable to the non-tradable sector, and encourage subcontracting and wage cuts in the tradable sector to preserve profit margins (Taylor 2004); (c) portfolio investment are often directed not to agriculture and labor-intensive manufacturing but to capital- and skill-intensive firms in the finance, insurance, and real estate sectors (ibid); (c) the 'disciplining effects' accompanying the liberalization of the capital account mostly had a deflationary effect by reducing tax revenue, public spending, and  national income; d) the way the bailout of bankrupt financial institutions was carried out consisted in a transfer (via taxes and inflation) from poor non-participants in the financial sector to (middle and upper class) participants, including large depositors, big borrowers, and financial institutions (Honohan 2005). Overall, an empirical analysis of the distributional impact of neoliberal policies in Latin America during the 1990s (Behrman et al. 2000) concludes that the external financial liberalization was the component of the Washington Consensus package with the strongest negative distributional impact.

480

7.7.3 Technology transfer

481

Globalization was accompanied by a rapid transfer of (costly but state of the art) technology to the developing countries and an acceleration of technological change in the advanced ones. The skill-biased technical change hypothesis suggests that in the industrialized countries the new technologies generate a demand for skills and a distribution of earnings more skewed than those emanating from old technologies. As a consequence, unless the educational system quickly supplies an adequate amount of workers with new skills, the demand for and wage of unskilled workers drop while that for skilled workers rises faster than their supply. As a result, wage dispersion rises in the sectors using the new technologies. Second, information technology reduces the cost of monitoring unskilled workers and minimizes labor shirking, so reducing the wage premium needed to ensure their efficient performance. Finally, especially in the service sector and in a few industrial branches, new technologies replace unskilled labor with physical capital and in so doing push up the capital share and overall income concentration.

482

In developing countries, import liberalization increased the access to Western-produced labor-saving technologies complementary to skilled labor, thus reducing the demand for unskilled labor, a change that worsened the distribution of income. This negative short-term effect may however be compensated over the medium term by the benefits of a more rapid industrial modernization.  

483

7.7.4 Migration and migrant remittances

484

During the globalization of the 1870-1914 period, when 60 million of mostly unskilled workers migrated from the European periphery to the New World, migration reduced income inequality in the European countries, as the ratio of unskilled wages to farm rents rose following a drop in labor supply due to migration (Lindert and Williamson 2001). The growth effects were also favorable. However, the recent migration tends to increase inequality in the countries of origin as – according to the ‘hum theory of migration’ - the unskilled poor are less likely to migrate than middle class people whose families are able to finance the high costs (between 3000 and 20,000 US$ per person) of informal migration. Remittances are therefore generally received by households in the 50th to 80th percentile of the income distribution, bypassing the people of the lowest rung. At the same time, outmigration of rare skilled workers may raise their wage in the countries of origin, leading to a rise in the wage premium and overall inequality. This is particularly true for the migration of highly skilled professional (such as doctors and nurses from Ghana) that represent a clear case of un-equalizing brain drain. Yet, also in the case of migration there are some discrepancies between theoretical predictions, and a review of the empirical literature. Docquier and Rapoport (2003) for instance argue that the standard theory does not offer conclusive evidence as to whether international migration increased or decreased economic inequality in the countries of origin. They suggest for instance that migration may not be un-equalizing in source countries when it is state-sponsored or when large migrant networks emerge in the countries of destination (as observed in El Salvador and Mexico during the last two decades (Cornia 2014). As for the long term growth effects, remittances may stimulate overall long term growth in source countries by lessening the balance of payments constraint, allowing the import of capital goods, facilitating the formation of human capital (as children staying behind have a greater chance to graduate from schools), and allowing poorer households to acquire and access productive assets and complementary inputs (see McCormick and Wahba 2001 for Turkey and Egypt). But migration might retard growth because of the brain drain and Dutch Disease it causes, and because of a contraction of domestic labor supply due to growing reliance on remittances as a source of livelihood. The overall evidence in this regard shows that remittances have a favorable effect on poverty, volatility and current consumption but have no effect on the investment rate, school enrolment rates and the long term growth rate of GDP (IMF 2005). The effect of migration in the countries of destination is controversial. There is limited evidence that migrants have a ‘displacement effect’ in low-skilled manual jobs, rather the evidence points to a ‘replacement effect’.  There seem to be however a possible effect on the level of skilled wages.

485

7.8 Inequality, Knowledge, and Policy  

486

7.8.1 Knowledge-based inequalities  

487

 Inequality is a multidimensional and multi-causal phenomenon; one of such causes relates to knowledge, and affects various dimensions of inequality, like income, health or education. Knowledge- based inequality is not new, neither intra-country nor among countries. What is new is the recognition of knowledge, technology and innovation as a major cause of inequality nowadays. As Charles Tilly put it, “a knowledge-based inequality prevails in contemporary world” (2005, 121).

488

One way of analyzing the type of causality that relates inequality to knowledge production and knowledge use is through the typology of social exclusion proposed by Amartya Sen (Sen, 2000). Social exclusion can be active (when a will to exclude is present) or passive (when exclusion occurs even if not explicitly wanted); it can also be constitutive (with particularly severe and lasting consequences) or instrumental (leading to important inconveniences). Combining these types of exclusion, a two by two matrix can be presented, the four cells representing (I) active and constitutive exclusion; (II) active and instrumental exclusion; (III) passive and instrumental exclusion and (IV) passive and constitutive exclusion. Each of these “social exclusion cells” are connected to relative deprivation; it can be shown that they recognize as one of their drivers the type of knowledge that is produced and how knowledge is used and distributed. Cell (I) can be exemplified by TRIPS and the exclusion it implies to accessing fundamental medicines. “The fundamental problem with the patent system is simple: it is based on restricting the use of knowledge. Because there is no extra cost associated with an additional individual enjoying the benefits of any piece of knowledge, restricting knowledge is inefficient. But the patent system not only restricts the use of knowledge; by granting (temporary) monopoly power, it often makes medications unaffordable for people who don’t have insurance. In the Third World, this can be a matter of life and death for people who cannot afford new brand- name drugs but might be able to afford generics.” (Stiglitz, 2007) Cell (IV) looks into the same type of problems from a “passive” perspective. Usually, when the choice of problems to be researched is made, or when the venues for innovative efforts are decided, nobody explicitly wants to exclude anybody. The point is that without a conscious inclusive perspective, the problems of those without agency, particularly without effective demand, will not be taken into account. In the realm of health, terms like “neglected diseases” or the 90/10 gap, coined by the WHO, express the exclusion, not actively pursued but nevertheless present, of a great part of the world’s population from the possibilities of a better health offered by new knowledge. Knowledge-related exclusion in Cell II can be exemplified by the possibilities to choose people - for jobs, for insurance, for fellowships or acceptance in educational institutions- from a wealth of personal data that may include as intimate data as their DNA. At the beginning of the 80s, the powerful German metal-workers’ trade-union conducted a long strike against the kind of data processing that the management was planning to use to better control production. They complain that such data processing would render the workers “transparent in face of the administration”. Besides the reasonable fears regarding Big Brother for the whole population, the possibility of rendering people as transparent as knowledge through an unregulated manipulation of personal data opens the road for active exclusion even though it may be of an instrumental type.  Finally, Cell III includes the “classical” type of social exclusion derived from unequal capabilities people, organizations and even whole societies have to produce and to use modern knowledge for solving problems. It is passive, because it is not exerted by a conscious will to exclude; it is instrumental because it may not be directly life threatening. But it is structural, in the sense that the type of “knowledge inequality” present in this cell stems from the productive structure and the relations of such structure with knowledge production and use. Christopher Freeman coined a term, “voluntary underdevelopment”, to describe the systematic preference for importing technical solutions over the effort to provide some of them from within, based on own efforts to develop scientific, technological and innovation capabilities. (Freeman, 1992) Part of the rationale of voluntary underdevelopment is the weak knowledge demand stemming from production in developing countries (Rodrik, 2007, Arocena and Sutz, 2010). Inequality follows these situations through the small proportion of highly qualified people in total employment, with its sequel of low wages and weak stimulus for personal investment in education. At a more aggregate level, the international division of labor -that shows little changes for the majority of developing countries- continues to put on one side primary producers or “maquila” type producers, and on the other side sophisticated manufacturing and services producers, again with all the sequels in terms of type of employment and associated wages.

489

Each cell represents a particular feature of the prevailing knowledge-based inequality. Behind each cell a specific type of power is excerpted through different actors and mechanisms: over global or national regulations, over research and innovation agendas, over economic structures. Fighting these knowledge-based inequalities needs building countervailing powers able to redress at least some of the more salient features of the current situation. It can be sustained, with good reasons, that even if the knowledge-based inequality represented in each cell needs a specific countervailing power to fight it, the intertwined nature of social exclusion manifestations will prevent major achievements if a more systemic and global perspective is not pursued. However, precisely due to the specific nature of the different types of knowledge-based inequalities, it is important to identify the particulars of the power influencing more directly each of them as well as the actors able to try to build alternatives.

490

7.8.2 Fighting active vs. passive social exclusion

491

 In terms of countervailing action, active social exclusion, including its knowledge-based consequences, is quite distinct from passive social exclusion. Fighting the knowledge-based inequalities stemming from TRIPS, for instance, calls for global action, even if in some occasions the political will and political strengths of some “continental” countries like Brazil and India were able to succeed, like in the case of AIDS’s generics. But when the will to exclude is not evidently present, and so a direct political action lacks a target, the building of countervailing action needs, first of all, the identification of the roots of passive social exclusion through knowledge and innovation.

492

One of such roots is related to the dominance of a heuristic of problem-solving based on the abundance of different types of resources. From such heuristics flourish a style of innovation that is unable to deliver solutions in contexts characterized by diverse types of scarcities. The examples are overwhelming, ranging from unaffordable prices to inadequate or nonexistent infrastructures. The vaccine against the bacteria Haemophilus type B, once developed, was subsidized through WHO and UNICEF to fight one of the deadliest illnesses for children under five years. Its massive use was prevented by the cost of manufacturing the vaccine, because it was a biological vaccine which requires a production procedure based on fermentation. Nobody was trying to get a synthetic vaccine, leading to a negligible cost of production once developed, because the scientific complexity of the endeavor was extremely high and, in the developed world, the market cost of the biological vaccine was affordable for private clients as well as for public health policies. In Cuba, such synthetic vaccine was developed, in a joint effort with a Canadian university, after fifteen years of research (Kayser, 2004). What is remarkable here is the “deviant” heuristics followed to solve the problem of how to get an affordable vaccine for an important illness. Innovations which operations require the existence of a cold-chain, a steady electricity supply or drinkable water, are not solutions at all when such infrastructures are not in place. Some times what is needed is to devise a solution that can pass from such infrastructures; in other cases, the infrastructure needs to be built, but in a totally new way to take into account context conditions.

493

We can call the ability to be scientific, technologically and innovative “heterodox” in the sense depicted above, as the capacity to innovate in scarcity conditions (Srinivas and Sutz, 2008). Such capacity has been recognized in the realm of grass-root innovation; the Indian Honey-Bee Network is an outstanding example of that; the story told by Jens Muller on Tanzanian’s blacksmiths solving-problems abilities is another (Muller, 2010). The Cuban vaccine example belongs to a different type: it is not rooted in people developing odd ways of solving-problems from their experience-based knowledge; it is a result of the rigorous application of formal scientific knowledge. Innovations rooted in a deviant logic respect the “canonical” and usually marketed way of doing things have got a new name: frugal innovations. As the name suggests, “frugal” means done with “dramatically less inputs” of any kind, even though achieving at least equally good results. (Bound and Thornton, 2012).

494

One important source of knowledge-based inequality is the huge differences in coverage and quality associated with access to fundamental goods and services. When the only goods and services available are of the canonical type, exclusion from its use follows to some extent. Public policies may soften such exclusion, but as the recent inclusive health policy in Brazil showed, imports of medicines, pharmaceuticals and medical devices led in seven years –from 2003 to 2010- to a huge deficit (Maldonado, 2011). Probably not every current good or service allows, at least for the time being, for a frugal alternative solution. But an exploration around those that constitute the fundamental core of public goods may have as an outcome alternative ways of producing and delivering them that amplify quality and access. It is important to bear in mind that alternative solutions can very often be found, where the balance between what is understood as a cost and what as an advantage of said solution may widely vary with context, that is, with the border conditions of the problem for which the solution is attempted. But alternative solutions need alternative approaches: as David Hess put it, “Behind the alternative product is an alternative knowledge and technology that had to be developed in order to bring it into existence” (2007:2).

495

The capacity to innovate in scarcity conditions is a cultural trait present in societies where scarcity is part of the National Innovation System landscape. Usually this cultural trait is rendered invisible by the overwhelming legitimacy of the canonical way of solving problems. Even when the excellence of an alternative frugal innovation has been proven, it is usually considered as a temporary solution until sufficient money is available to acquire the “good” one. To countervail this cultural bias is important for overcoming one of the facets of knowledge-based inequality. A systematic attempt to develop frugal innovations put forward by different public policies involved with the provision of public goods could be helpful in this regard. This calls for a new venue in STI policies.

496

7.8.3 Production of knowledge and inequality

497

Another root of knowledge-based inequality stems in part from the kind of knowledge that is produced, and from the knowledge that is not produced. Academic research agendas are molded upon a quite diverse set of influences, internal to the academic world as well as external to it. The latter express themselves in terms of effective demand: “Because political and economic elites possess the resources to water and weed the garden of knowledge, the knowledge tends to grow (to be ‘selected’) in directions that are consistent with the goals of political and economic elites” (Hess, 2007: 11). This does not mean that knowledge and innovation policies can only be a reflection of vexed interests, particularly because policy makers may represent and translate into policies the needs of non-elite parts of the population, frequently as part of social fights. However, what is not so frequent is to express this type of representation and translation in terms of STI policies. With little exceptions –even though some signs of change can be envisaged- STI policies follow the path “business as usual”, meaning that effective demand represented by political and economic power commands the external influences on the academic agenda. As a result, “The prioritization of research tends to create huge pockets of undone science that result in the systematic nonexistence of selected fields of research” (Hess, 2007:22). “Undone science” is another way of describing the consequences of the bias in research agendas due to external influences. It should be mentioned, however, that “external influences” may have quite different meanings in central and in peripheral countries –this denomination referring to countries that, respectively, have knowledge-based and innovation driven economies (de la Mothe and Paquet, 1996) and those that do not have such type of economies. In the former, terms like “academic capitalism” (Slaughter and Rhodes, 2004) describe the external influences dubbed more powerful over academic agendas; in the latter, though, more often than not what prevails, as already mentioned, is a marked lack of knowledge demand from the productive sector. In peripheral countries, then, undone science recognizes not so much the overwhelming presence of a biased knowledge demand but simply the lack of it.

498

“Undone science” can also be a result of forces that operate within academia. Researchers are institutionalized workers, who build their working situation –post-graduate studies, tenure, level achieved in the academic ladder, following tacit or unspoken rules as well as quite explicit ones. The impact of the academic evaluation system on the science that is effectively done has been object of close scrutiny (Hicks, 2004, Martin and Whitley, 2010). In a recent account of the strategies followed by some Dutch academic communities to cope with institutional evaluation it is said: “Our material shows how research activities become increasingly assessed and defined by their potential for translation into quantitative measures of quality. Other criteria of scientific quality, e.g. epistemic originality, societal relevance and social responsibility become redefined through their relations to quantitative indicators. We understand this to be in tension with policy goals to encourage innovative, societally relevant and responsible research” (de Rijcke and Rushforth, 2015).

499

Three features through which the prevailing academic reward system leads to undone science in relation to the problems of the most deprived part of the population are: (i) the time it takes to identify such problems through dialogues with different actors, leading to a lower rate of scientific productivity; (ii) the interdisciplinary nature of the approaches usually needed to address such problems, with the concomitant difficulties to harvest citations in well-established journals, and (iii) the usually local or national interest of the issues to be researched, leading to the same type of difficulties as (ii).

500

Again, as in the former point related to “heterodox” solutions, widening the scope of research agendas to include with full academic legitimacy social exclusion problems requires changes in current STI policies. For a long time, it was thought that STI policies’ main contribution to social inclusion was either a structural change in production leading to formal or good quality jobs hopefully diminishing the burden of the informal economy, or sustained economic growth leading to fiscal space for usually quite expensive social policies. These were the expectations of ECLAC for Latin America since the 1990s, when they produced their masterpiece “Productive transformation with equity”; they were not fulfilled, though. The present level of inequality in Latin America is importantly lower than it used to be, but the question is if this is the result of a structural transition towards a more dynamic and inclusive society that makes the most of knowledge and innovation, or the result of expensive social policies that will not be sustainable once the exceptional period of economic growths recedes. Actually, there is a growing recognition that “smart” growth lead by science, technology and innovation is not bearing fruits for everyone, and “equitable” growth is advocated as a new challenge (Lazonic and Mazzucatto, 2012).

501

A question may then be formulated: how can STI policies be devised in such a way that knowledge-based inequality is recognized as a problem to be fought and concrete policy instruments may be designed to put forwards such fight?

502

7.8.4 Devising STI policies to fight knowledge-based inequality

503

Probably, there is not a one-size-fits-all answer to this question. For instance, differences in the legitimacy of STI policies in national contexts would lead to different strategies to blend these policies with social concerns. In developing countries, STI policies have generally not met expectations, which has eroded their legitimacy. They are residual in the policy arena, receiving a very small proportion of the public budget and without a clear mandate in terms of goals to be fulfilled. On the other hand, at least in Latin America, social policies are highly legitimized and have been effective in sorting millions of people out of poverty, even though results have not been so successful in term of inequality. Social policies, besides working around the most obvious problems derived from deprivation, are a unique tool for identifying problems leading to multidimensional forms of inequality. Not all of these problems will be “knowledge-based”; moreover, not all of them would need the help of knowledge to reach a solution. But if the problems identified by social policies that do need knowledge and innovation to be efficiently addressed enter into STI policies radars, then a sort of virtuous circle may be established. Social policies may look into social problems thinking of knowledge and innovation as potential tools at their disposal: this is what the Secretary of Science and Technology of the Brazilian Ministry of Health is expected to do through a direct link to the Ministry of S,T&I. If STI policies assume as part of its mandate contributing to the solution of the problems detected by social policies, such problems, besides becoming visible, would become legitimated (i) as academic problems for research (if a parallel transformation of research evaluation is put in place) and (ii) as innovation problems for firms (probably through the market-creation effect of technology public procurement).

504

Social policies self-assumed also as STI policies and STI policies self-assumed also as social policies would be an improvement in relation to the silo-like policy design we have today, where knowledge and innovation are alien to social concerns and, even worse, social concerns are seldom taken on board by research and innovation agendas.

505

The question of who may be the stakeholders for these type of policies is a fundamental one. Perhaps this question should be put in a slightly different way: how can stakeholders for this type of policies be built?

506

A last word on possible approaches to the issue of inequality when looked from a STS perspective. The approach taken in this small text is that (i) given that access to a range of knowledge intensive goods and services is key to redress knowledge-based inequality,  (ii) “canonical” ways of providing such goods and services do not take into account the multidimensional scarcities that some people and some societies suffer, and (iii) the knowledge and innovations needed to be really useful as problem-solvers belong to a good extent to the realm of undone science, so different STI policies than the main-stream ones need to be devised. But other approaches are needed; for instance, redressing the knowledge-based inequality represented in Cell 2 requires that people access dignified working conditions that, in particular, open for them the road of continuing learning. Another issue to be tackled -that is absent in this text- is that of the participation of deprived people in the process of knowledge production oriented to solving their problems, starting with participation in the very definition of what problems are those. This is simply to state that a STS regard on inequality needs to be as multidimensional as inequality itself is.

507

8. Policy issues: the scope and limits of policy to affect within-country inequality

508

8.1 National policy issues for addressing within-country inequality 

509

As described in previous sections, the world is an unequal place. Income and wealth inequality between and within countries is still quite large and pervasive and, in a substantial number of countries, inequality has been on the rise. Unequal distribution of assets translates into unequal opportunities. Unequal opportunities translate into earnings inequality. Concentration of power and wealth translates into weak, fragmented, and unfair social contracts. The concern with inequality is based on notions of justice and fairness (Rawls, 1971) and because of the negative consequences high inequality has on other outcomes we care about.  High inequality, for instance, can result in lower economic growth (Ostry et al., 2014; World Bank, 2005), rich-biased policies (Esteban and Ray, 2006; Stiglitz, 2012), persistently low social mobility (Coleman, 1974), and to weaker incentives for cooperation and coordination, to political polarization, and to the weakening of the social contract (World Bank, 2017). 

510

By adopting the Sustainable Development Goals (SDGs) in September 2015, countries worldwide committed to make the world a fairer place. The international community committed to eradicating poverty and hunger and achieving healthy lives, quality education, gender equity, and sustainable development. Countries have also committed to promoting full-employment growth, decent work, peaceful societies and accountable institutions, as well as strengthening global partnerships for sustainable development. Notably, in goal 10, countries committed, for the first time, to reducing inequality.

511

Societies have different means to change inequality of both opportunities and outcomes. Some of the actions should focus on improving the conditions of the poor, the vulnerable, and the socially excluded. That is, on increasing the assets—in particular, human capital-- the opportunities, and living standards of those at the bottom of the social ladder (Atkinson, 2015; Basu, 2013; World Bank, 2001).  A second set of policies should be geared at supporting the growth and sustainability of a strong middle class (Atkinson, 2015; Ferreira, et al., 2013). Finally, given the large—and in many countries increasing—concentration of income and wealth at the top, policies should be aimed at curbing the excesses of concentration of income and wealth among the rich (Atkinson, 2015; Piketty, 2014). 

512

Actions will need to focus on leveling the playing field, setting boundaries on market outcomes, and redistributing income and wealth through taxes and transfers –including the provision of services. In all three instances, the power of the state (individually and through multilateral mechanisms) to redistribute assets, income, opportunities and power through laws, regulation, and fiscal policy can play a key role. Policies which focus on the poor and middle classes must ensure to include measures designed to protect the losers from the undesirable consequences of economic progress as well. While globalization and technological change can bring higher growth, improve the lot of the poor in large parts of the world, and reduce inequality between countries, they can also generate significant dislocation and downward mobility for vast sections of the population. Development is always uneven and generates tensions and demands for redistribution of resources and power (Ray, 2016; World Bank, 2017).

513

8.1.1 Building assets and enhancing opportunities of poor people, and promoting social inclusion

514

Increasing the assets that individuals have, the way that they can make use of them, and the returns they obtain is a central aspect to improve the distribution of productive opportunities.  Box 3.1 describes a framework, based on the assets approach, to organize the discussion around the different types of instruments of redistribution available to policymakers. As shown, policies can contribute to redistribute productive opportunities via the provision of public goods and services (that is, before the market, or ex ante), as well as by enhancing access to markets. Policies can also contribute to redistribute income directly, through the fiscal system of taxation and transfers (ex-post redistribution). Improving the way that institutions function can have effects on both the redistribution of opportunities, as well as in terms of the reduction of inequality in outcomes. In particular, enhancing the way that the system is responsive—from the design of policies to the allocation of resources— to the needs and interests of actors who tend to be left out of the bargaining process can have an impact both on current and intergenerational inequality.

515

Box 3.1: Policies that can impact redistribution: the asset based approach

516

Policies can affect inequality through various mechanisms. An adaptation of the assets-based approach (Attanasio and Székely, 1999; Bussolo and Lopez-Calva, 2014; and more recently, Lopez-Calva and Rodriguez-Castelan, 2016) is a useful framework to think about the different channels, and their interaction. In the short run, the income-generating capacity of individuals (and their related contribution to aggregate growth) is determined by the assets that households have and the opportunities that they have to exploit these assets, given market conditions. Indeed, inequality and growth can be conceptualized as being jointly determined (Ferreira 2010; Chenery 1974). In the long run, the ability of individuals to accumulate assets and to use those assets productively is shaped by policies—such as those related to the provision of goods and services.

517

We can think about assets in terms of three elements: the stock of assets that people possess, the rate at which these assets are used to generate income, and their returns—such as wages or interest rates. The public provision of goods and services allows individuals to increase their stock of assets, such as in terms of human capital (for example in the form of health, education, or skills) or physical capital, (such as machinery, seeds, or land). These policies can help equalize opportunities across individuals, independently of their initial circumstances. Policies related to the provision of public health and education, policies that provide job training, or policies that encourage the supply of affordable credit and that facilitate savings can all contribute to redistribution ex-ante by enhancing the ability of poorer households to increase their stock of assets—including by influencing the decision of individuals to accumulate those assets.

518

Other policies have the potential to improve the distribution of economic opportunities, increasing households’ ability to exploit their assets, such as policies that implement land reform, policies that aim to improve access to markets, and policies to strengthen labor markets in order to promote employment. These policies can help equalize the opportunities of individuals to use their capital and labor to generate income—for instance, utilizing their skills to participate in the labor market, exploiting their machinery, or their land for agricultural production. Promoting an environment of investment and innovation and a stable, growing macroeconomic environment can expand access to opportunities and job creation. Minimum wage and other labor policies can affect labor market income directly and indirectly, through their effect on the amount of labor supplied, and demanded.

519

Policies also have an effect on the returns that households obtain for their assets, such as wages, interest rates, rents from property, or prices of land. The direct role of fiscal policy in redistributing income is discussed further below, but both it, and monetary policy also affect the returns to factors of production, which in turn have a feedback impact on the decision of individuals to use their assets more intensely or not.a

520

Two other elements are necessary to look at policies that can bring about a reduction in inequality. The (market) income that households generate based on their assets (their stock of assets, how intensively they are used, and the returns obtained) is complemented by the transfers that households receive, which can be privateb or—the focus of our interest—public. Public transfers include the benefits provided by the government that complement individuals’ income such as conditional cash transfer programs, pensions, unemployment insurance, or programs that provide disaster relief and other transfers to mitigate the effects of shocks.c These social protection systems can contribute to a reduction in inequality of outcomes, redistributing resources toward the most vulnerable. Finally, households’ full income generation capacity is also affected by the other side of the fiscal system, that is, taxation. Social spending necessarily requires that resources be collected in order to be redistributed. Fiscal policy in this way influences inequality ex-post through both transfers and taxes.  

521

Countries’ fiscal systems have different redistributive capacities depending on their structure of direct and indirect taxes, transfers and subsidies. This redistributive capacity is quantifiable, by comparing the average measure of inequality (as captured by the Gini coefficient) based on the market income of individuals before the fiscal system, and after, once the effect of taxes and transfers is manifested. As shown in the sections below, the redistributive potential of fiscal redistribution in many developing countries appears to remain untapped.

522

As discussed in the World Bank’s World Development Report (WDR) 2017 Governance and the Law, how much countries redistribute can be understood from different viewpoints (World Bank, 2017). It may be reflecting the incentives that governments have to collect and redistribute resources—where more redistribution is associated with more checks and balances (Besley and Persson, 2011). It can be a manifestation of the preferences for redistribution of a given society. It is also a result of the relative ability of actors to influence policymaking and the allocation of resources.

523

 _______________________

524

a The framework also considers the set of prices of the basket of goods and services that households consume, which is also influenced by fiscal and monetary policy.

525

b Such as domestic and international remittances, and in-kind transfers from other households.

526

c In fiscal incidence analysis, the provision of public goods and services (such as education or health) and other subsidies, are also considered transfers, of the in-kind nature (see CEQ methodology, CEQ, 2016).

527

Expanding the human, physical, natural, and financial assets that poor people own or can use can be accomplished through, for example, land reform, programs which distribute shares from privatization of public enterprises among the population and reforming inheritance laws. One key way of leveling the assets playing field would be to put in place a minimum inheritance (capital endowment) paid to all at adulthood (Atkinson, 2015). Provision of housing subsidies for low-income groups can also serve to provide the poor with an important asset. And, over time, more education can prove to be the critical asset for the poor. Efforts are needed to reduce the large inequalities in access to education and training and to upgrade the skills of the poor.

528

Building human capital starts within the household. Policies aimed at providing information and access to reproductive health so that households can make informed and conscious decisions about their desired number of children and prevent unwanted teenage pregnancies are also key ingredients of human capital development (Azevedo, et al., 2012). Human capital accumulation may also suffer if poor households´ infants and young children are malnourished. Crucial are early-intervention programs in health and nutrition (for example, mother and child health programs, vaccinations and other health interventions) and basic infrastructure investment (running water, electricity, transportation) because of the synergies at work between sound nutrition and people's ability to use new learning technologies (distance learning institutes, distance high-school education). Reforms to the institutional apparatus for social services delivery need to make sure that the poor have access to these services (World Bank, 2003). Improving the quality of services requires ensuring the participation of poor communities and households in choosing and implementing services and monitoring them to keep providers accountable. Community-based schemes to protect water resources and other elements of the natural environment should be supported.

529

Perhaps as important to building the human capital of the poor are programs that serve to increase the demand of households for health and education services. Actions to increase demand include improvements in the quality and availability of social services and compensation of the poor – through direct transfers – for the complementary costs (e.g., transportation, school materials, and so on) as well as the opportunity cost of the time that household members spend in school or in health facilities. They also include giving parents additional incentives to invest in the education and health of their children. Unconditional and conditional cash transfers (e.g., stay in school programs, human development programs, and so on) are two instruments that evidence shows can help to increase the demand of households for education and health of their offspring (Box 3.2).  Through their short-term effects as well as their longer-term impacts they can also contribute to reducing inequality (Box 3.3).

530

Box 3.2: The impact of Conditional Cash Transfer (CCT) programs

531

Conditional cash transfer programs, or CCTs, provide cash benefits to poor households as long as specific conditions, such as regarding school attendance and health/nutrition checkups, are met. Typically, these programs have two main purposes: the alleviation of poverty in the short term, facilitated through the cash transfer component; and the reduction of poverty over the longer term, by incentivizing the accumulation of human capital (see Box 3.1 on assets).

532

A substantial amount of literature has documented the positive impacts of conditional cash transfer programs. The estimations of Cardoso and Souza (2003) suggest that Bolsa Escola in Brazil had a positive and significant effect on school attendance—though no effect on child labor. Glewwe and Olinto (2004) find that the PRAF II program in Honduras has positive impacts on educational outcomes, increasing enrollment rates and attendance and reducing dropouts (however, again with no effect on child labor force participation). Behrman et al. (2005) look at the impact on health outcomes of PROGRESA in Mexico, and—controlling for unobserved heterogeneity—find a large and significant effect of the program’s nutritional supplements on child nutrition. In Colombia, the evaluation of Attanasio et al. (2005) of Familias en Acción finds effects of the program on increased consumption of food and clothes; increased school attendance at the secondary school level and the likelihood of being up-to-date on healthcare visits. Schady and Araujo (2008) study the impact of the Bono de Desarrollo Humano (BDH) in Ecuador, finding a large, positive impact on school enrollment. Fiszbein and Schady (2009), in a thorough review of the evidence, particularly from impact evaluations, find that transfers overall have increased consumption levels and contributed to the reduction of poverty, while the potential undesirable effects of such programs, such as declines in labor market participation have been modest.

533

Nonetheless, recent studies are beginning to throw into question how significant the positive effects of CCTs have been over the longer term. Some long-term studies reveal lasting impacts; such as those found by Barham et al. (2014) on the effect of Nicaragua’s Red de Protección Social on educational attainment; or the long-term effect of an antipoverty program that as an asset transfer component in India, as assessed by Banerjee et al. (2016). Yet, many others reveal discouraging results. For example, studying the 10-year impacts of BDH in Ecuador, Araujo et al. (2016) find a small increase in the probability that girls graduate from secondary school, but no effect on the probability of attending tertiary education or on working. Other long-term impact evaluations have shown small effects on beneficiaries’ labor market conditions (Rodriguez-Oreggia and Freije, 2009; Molina-Millan et al., 2016).   

534

Why could CCT programs be having a low impact on addressing long-term constraints? The answer might lie in the underlying assumptions. CCTs are designed under some key implicit assumptions: i) co-responsibility (conditionality) is effectively enforced; ii) the selection of beneficiaries is based on objective indicators and free of manipulation; iii) the quality of services provided—education and health—is high enough to affect the beneficiaries’ productive capacity in the future; and iv) the economy will generate quality employment opportunities for the beneficiaries when they enter the labor market (UNDP, 2010). These assumptions do not always hold in reality. Considering these constraints can make CCTs—and other instruments—more effective in breaking the cycle of poverty.

535

There is also growing research on the efficiency of CCTs relative to other unconditional programs. In a review of data from different studies, Baird et al. (2013) find that explicitly conditional interventions that sanction noncompliance have larger effects than those with no conditions or without enforcement. Ozler et al. (2016), on the other hand, find that even though the effects of an unconditional transfer program on health and female empowerment in Malawi were short-lived, children born to the beneficiaries of these programs had a higher height-for-age z-scores. CCTs, conversely, had an impact on educational attainment, but no effect on health, labor market outcomes, or empowerment. These results, the authors suggest, point both to the promise and limitations of CCTs. It has also been argued that, as transfers are conditioned on the consumption of normal goods (education or healthcare) relatively richer households are more likely to consume more; while, if conditionality is understood as a cost at the margin, it may be making the poorest opt out. Rodriguez-Castelan (2016) proposes a framework that models the decision making of households to participate in CCTs, depending on whether conditionality exists. Comparing the distributional effect of a CCT program relative to an unconditional one, the paper suggests that the latter would be favored under a sufficiently high degree of poverty aversion.           

536

Box 3.3: The impact of Cash Transfer (CCT) programs on (gender) inequality

537

Cash transfer programs have been adopted and practiced in several countries in Latin America (such as Brazil, Peru, Mexico, and Nicaragua), Asia (including India, China, and Indonesia) and in African countries such as Zambia, South Africa, Ethiopia and Tanzania (DFID, 2011), among others. In such countries, cash transfers have played an important role in reducing income inequality through the transfer of resources from higher income households to lower income households via the tax system, and in addressing poverty, vulnerability, and development challenges (Gertler (2005). Research evidences, for example, that the Gini coefficient in Brazil declined by 5.2 points between the early 1990s and 2008 (Holmes et al, 2011). Soares et al (2007) further state that the Bolsa Familia program in Brazil was responsible for 21% of the total fall in the Gini coefficient between 1995 and 2004, and that a total of 12-14% of the reduction in inequality between 2001 and 2004 was attributed to Bolsa Família.  Moreover, between 1996 and 2006, the Oportunidades cash transfer program in Mexico led to a reduction of poverty gap in rural areas by 19% whereas Bolsa Familia facilitated poverty reduction by 12% between 2001 and 2005  (Dercon, 2011). The transfers have also evidenced positive impacts on the health and nutritional status of children in situations of chronic food insecurity as in Ethiopia and on the livelihood security of poor households (DFID, 2011).

538

African countries have also witnessed positive impacts of cash transfers on income inequality and poverty. The programs have further improved human capital by enhancing education and health outcomes, strengthen income generating capabilities, and enhanced inclusive educational opportunities for poor children. The South African Child Support Grant (CSG) for the period 2002-2004 contributed to the reduction in hunger among children receiving the CSG compared to those not receiving it (Samson et al, 2011). The Gini coefficient in this country was 7% in 2005/06 points lower as a result of the national social transfer program (Statistics South Africa, 2008). In Namibia, cash transfers have also documented great impact on poverty at lower levels of poverty significantly reducing the incidence of poverty by 22% and 10% at the lower and upper bound poverty line respectively (Levine et al, 2009). Social transfers in Namibia have also played role in reducing inequality.  The Mchinji Social Cash Transfer Pilot Scheme in Malawi documented significant positive health and education impacts over the period 2007-2008 (Miller et al, 2011). The authors evidenced a 10% reduction in morbidity for children below 19 years, and a 4.2 % increase in school enrolment for children aged 6 to 18 years.

539

Younger, Myamba, & Mdadila (2016) used methods developed by the Commitment to Equity and data from the 2011/12 Household Budget Survey to assess the effects of government taxation, social spending, and indirect subsidies on poverty and inequality in Tanzania. Findings indicated that Tanzania redistributes more than expected given its relatively low income and inequality. This was attributed to both direct and indirect taxes that were documented to be more progressive than in other countries. The authors further assessed, among other aspects, the effect of the conditional cash transfer program in Tanzania and found it to have an excellent targeting mechanism, and concluded that that if the CCTs were expanded to half a percent of GDP, a fairly typical amount for lower-middle income countries, poverty could be reduced by about 1.5 percentage points. Moreover, Taylor (2015) evaluated seven cash transfer programs in Africa and evidenced that a dollar transferred to eligible households causes considerably more than a dollar in income in the local economy, indicating that cash transfers have the capacity to create income multipliers and have the ability to stimulate growth in developing economies.

540

Moreover, evidence indicates that where transfers are paid directly to women as in Latin America such transfers have potential to increase women’s power to make decisions on household expenditure, and providing them with financial security, self-esteem and social status (Ellis, 2008; DFID, 2011). This means that transfers paid to women can potentially increase their control over household resources they have and have implications for women’s empowerment.

541

Cash Transfer Program in Tanzania

542

The Government of Tanzania started, as part of its Tanzania Social Action Fund, a Productive Social Safety Net (PSSN). This program focuses on poor households in village communities. In the light of the Household Budget Survey (HBS) 2012 findings and in line with the reduction of the case load of extreme poverty, the program plans to benefit about 1 million households by the end of program. The program aims to enhance the livelihoods and reduce the risk of poor and vulnerable households. This has become TASAF’s most famous social protection program thus far.

543

Even when the program operational manual does not state explicitly that payments should be made to women the woman is the default recipient of the program cash transfer in mixed households.  It is worth pointing out on a gender sensitive action in which during cash transfer pilot phase, TASAF deliberately modified the mode of cash transactions to ensure that women receive the cash on behalf of household beneficiaries.  This was a later modification after documenting cases of male recipients inappropriately spending CCT funds, principally on alcohol and mistresses. Such misuse of program cash was not evident in some country CCT programs including Viet Nam (Humphries, 2008), Lesotho (pilot) (Slater and Mphale, 2008), Nicaragua (Maluccio and Flores, 2005), Colombia (Attanasio and Mes- nard, 2006) as well as Mexico’s Progressa (Rubalcava et al., 2004) and the Bolsa Familia in Brazil.

544

There is also a learned recognition that women are more likely than men to use resources for the benefit of the family, spelled out in Tanzania’s National Social Protection Framework (2012) and DFID (2011). Analyzing different concepts of risk and vulnerability, Sabates-Wheeler and Kabeer (2003), for example, acknowledged that the informal systems of social assistance and security in the African society disproportionately places on women the entire burden of looking after the family, especially the elderly and children in the case of sickness, conflicts and poverty. It is unfortunate that the society of Tanzania, as is for most of Africa, does not recognize this effort and contribution, thus leaving women with little or no support from existing forms of social and economic protection.

545

Women with young children in Latin America are not only trusted to be the recipients but are also believed to use the resources to benefit the whole family. As an example, it was evidenced that the Progresa/Oportunidades CCT program in Mexico gave cash to women only, which eventually increased power to make decisions on household expenditure, and provided them with financial security, self-esteem and social status (Ellis, 2008; DFID, 2011). CCT programs in Brazil have also demonstrated improved women’s status due to a regular income and participating in the labor market. This was evidenced in Suarez, M. et al. (2006) and Soares, Ribas, and Osorio, (2007) who analyzed gender inequalities and evaluated the impact of the Bolsa Família CCT program in Brazil. Analyzing lessons to be learnt from Malawi’s social cash transfer pilot scheme Schubert and Huijbregts (2006) also evidenced reduced likelihood of female and child-headed households engaging in risky survival behaviors, such as prostitution. Results from an evaluation of pilot cash transfer in Viet Nam Showed that some of the women experience increased financial security, and that the tension in the household was reduced as the result of cash transfer (Humphries, 2008).

546

Can Women Equally Benefit through PSSN?

547

TASAF looks at gender in terms of a shared benefit perspective, that is, the empowerment of beneficiaries, both men and women, together with people with special needs. The program adopts a household approach whereby the money is for the whole household but it should be collected by a woman unless there is no existence of a woman in the household because they see money as safe in the hands of the women, who the ones cooking now and will always continue cooking for the household. Evidence is yet to be generated to assess whether or not the PSSN design will create equal benefits for women as for the men. Hence REPOA is currently conducting an evaluation study to assess the impact of PSSN on women’s empowerment. The baseline findings of this study, which was conducted in 2015-2016, are yet to be published by (forthcoming Myamba et.al 2016). The follow-up survey and evaluation will be conducted in the year 2017. 

548

On the positive side participants of the baseline study were optimistic about women benefiting through involvement and consideration as recipient and managers of the family PSSN program cash.  Even with these positive perceptions about women’s benefits through PSSN program, participants reported on the other hand not benefiting as much as the program was intended due to male domination and oppressive behavior which is ingrained in the culture of patriarchy.

549

Special action is required in many societies to tackle socially based asset inequalities.  These inequalities can take several forms. For instance, discrimination (explicit or implicit) in access to housing and land against specific ethnic, racial, religious groups or people of different sexual orientation. A second example are the marital, inheritance, and property regimes that treat women unfairly (Deere et al., 2013). Discriminatory laws and regulations should be eliminated and discriminatory practices should be punished. In specific contexts, desegregation and affirmative action laws will be indispensable to correct long-standing exclusion and oppression of certain groups. Since poor people and the socially-excluded lack the resources and the information to access the legal system, measures such as legal aid and dissemination of information on legal procedures are especially powerful instruments for creating more inclusive and accountable legal systems.  Of course, state action can also be essential to change the social norms that contribute to the perpetuation of socially based asset inequalities (see Box 3.4).

550

Box 3.4: Shifting Social Norms

551

The ability of policies to expand opportunities can be hindered if deeply rooted social norms, such as those related to racial or gender discrimination—are not taken into account. De jure reforms are often at peril of not being implemented if they are at conflict with prevailing norms, including customary law. Legal reforms to improve the rights and opportunities of women, for instance, often fail if norms that sustain existing asymmetries in bargaining power remain unchanged (Milazzo, 2016). Land titling programs looking to improve the access of women to land, for example, can be made ineffective if women are afraid to claim their titles for fear of social sanctions and backlash from their relatives and community (World Bank/FAO/IFAD 2009; Giovarelli et al., 2005).

552

Consider the case of women representation in the political arena. Low representation of women in national parliaments is associated with entrenched beliefs about the ability of women to perform effectively in the political sphere. This lack of female representation, in turn has been linked with unfavorable effects on corruption and the introduction of inclusive policies. While the law cannot change norms in themselves, it can provide incentives to embrace a new law, as well as support the process of internalizing new norms. Enforcement can help jumpstart this process. In 1993, a constitutional amendment mandating gender quotas was implemented in village councils in India. The attitudes of voters towards women were negative during the first term. Yet, two terms of repeated exposure later, the perception of men about the ability of women had improved significantly (Beaman et al., 2009). The aspiration for education of adolescent girls and their parents increased (Beaman et al., 2012), while female entrepreneurship in the manufacturing sector rose (Ghani et al., 2014). The amendment, which mandated the reservation of one-third of local government council positions to women, also led to reduced incidence of corruption (Beaman et al., 2012).

553

The adoption of certain norms is more elastic to a higher level of development. Some norms are more persistent, including those based on religious or philosophical principles. Others are faster to respond to change. The adoption and effectiveness of regulations on child labor, for example, has been found to be associated with income levels. To the degree that households are less dependent on the income of children, the effectiveness of formal regulation rises (Basu, 1999). Sometimes, however, regulations backfire. Bharadwaj and Lkdawala (2013) find that child labor regulations in India led to a reduction in child wages, and to a shift to increased child labor among poor families. 

554

Source: Prepared for this chapter based on World Development Report 2017 Governance and the Law (World Bank, 2017).

555

Poor people tend to have lower access to infrastructure and are more likely to experience the impact of environmental degradation (World Bank, 2001; Hallegate et al., 2016). Thus, efforts to improve living conditions at the local level, from investment in water and sanitation to neighborhood improvement programs and environmental clean-ups can particularly benefit the poor. Improved housing and transportation or the provision of clean water and services, among other services, can have a direct effect on the quality of life, health, and productive opportunities of households. In addition, to the degree that investment and services increase the value of property and land—given deeds and land titling—they have the potential to increase poor households’ collateral and access to credit. The costs of these investments can be considered as direct transfers to the poor (since recovery costs are only taken into account in the operation of the services).

556

Policies that improve access to market opportunities for the poor encompass a wide range of areas. These go from the provision of infrastructure, such as in the form of roads to connect remote and underserved regions, transportation, or programs to reduce crime and violence—ensuring a safe commute to work—to enhancing the access of poor households to technology, or land titling and other means to facilitate their ability to use assets as collateral. They also include addressing market failures more directly, such as in the credit market; and constraining discriminatory practices such as in the judiciary system, or the labor market. Overall, improving access to markets, as other strategies to tackle inequality, implies increasing the bargaining power of the poor. 

557

Aggregate shocks such as economic crises, natural disasters, widespread epidemics, unemployment, illness, and death of breadwinner do more than make the poor transitorily poorer. They can also generate poverty traps and dampen growth. In order to survive, the poor may pare their productive and human capital or simply stop investing in these resources. Coping with risks and adverse shocks requires a mix of measures to deal with economy-wide or region-wide risks and to help poor people address individual adverse shocks. Countries must have the right kind of mechanisms to cushion the damage that such crises can have on the human capital development of the poor. These mechanisms can include arrangements to protect pro-poor public spending when austerity policies need to be put into practice. Rather than improvising, countries need social safety nets such as temporary or emergency employment programs, early childhood programs and cash transfer programs that can be quickly deployed. Micro-insurance programs can complement microcredit programs. Public work schemes can expand in response to local or national shocks. Food transfer programs and social funds to help finance projects identified by communities can also be effective in coping with disaster.

558

Gender inequality, as a persistent form of norms and values-based social inequities, demands specific attention. Even though gender gaps have been narrowing across many domains, unequal gender relations remain pervasive around the world. In nearly all countries, most women encounter disadvantages related to the control of material resources—such as land tenure, which in developing countries often remains conferred to men—and frequently face more insecurity than men. The limited autonomy of women can have important negative effects on children’s education and health. Thus, in addition to normative concerns, improving gender equity can have instrumental benefits for poverty reduction. Progress has been achieved, such as in education or health, but critical gaps remain. Approaches that integrate different aspects—legal, political, and direct public action—tend to be effective. Evidence shows that increasing productive opportunities such as through the provision of farming inputs and microfinance can lead to higher output, more autonomy, and improved nutrition in children. 

559

Social structures are behind many of the dynamics that generate and aggravate poverty—or, alternatively, that can help reduce it. Exclusionary social structures, such as class stratification or gender divisions limit upward mobility. There are some mechanisms that can help shift these dynamics. Policies that support the participation of socially excluded groups and individuals—including by fostering debate around stigma and exclusionary practices—can help remove these obstacles. Affirmative action policies can help groups that face systematic discrimination. Forums, both formal and informal, that bring groups together, can be helpful to mitigate social fragmentation, and to guide dissatisfaction into the political process—and away from social conflict. Finally, tackling biases—such as ethnic, racial or gender—in legislation and in how the legal system operates is central to address inequality; as is promoting the representation of previously excluded groups in community and national organizations.

560

Social capital, rooted in social norms and networks, is an important asset for individuals to exit poverty. It can be promoted by backing existing networks of poor people and providing links to broader markets, public institutions and organizations. To boost the potential of these networks it is necessary to improve the institutional frameworks—legal or regulatory—where the groups that represent poor households operate. Additionally, given that poorer households tend to organize locally, it is important to enhance their ability to advocate for policies at higher levels (state, region, national). This can be achieved by connecting local organizations to broader ones.

561

Regarding fiscal policy, governments should design their tax and transfers system so that the after taxes and transfers incomes (or consumption) of the poor are not lower than their incomes (or consumption) before fiscal interventions. At present, evidence shows that although the existing combination of taxes and transfers is equalizing (albeit to various degrees), what the poor pay in taxes (direct and, especially, consumption taxes) surpasses what they receive in cash transfers especially in low and low-middle income countries (Lustig, 2016; and Lustig, forthcoming). In addition, governments’ spending on education and health should strive for adequate coverage and quality of basic services for the poor.

562

To combat exploitative forms of child labor and discrimination, and foster fair practices in the labor markets, governments should adhere to the core labor standards set out in the Declaration on the Fundamental Principles and Rights at Work adopted by the members of the International Labor Organization. They include freedom of association and the right to collective bargaining, elimination of forced labor, effective abolition of child labor, and the elimination of discrimination in employment and occupation.

563

Some of the policies aimed at changing the distribution of assets (such as inheritance laws, land reform, minimum inheritance, anti-discrimination and affirmative action laws) are likely to be met with a lot of controversy and face a series of political obstacles to overcome. Societies need to reach collectively supported decisions on how to allocate the resources for poverty reduction in an efficient and equitable manner so that they reach the poor rather than being absorbed by other sectors.

564

8.1.2 Building a strong and resilient middle-class

565

Middle classes flourish with labor-intensive economic growth. The growth of jobs and labor incomes has two main sources: accumulation of resources, that is, investment; and efficiency, how well resources are put to use, which is largely driven by technological innovation. Private investment can be promoted by reducing risk through stable fiscal and monetary policy, healthy financial systems, and reliable and transparent business environments. It also requires a sound institutional environment, where rules are applied impersonally and systematically, and with incentives in place to tackle corruption, including the forms of corruption associated with vested business interests (such as favored monopolies, special deals and kickbacks) (World Bank, 2017). Public investment—particularly in terms of infrastructure, communication, and labor force skills—complements private investment in creating productive opportunities and enhancing efficiency.

566

Microenterprises and small businesses are frequently more at risk of being subject to bureaucratic badgering and unequal treatment favoring the well-connected. Measures that reduce the sources of market failures can contribute to the effective participation of these enterprises in the market. For example, in the form of financial deepening to promote access to credit, or by reducing the transaction costs of reaching international markets—such as through export fairs, enhanced access to technology, specialized training, and roads. Importantly, institutional reforms—such as lowering restrictions on the informal sector, addressing land tenure or enabling regulation that encourages smaller investments—can contribute to provide a conducive business environment for smaller firms and individuals.

567

International markets offer a huge opportunity for job and income growth— in agriculture, industry, and services. All countries that have had major reductions in income poverty and a vibrant middle class have made use of international trade. But opening to trade and investment can create losers as well as winners, and it will yield substantial benefits only when countries have the infrastructure and institutions to underpin a strong supply response. Thus the opening needs to be well designed, with special attention to country specifics and to institutional and other bottlenecks. The sequencing of policies should encourage job creation and manage job destruction. Explicit policies should offset transitory costs for the group of people who lose as a result of globalization. The opening of the capital account has to be managed with utmost prudence—in step with domestic financial sector development—to reduce the risk of high volatility in capital flows.

568

While globalization and technological progress appear to have reduced between-countries inequality, their impact on within-country inequality has been in many countries un-equalizing. More importantly, these forces have resulted in dislocation and downward mobility of significant portions of the middle classes and the working poor.  In addition, the benefits of globalization and technological progress have been disproportionately reaped (in some cases dramatically so) by those at the top of the social ladder.  Raising the skill level of a country’s labor force will render that country more capable to benefit from globalization and technological progress. The building of harmonious societies, however, requires –with increasing urgency-- the introduction of drastic corrections on the outcomes that markets produce.  There is increasing evidence that markets are not, and will not, generate sufficient employment at living wages for all.  And markets will continue to make the rich richer if left unchecked.  Given this background, now more than ever it is essential for societies to implement (in some cases, re-introduce) boundaries on market outcomes. 

569

The first boundary on market outcomes should be placed on employment.  A key determinant of employment growth is the characteristic of technical change: whether labor-intensive or labor-displacing, and intensive in which type of labor—skilled or low-skilled.  Even if feasible, it may not be desirable to direct technical change away from labor-saving technologies: after all, the latter have allowed mankind to enjoy increasing amounts of leisure over time.  The problem is that if the income derived from the capital used to produce goods and services is not widely distributed, labor-saving technological change may leave vast portions of the population unemployed or underemployed (i.e., working for fewer hours or in lesser paid jobs than the individual is willing to work for).  To tackle this problem, countries should consider the implementation of some form of guaranteed work. If resources permit, “the government should adopt an explicit target for preventing and reducing unemployment and underpin this ambition by offering guaranteed public employment at the minimum wage to those who seek it. … [I] employment, and meeting the qualifications (see below), are guaranteed a position for a minimum number of hours per week (say, thirty-five hours) paid at the minimum wage working for a public body or an approved non-profit-making institution.”[43]

570

A second boundary on market outcomes should be placed on wages. As we know from economic theory, supply and demand do not fully determine the market wage; they only place bounds on the wage, allowing scope for bargaining about the division of the surplus. But the bargaining process is not just a matter for individuals acting isolated from one another.  Here the trade unions and collective bargaining can play a crucial role in helping produce a fairer distribution of the surplus and a more equitable distribution of wages themselves. For reasons which cannot be discussed here, labor unions have been losing their grip over time.  Unionization rates have been on a decline in practically all countries in the world.  Governments should support union-friendly legislation which at the same time does not foster egregious inefficiency and, above all, clientelism and corrupt practices. Particularly important are the rules that govern collective bargaining and resolution of labor disputes. Minimum wages set at levels corresponding to a living wage is also an important instrument in setting a lower bound for the division of the surplus.

571

As argued in Chapter 8, successful welfare states rely on universalistic programs.  However, “universalistic” should not be equated with providing everybody the same lump-sum transfer.  It means that everybody has the same right to the insurance aspect of universal welfare spending.  Thus, if one of the commitments of a welfare state is that all individuals should enjoy a minimum income level, the insurance aspect implies that those whose pre-fiscal income is below that level—and not the entire population—are entitled to receiving a transfer.  Universalism means that every individual –rich or poor—should have access to free or affordable education and healthcare of acceptable quality as well as to the social insurance mechanisms to cope with idiosyncratic shocks such as unemployment, illness, and death of the breadwinner, to name a few. It is important to recognize, however, that universalism is achieved over time.  Today’s most exemplary welfare states did not have universal programs from the start (see for example, Lindert, 2004). Moreover, poor countries cannot be expected to have universal coverage for all their programs on day one because both their resource and institutional capacity are real limitations.  Nonetheless, to be politically sustainable and contribute to equalizing opportunities, the welfare state needs to award adequate social protection to all. It also needs to provide basic education and health services of sufficient quality so that the middle-class does not feel compelled to opt out. Box 3.5 provides an overview of recent trends in social insurance systems in Latin America.

572

Box 3.5: Markets and social insurance: beyond the market

573

Many important markets, especially for risk, do not exist (Stiglitz, 2004, p. 38). The present and future well-being of all individuals is subject to risks such as illness, the multiple restrictions involved in meeting the care needs of children and the frail or those with disabilities, periods of unemployment and underemployment, and the radical decline in (or loss of) income during old age, whose impact varies in both duration and intensity. Although asymmetrical socioeconomic circumstances are a crucial determining factor, they are, to differing degrees, beyond individuals’ control. Insurance makes it possible to address what has been dubbed the “welfare economics of uncertainty”. This is a reference to uncertainties surrounding access to the necessary protection, the length of time protection will be required, the costs involved, and the degree to which personal well-being and income will be affected. By its nature, therefore, this is a demand for services that is often irregular and unpredictable.

574

Pricing systems have a limited application, since they are unable to deal properly with certain risks, with the result that market insurance is restricted in both coverage and amount. For example, the limitations of private health insurance mean that large medical expenses —the very ones it would be most desirable to insure— go uncovered. Compulsory and solidarity-based insurance, by including and retaining people at low risk, makes it possible to operate by a logic different from that of private insurance and achieve stable risk differentiation and operates with a long-term perspective: because guarantees are applied generally and not to subgroups categorized by risk, individuals are not reclassified if their risks increase and coverage is generally established in more generic terms (Arrow, 1963, 2000).

575

Recent trends in Latin America

576

In the context of a favorable economic situation, and owing to labor market and social protection policies introduced by countries in Latin America, there have been some major advances in extending social protection coverage. Unfortunately, some countries lag behind and coverage has stagnated in them. In recent years Latin America has seen a recovery in employment and an increase in formalization. Various factors have driven the increased formalization of the labor market and the expansion of pension and health care coverage: substantial improvements in the quality of employment, the easing of criteria governing entitlement to contributory coverage (more flexible forms of contribution) and the strengthening of the roles of the State in relation to supervision and regulation of the labour market. The formalization of the labour market has been a key strategy to extend social protection and ensure decent and quality employment, thereby strengthening associated contributory mechanisms. Formalization policies, which are closely linked to contributory schemes, can play a crucial role in fighting poverty, vulnerability and inequality (ECLAC, 2015)

577

In terms of equity, the architecture of social protection systems determines their degree of fragmentation, segmentation and stratification, which does not depend directly on whether their resources are contributory or non-contributory. The specific manner in which they are combined is also crucial, which depends on the design of this architecture, which includes, among other things: the progressive or regressive nature of financing, resource allocation, the definition of benefits, the quality of benefits, how effectively and efficiently resources are used, the regulation of public-private combinations, and the legal protection of entitlements (ECLAC, 2014). In the case of health care, it is crucial to ensure the alignment of benefits, quite beyond the sources of funding that are at stake: the greater the institutional fragmentation, the more complex will be any progress towards such convergence.

578

Some countries in the region have made efforts to deal with the fragmentation of social protection by means of mechanisms designed to bring about forms of convergence and inclusion; for example, non-contributory financing may be linked to contributory financing under similar health-care coverage rules, as is done by the health-care system in Costa Rica. Convergence between the provisions of the contributory and non-contributory health-care regimes is also taking place in Colombia, catalysed by Constitutional Court ruling T-760/08, in a context where the right to health has increasingly been judicialized. In the case of Uruguay, public-sector health-care provision coexists with contribution-financed health-care institutions, but in 2007 the National Integrated Health System (SNIS) was set up to coordinate public- and private-sector providers of integral health care, financed by the National Health Fund, which covers formal workers and their dependents. With regard to pensions, Uruguay has a mixed system comprising an unfunded system and a compulsory individual capitalization system for those with a certain level of income; whereas retirement saving fund management companies (AFAP) manage private pension funds, the vast bulk of old-age benefits are managed by the Social Security Bank, a public-sector body. In Costa Rica, non-contributory and most contributory pensions are administered by the same institution, the Social Security Fund (CCSS).

579

But the dominant trends are different. The goal of extending health coverage by gradually unifying the rules of social protection systems either through social insurance or under the public system has been undermined by inertia and by corporate interests. One good example of the difficulties involved in building up a national health system is the case of Mexico, where the main institutions in charge of health care and health insurance have been the cornerstone of corporate arrangements between the State and politically organized social groups, via the State bureaucracy and unions

580

Most of the advances made in extending coverage remain far from surmounting the fragmentation and segmentation which, to varying degrees and extents, affect social protection systems in the region. Even the Single Health System (SUS) of Brazil, which began as a great universalist project, is having to develop under unequal conditions, as the different subsidies for private health plans, with tax deductibility for private schemes, among other instruments. It also provides public goods such as blood banks at no charge. All this is strongly supported by the middle classes and major union leaderships (Ocké-Reis, 2012). The result is that use of the SUS and private health plans is clearly segmented by income (Sojo, forthcoming).

581

Pensions in Argentina are another emblematic example of institutional fragmentation. Such is the degree of segmentation and fragmentation that it is really impossible to speak of a system: rather, there are a variety of pension programmes that differ in their financing, coverage rules, conditions of access to benefits and type of administration, the outcome not of consistent political decisions, but of a succession of piecemeal solutions (Bertranou et al., 2009).

582

On the other hand, dualist public-private mixes in Latin America, although funded by compulsory contributions and public cross-subsidies, are guided by the profit principle, and insurance is subordinated to this, so that they depart from social security principles and operate on a market insurance logic. Different social protection systems are suffering from a crisis of legitimacy, manifested for example in very high levels of judicialization or high rates of evasion where payment of contributions is concerned. To increase legitimacy, there is a need to improve benefits and reduce inequality gaps. However, this means consolidating redistributive alliances that emphasize convergence of provision, universality for certain levels of protection and specific spheres of solidarity.

583

This underlines the importance of visions of universal coverage that propose a convergence of benefits and to move towards greater alignment and to reducing the stratification of benefits in the different branches of social protection that involve both contributory and non-contributory resources. The strategy for making social protection universal in the region cannot focus exclusively on increasing resources – it must also address their form: institutional change is a crucial part of the locus of innovation (Sojo, 2015). There is a need in Latin America to forge “redistributive alliances”: coalitions, as per Baldwin’s categories (Baldwin, 2003), that are based on the transversality of risk categories, over and above primary income distribution, and that can open up fissures in the economic logic whereby actors are structured (where private-sector pension fund management companies are concerned, for example). Historical experience shows that social policy has gone beyond politically and economically functional minimums when the middle classes have become strongly involved (Baldwin, 2003, p. 9): “Agreements to reapportion risk beyond any functionally minimal level have been possible, therefore, only with the brokering of coalitions of redistributive winners that were politically powerful enough to shift burdens to the losers” (p. 36). Another lesson is that “even without a shift or an increase in risk incidence, formerly unsolidaristic groups have acquired a stake in redistribution through their faltering capacity to carry burdens unaided” (p.15). The feasibility of some reforms towards a greater convergence of benefits or more redistribution could increase, considering that some middle classes and privileged sectors are discontent as some reforms that individualized risks have matured, as is happening in Chile. Thus, middle- and even high-income sectors dissatisfied with the replacement rates being provided by profit-driven contributory pension systems or with health risk selection in likewise profit-driven contributory systems can become significant political subjects (Sojo, 2014).

584

Source: Prepared by Ana Sojo for this chapter.

585

8.1.3 Curbing the excesses of income and wealth concentration at the top

586

Evidence of the high—and in many countries growing—concentration of income and wealth at the top have led to consider boundaries on market outcomes for the rich.  This consideration arises not only because more resources collected from the rich could be used to better support the universalistic welfare state.  They also arise from believing that this excessive concentration is the result of political rent-seeking-- the incomes of the rich do not reflect their contributions to society—and/or that it fosters political and social instability. 

587

There are three main channels that could be used to curb incomes and wealth at the top.  The two most typical involve taxation: increase the progressivity of the tax system and apply high inheritance taxes and taxes on inter-vivos gifts for the rich.  A third measure is more unconventional:  adopt a code of practice for pay.  This is advocated by Atkinson (2015, chapter 5).  This approach can be taken even further through punishing those who do not abide. As an example of the latter, in December 2016, the city of Portland in the USA adopted a tax penalty on corporations that pay their CEOs more than 100 times what they pay typical workers.[44]

588

International actions

589

The policy spheres discussed above apply to actions which are most often taken at the national or subnational levels. In addition, there are a set of initiatives which can be taken in multilateral forums or by advanced nations bilaterally and that affect the evolution of inequality, especially between countries, in the world. International organizations and advanced countries can make a contribution to improving the lot of the poor in the developing world through at least three main channels. First, richer nations can make capital available for the capital-poor countries through grants and long-term loans at concessional rates (e.g., IDA funds at World Bank); reducing current official debt levels (such as the Highly-Indebted Poor Countries initiative); providing financial safety nets in the face of adverse shocks; and, direct bilateral aid. Second, opening markets in advanced countries for agricultural products and promoting freer trade or extending benefits of preferential agreements can help boost developing countries’ exports, improve access to modern technology, and encourage private capital inflows. Imposing non-onerous terms in transfers of intellectual property rights can help poorer countries gain easier access to modern technology. Third, multilateral institutions can assist countries in the design of sound policies, and through their lending program and policy dialogue influence the policies and allocation of resources by individual countries to better target the poor, set appropriate boundaries in market outcomes, and curb excesses of income concentration and wealth accumulation at the top. Through bringing positive pressure to bear upon key policy decisions, international development organizations can tilt the scales in favor of the adoption of pro-poor programs and policies.  The specific roles that aid and trade can play are highlighted in Boxes 3.6 and 3.7. 

590

Box Ch. 3.6: Can aid help reduce inequality?

591

Foreign aid could help mitigate inequality within recipient countries if two critical conditions were met. Donors would have to allocate aid in line with their rhetoric on pro-poor growth, by targeting the most disadvantaged population groups. At the same time, the authorities in the recipient countries would have to ensure that aid actually reaches the poor. Both conditions are likely to be violated at least to some extent. From the literature on aid allocation across recipient countries, it is well known that donors pursue a mix of motives, being motivated partly by developmental concerns and partly by commercial and political self-interest (e.g. Hoeffler and Outram 2011). Commercial donor interests may have as a consequence that foreign aid, e.g. in the area of physical infrastructure, is concentrated in industrial clusters rather than remote areas where the poorest people are living. Likewise, using aid as a means to buy political support by the local elite implies that it favors the rich rather than the poor within a particular country. On the recipient side, aid may be used to provide goods and services that benefit the poor, but it has also been shown to induce rent-seeking (e.g. Reinikka and Svensson 2004). The latter tends to be inequality-increasing as rents are typically captured by local elites endowed with a disproportionate share of a country’s economic and political power (Angeles and Neanidis 2009).

592

Given these counteracting factors, the question of whether foreign aid has on balance reduced within-country inequality is an empirical one. The evidence so far is limited and ambiguous.Herzer and Nunnenkamp (2012) find that foreign aid exerts an inequality-increasing effect on income distribution. According to Chong et al. (2009), there is no robust association between aid and inequality. Shafiullah (2011) as well as Hirano and Otsube (2014) conclude that aid reduces income inequality.

593

The mixed results of these studies may partly be due to differences in country samples and time periods as well as differences in the panel data techniques employed. Yet, the most recent analysis by Hirano and Otsube also points to a more substantive explanation. The authors detect a considerable heterogeneity of the estimated impacts across aid sectors. Specifically, aid given to the social sector, which increased disproportionately over the period covered by their study, is shown to have the strongest and most robust inequality-reducing effect.

594

This finding is in accordance with a strand of the literature that uses disaggregated aid data, investigating the impact of social sector aid on various outcomes related to the Millennium Development Goals. Dreher et al. (2008) show that aid for education has increased primary school enrollment and primary completion rates, while Mishra and Newhouse (2009) show that aid for health has lowered infant mortality. Two recent studies for Uganda and Malawi (Odokonyero et al. 2015; De and Becker 2015), which are based on geocoded data at the sub-national level and thereby mitigate some of the methodological problems that arise in the cross-country setting, corroborate the previous findings.

595

Overall, the incentive problems of both donors and recipients notwithstanding, the part of foreign aid dedicated to the social sector appears to be effective in improving social indicators that matter for the poorest segments of the populations in recipient countries. This is even though the targeting of social sector aid towards primary services – while having improved – still leaves much to be desired. Further improvements in targeting may be seen as a realistic next step towards increasing the inequality-reducing potential of foreign aid.

596

Box Ch.3.7: Trade, Trade Policy, and Inequality

597

International trade theory suggests several channels through which openness to international trade would affect within-country inequality. The most well-known channel is the Stolper-Samuelson mechanism derived from the Hecksher-Ohlin trade model. This channel can however not explain the rise in wage inequality between skilled and unskilled workers in countries such as the United States that coincided with trade reforms implemented in many developing countries during the 1980s and 1990s (Pavcnik 2011). While the Stolper-Samuelson mechanism suggests that increased relative demand for skilled labor in skilled-labor-abundant countries occurs as a result of shifts in relative demand for skilled labor across industries, what mainly happened in the United States and other developed countries was an increase in relative demand for skilled labor within industries.

598

In addition, various studies have documented that, contrary to the predictions of the simple Hecksher-Ohlin model, many developing countries that liberalized their trade in the 1980s and 1990s also observed an increase, rather than a decrease, in wage inequality between skill groups (e.g. Goldberg and Pavcnik 2007). By contrast, the earlier experience of East Asia, (Winters et al. 2004).

599

According to the consensus view that has emerged, skill-biased technical change was the main driving force of rising wage inequality in both developed and developing countries (Pavcnik 2011). Yet, recent research has also pointed to channels other than the Stolper-Samuelson mechanism through which trade could have contributed to wage inequality. One potential channel relates to outsourcing and the concomitant rise in traded intermediate inputs. If firms relocate unskilled-labor-intensive parts of their production to unskilled-labor-abundant countries, the skill intensity of production and thus the relative demand for skilled labor is likely to rise in both developed and developing countries. This channel has for example been shown to be empirically relevant for the case of outsourcing from Hong Kong to mainland China and from the United States to Mexico (Hsieh and Woo 2005; Feenstra and Hanson 1997).

600

Another potentially important channel operates through the differential effects of trade on the wages of workers across heterogeneous firms within industries. If more productive firms select into exporting and expand their activities whereas less productive import-competing firms contract, this tends to raise demand for skilled labor and widen wage disparities. Various empirical studies (e.g. Amiti and Davis 2012; Verhoogen 2008) provide empirical support for the relevance of this channel.

601

Finally, some developing countries such as Columbia and Mexico tended to protect industries employing unskilled workers prior to liberalization (Goldberg and Pavcnik 2007), which provides a further explanation for rising wage inequality but also points to the broader conclusion that the effects of trade on inequality depend on country-specific factors such as the structure of protection and the type of trade a country is engaged in.

602

As concerns the distributional impact of existing and planned trade agreements between developed and developing countries, empirical evidence is scarce and inconclusive.Hanson (2003) finds that Mexico’s membership in NAFTA led to overall wage gains that were largest for more-educated workers living close to the United States and smallest for less-educated workers living in the country’s south. By contrast, ex-ante evaluations of the Trans-Pacific Partnership (TPP) point to slightly higher potential gains for unskilled workers as compared to skilled workers in the participating developing countries (Petri and Plummer 2016).

603

Overall, irrespective of the specific channel, developed countries opening up to foreign trade with less-developed countries have to be prepared for rising wage inequality but should have the means to compensate the losers. For developing countries, trade liberalization is not the silver bullet for achieving growth with equity a naïve application of the Heckscher-Ohlin model might suggest.

604

Governance, policy-making and implementation, the political economy of policies[45]

605

As discussed, policies that contribute to the reduction of inequality have the potential to increase productivity and boost growth by redistributing productive opportunities. Moreover, by reinforcing voluntary compliance, these policies can contribute to strengthening the social contract. Certainly, beyond normative concerns, there are substantial instrumental reasons to care about equity. Yet, governments often fail to adopt pro-equity policies; and, when they do, these policies often fail to reach the desired goal. Why? An important part of the answer has to with the fact that even though these policies have the potential to increase efficiency, they may affect certain groups unfavorably, particularly over the short term.

606

As emphasized in WDR (2017), policymaking does not take place in a vacuum. Instead, it occurs in a process where individuals and groups—who can have very different amounts of power—interact within changing rules as they pursue what can be very different interests (what the Report calls the process of governance). Actors who might be negatively affected by pro-equity policies—such as in terms of rents, income or their ability to maintain influence—may try to block their adoption or implementation. For example, potentially affected groups may attempt to undermine policies that seek land reform. Or public officials may undermine administrative reforms that seek to improve public services, if it cuts into their discretionary control over resources. 

607

Who is able to bargain in the policy decision process is fundamental for the adoption and implementation of policies. The entry barriers and the distribution of power among actors—from policy makers, bureaucrats, civil society groups, and the private sector to individual citizens—determine who gets to participate in the policy arena, and whose voice is more heard. The bargaining power that actors have emanates from an assortment of sources such as social norms, formal rules, control over resources, or the ability to mobilize others. In unequal societies, the capacity of different actors to influence decision-making tends to uneven as well, feeding back into inequality.

608

A common manifestation of power asymmetries is clientelism—where benefits are exchanged in return for political support—which can seriously undermine the effectiveness of pro-equity policies. There are two main forms in which clientelism affects the selection and operation of policies (see Figure 3.7). In the first one, clientelism distorts the relationship between public officials and voters, such that, rather than a dynamic where public officials are the agents of voters (who, in turn, sanction agents), public officials look to “buy­” votes from citizens in exchange for (usually) short-term benefits. That the poor and disadvantaged tend to have higher discount rate for the present, makes them more vulnerable to these short-term bargains. A second form of clientelism takes place when public officials become subject to the interests of influential groups, in exchange for political support. Politicians may then favor the interest of public service providers—such as energy or telecom suppliers, or teacher unions—over those of consumers or students. This can have important costs for society. In order to exert influence, providers may withhold services such as through absenteeism, or they might engage in corrupt practices, or extract rents, diverting public resources. These distortions can impair the provision of basic services, with serious impacts on equity and development.

609

Figure 3.10:  The politician can become an agent of the provider in clientelist settings

         Classic case                                       Clientelism 1                            Clientelism 2

Source: World Development Report 2017: Governance and the Law (World Bank, 2017).

610

Another cornerstone for the effectiveness of pro-equity polities refers to the importance of cooperation, particularly the willingness of individuals to contribute to the funding of public goods. In the absence of credible systems of rewards and penalties, citizens may have incentives to behave opportunistically, enjoying public services without paying taxes. The willingness, or compliance, to continue funding public services can be weakened by the perception that others are free-riding. Other factors can debilitate compliance. Actors that are excluded from the benefits of policies tend to be unwilling to contribute to the provision of public goods. Or, if individuals believe that they are being cheated, for instance, in the form of low quality public services, compliance can further debilitate. Lack of compliance tends to deteriorate the social contract. Conversely, solving cooperation problems—including that all actors involved, from citizens to the providers of services, sustain their end of the deal—can contribute to the effectiveness of policies.   

611

The “optimal” policies on paper that do not acknowledge political equilibria can be inefficient to adopt and implement, if there are powerful actors who gain from the status quo. In those cases, “second best”—but feasible—policies may be preferable. Indeed, successful reforms often go beyond best practices, addressing incentives, and adapting, in order to solve specific problems that are in the way of development.

612

Appendix

613

A1. Tables on Measurement Challenges

614

Table 3.12: Change in Inequality 1985-1995 to 2000-2010

615
 

IMF Fiscal

PovcalNet

SWIID 5.0

Country

Monitor

First

Last

Average

First

Last

Average

Côte d'Ivoire

5.0

7.2

4.8

6.5

5.6

5.0

6.5

Ghana

2.4

7.4

4.6

6.3

3.0

5.3

4.7

Kenya

-6.2

-9.8

5.6

-2.1

-7.1

-6.3

-6.9

Madagascar

-1.0

1.4

-2.0

0.2

3.8

4.4

4.8

Niger

-6.2

4.3

-10.4

0.4

7.1

-13.0

-0.6

Senegal

-7.8

-12.9

-2.3

-7.6

-12.7

-2.3

-7.0

Tanzania

-3.1

0.8

3.8

2.3

-5.0

-1.0

-1.0

Zambia

-13.5

-10.5

2.0

-3.5

-6.0

0.4

-2.6

616

Source: Ferreira, Lustig and Teles (2015). Table 5.

617

NOTE:  Columns labeled “First” use the first available datapoint in the time periods 1985 – 1995 and 2000 – 2010, respectively.  Columns labeled “Last” use the first available datapoint in the time periods 1985 – 1995 and 2000 – 2010, respectively. Columns labeled “Average” use the average across all available years in each time period.

618

Table 3.13: Gini Coefficient Frequencies in Microdata-based Datasets

619

(CEPAL, LIS, SEDLAC, OECD IDD, and WDI/ PovcalNet)

620

Region

Number of Country-Years with Primary Source Data

Total Number Primary Source Observations

Earliest Observation

Most Recent Observation

East Asia and Pacific

112

114

1981

2011

Eastern Europe and Central Asia

301

334

1984

2011

Latin America and Caribbean

378

832

1974

2013

Middle East and North Africa

51

53

1979

2010

South Asia

39

39

1978

2012

Sub-Saharan Africa

139

139

1980

2011

Western Europe and North America

316

403

1967

2010

Grand Total

1336

1914

1967

2013

621

NOTE: Statistics as of December 2014.  WDI Ginis are derived from PovcalNet.

622

Source: Ferreira, Lustig and Teles (2015). Table 1.

623

Table 3.14: Description of Datasets

624

Panel A: Microdata-based Datasets

625

Dataset

CEPAL

LIS

OECD IDD

SEDLAC

WDI PovcalNet

Dataset Summary 

Inequality Indicators (Gini (G), Theil (T), Atkinson (A), Others (O))

G,T,A,O

G,T,A,O

G,O

G,T,A,O

G,T,O

Statistical Significance Indicators (i.e., standard errors or confidence intervals) (Always (A), Sometimes (S), Never (N))

N

N

S

A

N

What is the unit of time: month (M), quarter (Q), year (A)? Are corrections made for inflation?  (Yes(Y)/No(N))

M(N)

A(Y)

A(Y)

M(Y)

Varies (Y)

Description of Welfare Concept

       

Income (I) or consumption (C )

I

I

I

I

varies

Monetary (M) or total (T)? If ‘total’, does it include auto-consumption (Yes(Y)/No(N)) , imputed rent (Yes(Y)/No(N))?

T(Y,Y)

T(Y,N)

M

T(Y,Y)

V-NS

Includes estimates before taxes and transfers?  (Yes(Y)/No(N))

NS

N

Y

N

NS

Includes estimates after taxes and transfers?  (Yes(Y)/No(N))

NS

Y

Y

Y

NS

Equivalization of household incomes: per capita (PC), or alternative equivalence scale (E)?

PC

E

E

PC & E

PC

Are differences in prices by region (rural, urban, etc.) accounted for?(Yes(Y)/No(N))

NS

N

N

Y

V-NS

Adjustments to the original data source ( for harmonization purposes) 

       

Correction for under-reporting (Yes(Y)/No(N))

Y

N

varies

N

N

Is documentation sufficient to replicate results? (Yes(Y)/No(N))

N

N/A

N

N/A

N/A

Adjustment for top coding?  (Yes(Y)/No(N))

N

Y

N

N

NS

  Is documentation sufficient to replicate results? (Yes(Y)/No(N))

N/A

Y

N/A

N/A

N

Elimination of extreme values  (Yes(Y)/No(N))

N

N

V-NS

N

NS

Is microdata made available through the dataset provider? (Yes(Y)/No(N))

N

Y

N

N

N

626

* Consumption is the preferred welfare definition for WDI.  Inequality induces based on income are also listed in certain cases.

627

NOTES:  N/A denotes Not Applicable.  NS denotes Not Stated; in these cases the information is not documented by the dataset provider. V-NS denotes Varies by observation, not stated at observation level. SOURCES: CEPAL: CEPALSTAT (2013a, 2013b, 2013c); LIS: LIS (2013a, 2013b, 2013c); IDD: OCED (2013a, 2013b, 2014a, 2014b); SEDLAC: Socio-Economic Database for Latin America and the Caribbean (2013a, 2013b); WDI:  World Bank (2013c, 2014

628

Panel B: Secondary Source Datasets

629

Dataset

ATG

WIID

Dataset Summary 

Inequality Indicators (Gini (G), Theil (T), Atkinson (A), Others (O))

G

G, O

Statistical Significance Indicators (i., standard errors or confidence intervals) (Always (A), Sometimes (S), Never (N))

N

N

Methodology for Aggregating Data

 

Adjusts primary source data?(Yes(Y)/No(N))

N

N

Is original source of data clearly noted?(Yes(Y)/No(N))

Y

Y

Are welfare concepts clearly noted? (Yes(Y)/No(N))

N

Y

If multiple datapoints are available for the same country and year, are some sources of data given priority?(Yes(Y)/No(N))

Y

Y

If multiple datapoints are available for the same country and year, is a "first-best" datapoint selected? (Yes(Y)/No(N))

Y

N

Are secondary-source databases used as inputs? (Yes(Y)/No(N))

Y

N

If secondary-source based databases are used as inputs, is a methodology in place to prevent over-representation if the same datapoint appears from both a primary and secondary source? (Yes(Y)/No(N)/Not Stated (NS))

NS

N/A

630

Panel C: Imputation-based Datasets

631

Dataset

SWIID

Dataset Summary 

Inequality Indicators (Gini (G), Theil (T), Atkinson (A), Others (O))

G

Statistical Significance Indicators (i., standard errors or confidence intervals) (Always (A), Sometimes (S), Never (N))

A*

Methodology for Imputations

Is the description of imputation methods sufficient to replicate the dataset? (Yes(Y)/No(N))

Y

Has methodology been subject to scrutiny by experts in the field of imputation?  (Yes(Y)/No(N))

Not Clear

Are secondary data sources used as inputs? (Yes(Y)/No(N))

Y

If secondary data sources are used as inputs, is a methodology in place to prevent over-representation if the same datapoint appears from both a primary and secondary source? (Yes(Y)/No(N)/Not Stated (NS))

NS

Is there a systematic validation process in place with country/region experts? (Yes(Y)/No(N))

Not Clear

If multiple datapoints are available for the same country and year, is a "first-best" datapoint selected or are all datapoints used in the imputation method? (First Best (FB)/ All)

All

 Is the welfare concept clearly defined?   (Yes(Y)/No(N)) Can it be deduced from the documentation on the primary sources?  (Yes(Y)/No(N))

N, Y

632

Source: Ferreira, Lustig and Teles (2015). Table 2.

633

*SWIID v5.0 provides 100 imputed values of each of its indicators.  Stata code on the SWIID website allows users to either calculate means, standard deviation, and confidence intervals or to use the 100 values to multiply impute their analysis.   Earlier versions of the SWIID provided the mean and standard deviation.

634

NOTES:  N/A denotes Not Applicable.  NS denotes Not Stated; in these cases the information is not documented by the dataset provider. V-NS denotes Varies by observation, not stated at observation level. 

635

SOURCES: ATG: Milanovic (2013); WIID: World Income Inequality Database (2014); SWIID: Solt (2009, 2014).

636

Table 3.15: Sources Used by Secondary Source Datasets: All the Ginis, SWIID, and WIID

637

Sources Used

Secondary Datasets

All the Ginis

The Standardized World Income Inequality Database (SWIID)

World Income Inequality Database (WIID)

Group 1: Datasets that Calculate Indices with Microdata

CEPALSTAT

 

X

 

Luxembourg Income Study (LIS)

X

X

X

OECD IDD

     

Socio-Economic Database for Latin America and The Caribbean (SEDLAC )

X

X

X

World Development Indicators(WDI) and PovcalNet

X

X

X

Group 2: Datasets that use Secondary Sources

All the Ginis

 

X

 

The Standardized World Income Inequality Database (SWIID)

     

World Income Inequality Database (WIID)

X

X

 
638

Source: Ferreira, Lustig and Teles (2015). Table 3.

639

A2. Alternative estimates of inequality trends using Global Consumption and Income Project

640

Table 3.16: Using Global Consumption and Income Project

641
           

OECD

European

Asian

Latin

MENA

South

South

Sub-

World

 

Transition

Transition Economies

America

East

Asia

Saharan

 

Economies

Asia

Africa

 

1980s (or earlier available year) and 1990s

 

Specific period for

1980-2001

1990-1998

1980-2000

1980-2002

1980-2000

1980-1995

1980-2000

1980-1995

   

Each region/3

 

 Rising inequality

13

23

2

7

3

3

5

10

66

62%

 No change

0

1

1

3

0

0

0

6

11

10%

 Falling inequality

8

0

0

8

5

4

0

5

30

28%

Total

21

24

3

18

8

7

5

21

           107

 

2000-2010 (or similar period)

 

 Specific period for each region /3

2000-

1998-2010

   2000 –  2009

2002-

2000-2007

1995-2009

2000-

1995-2007

    

 

2010

2010

2010

 

Rising inequality

9

11

3

3

2

4

0

7

39

36%

No change

6

2

0

0

3

0

1

4

16

15%

Falling inequality

6

11

0

15

3

3

4

10

52

49%

Total

21

24

3

18

8

7

5

21

107

 
                       
642

Note: These data are based on income.Consumption surveys have been adjusted to reflect that they tend to show lower inequality.Shaded cells indicate that the numbers are substantially different to Table 1 in the text.

643

Source: GCIP

644

Table 3.17: Average Regional Inequality 2000-2010 using CGIP

645
       

Gini Coefficienta

Regionb

2000

2005

2010

World (only countries which Lustig lists for regions)

0.452

0.453

0.448

World (About 155 countries for which GCID has data)

0.476

0.477

0.47

Advanced Economies………… 

0.319

0.324

0.324

East Asia and the Pacific……… 

0.507

0.515

0.519

Eastern Europe and Central Asia

0.406

0.429

0.425

Latin America and the Caribbean

0.528

0.517

0.49

Middle East and North Africa

South Asia… 

0.454

0.425

0.418

Sub-Saharan Africa…………..    

0.606

0.59

0.597

Income Categoryc (Using only countries listed by Lustig)

Low Income Countries………  

0.606

0.58

0.58

   Lower Middle Income Countries…   

0.493

0.494

0.481

   Upper Middle Income Countries…   

0.467

0.483

0.479

Total Middle Income Countries 

0.431

0.423

0.413

High Income ountries……….

0.334

0.34

0.34

         

* Countries included in various regions and income levels are the only the ones which Lustig has listed. The coverage is uneven by region and income levels, with good coverage of LA but fewer countries in SSA, SA.

Note: Levels of inequality are reported to be much higher in this Table than in Table 2 above due to the fact that an adjustment has been applied to consumption surveys (which are prevalent in developing regions except Latin America) to reflect that tend to show substantially lower inequality.

646

Figure 3.11: Gini Level and Change

647

Source: GCIP.

648

References

649

Aaberge, R., Mogstad, M. and Peragine, V. 2011. Measuring long-term inequality of opportunity. Journal of Public Economics 95(3), pp 193-204.

650

Aaberge, R. and A. Brandolini. 2014. “Multidimensional poverty and inequality,” Bank of Italy Discussion Paper. No. 976.

651

Acemoglu, D. 1998. “Why do new technologies complement skills? Directed technical

652

change and wage inequality”. Quarterly journal of economics, 1055-1089.

653

Acemoglu, D. and J. Robinson. 2009. Economic Origins of Dictatorship and Democracy, Cambridge: Cambridge University Press.

654

Acemoglu, D., and Autor, D.H. 2011. “Skills, Tasks and Technologies: Implications for

655

Employment and Earnings” In: D. Card, and O. Ashenfelter (Eds.), Handbook of Labor

656

Economics, Elsevier, New York, 1043-1171.

657

Acs, G. 2011. “Downward Mobility from the Middle Class: Waking Up from the American Dream,” Pew Charitable Trusts, http://www.urban.org/sites/default/files/alfresco/publication-pdfs/1001603-Downward-Mobility-from-the-Middle-Class-Waking-Up-from-the-American-Dream.PDF. Accessed on June 22, 2016.

658

Adams, R. H. (2004). Economic growth, inequality and poverty: estimating the growth elasticity of poverty. World Development, 32(12), 1989-2004.

659

Adam, K. and P. Tzamourani (2015) Distributional Consequences of Asset Price Inflation in the Euro Area, WP, University of Mannheim.

660

African Population Policy Research Center. 1998. Fertility Decline in Kenya: Levels, Trends, Differentials.  Nairobi: APPRC.

661

Aghion, P., and Bolton, P. (1997). A theory of trickle-down growth and development. The Review of Economic Studies64(2), pp.151-172.

662

Agnello, Luca, Sushanta K. Mallick, and Ricardo M. Sousa. "Financial reforms and income inequality." Economics Letters 116.3 (2012): 583-587.

663

Agrawal, A. 2010. Health Outcomes in India: Distribution, Determinants and Policy Options, Unpublished Ph.D. Dissertation, Mumbai: Indira Gandhi Institute of Development Research.

664

Ahituv A. (2001). Be fruitful or multiply: On the interplay between fertility and economic development. Journal of Population Economics, 14, pp.51-71.

665

Akerlof, G.A. and R.E. Kranton. 2000.  “Economics and Identity,” The Quarterly Journal of Economics 115: 715-53.

666

Alesina, A., and Rodrik, D. (1994), “Distributive Politics and Economic Growth,” Quarterly Journal of Economics, 109(2), 465–490.

667

Alesina, A. and E. Glaeser. 2004. Fighting poverty in the US and Europe: A world of difference.  Oxford: Oxford University Press.

668

Alvaredo, Facundo and Thomas Piketty. (2010). “The Dynamics of Income Concentration

669

in Developed and Developing Countries: A View from the Top.” Chapter 4 in Lopez-Calva,

670

Luis F. and Nora Lustig, editors, Declining Inequality in Latin America: A Decade of

671

Progress? Brookings Institution Press and UNDP.

672

Alvaredo, F., A. Atkinson, T. Piketty, and E. Saez. 2013. 'The top 1% in international and historical perspective” Journal of Economic Perspectives 27(3): 3-20.

673

Alvaredo, F., N. Lustig, and T. Piketty. 2016. “Income and Wealth Inequality.” Chapter in report by High Level Expert Group on Measuring Economic Performance and Social Progress.

674

Amiti, M., and D. R. Davis (2012). Trade, Firms, and Wages: Theory and Evidence. Review of Economic Studies 79(1): 1–36.

675

Anand, S., & Segal, P. (2008). What do we know about global income inequality?. Journal of Economic Literature46(1), 57-94.

676

Anand, S., & Segal, P. (2014). The global distribution of income. Atkinson, A. B., & Bourguignon, F. (Eds.). Handbook of Income Distribution SET vols. 2A-2B. Elsevier.

677

Anderson, E., (1999), “What is the Point of Equality?”, Vol. 109, No. 2.

678

André, C., and Platteau, J-P., (1998), “Land Relations under Unbearable Stress: Rwanda Caught in the Malthusian Trap,” Journal of Economic Behaviour and Organisation, Vol. 34, pp. 1–47.

679

Andrews, D., Jencks, C., and Leigh, A. (2011). Do rising top incomes lift all boats? B.E. Journal of Economic Analysis and Policy, 11, Contributions Article 6.

680

Angeles, L., and K.C. Neanidis (2009). Aid Effectiveness: The Role of the Local Elite. Journal of Development Economics 90: 120-134.

681

Anyanwu, J.C, A.E. O. Erhijakpor and E.Obi_(2016), ‘Empirical Analysis of the Key Drivers

682

of Income Inequality in West Africa’, African Development Review, vol 28, No.1, 2016, 18

683

38.

684

Araujo, M. Caridad, Mariano Bosch, and Norbert Schady. 2016. “Can Cash Transfers Help Households Escape an Inter-Generational Poverty Trap?” NBER Working Paper 22670, National Bureau of Economic Research, Cambridge, MA.

685

Arocena, R. and J. Sutz (2010) ‘Weak knowledge demand in the South: learning divides and innovation policies’, Science and Public Policy, 37 (8) 571–582.

686

Aronson, J., Cohen, G., & Nail, P. R. (1999). Self-affirmation theory: An update and

687

appraisal. In J. Mills, & E. Harmon-Jones (1999), Cognitive dissonance: Progress on a

688

Pivotal Theory in Social Psychology. Washington, DC: American Psychological

689

Association.

690

Arrow, Kenneth. 1963. “Uncertainty and the welfare economics of medical care”, The American Economic Review, Vol. LIII, No. 5, December.

691

_______. 2000. “Insurance, risk and resource allocation”, Foundations of Insurance Economics. Readings in Economics and Finance, G. Dionne and S. E. Harrington (comps.), Boston, Kluwer Academic Publishers.

692

Asian Development Bank. 1997.  Emerging Asia.  Manila: ADB.

693

Atkinson, A. (1970) On the measurement of inequality, Journal of Economic Theory, 2(3), pp. 244–263.

694

Atkinson, A. 1995. “On Targeting Social Security: Theory and Western Experience with Family Benefits,” in D. van de Walle and K. Nead (eds.), Public Spending and the Poor: Theory and Evidence.

695

Atkinson, A. (1997) Bringing income distribution in from the cold, Economic Journal, 107, pp. 297–321.

696

Atkinson, A. B., & Brandolini, A. (2001). Promise and pitfalls in the use of" secondary" data-sets: Income inequality in OECD countries as a case study.Journal of economic literature39(3), 771-799.

697

Atkinson, A.B. (2008). The Changing Distribution of Earnings in OECD Countries. Oxford University Press, New York.

698

Atkinson, Anthony, B. 2015. Inequality: What can be done? Harvard University Press, London.

699

Atkinson, A. and A. Brandolini. 2010. “On Analyzing World Distribution of Income,” World Bank Economic Review 24 (1): 1-37.

700

Atkinson, G. (1995) Efficiency versus equity? A false dichotomy?, in: C.M.A. Clark (Ed.) Institutional Economics and the Theory of Social Value: Essays in Honor of Marc R. Tool (Amsterdam, Kluwer).

701

Attanasio, Orazio P., and Miguel Székely. 1999. “An Asset-Based Approach to the Analysis of Poverty in Latin America.” Working Paper R-376, Inter-American Development Bank, Washington, DC.

702

Attanasio, O., E. Battistin, E. Fitzsimons, A. Mesnard, and M. Vera-Hernandez. 2005. “How Effective Are Conditional Cash Transfers? Evidence from Colombia.” Briefing Note 54, Institute for Fiscal Studies, London.

703

Autor, D., Katz, L., & Kearney, M. (2006). “Measuring and interpreting trends in economic

704

inequality.” In AEA Papers and Proceedings, 96(1), 189-194.

705

Autor, David, H., Kerr, W. R., & Kugler, A. D. (2007). “Does employment protection reduce

706

productivity? Evidence from US states.” The Economic Journal, 117(521), 189-217.

707

Autor, D. H., Katz, L. F., and M.S. Kearney (2008).”Trends in US wage inequality: Revising

708

the revisionists.” The Review of Economics and Statistics, 90 (2), 300-323.

709

Azevedo, Joao Pedro; Favara, Marta; Haddock, Sarah E.; Lopez-Calva, Luis F.; Muller, Miriam; Perova, Elizaveta. 2012. Teenage Pregnancy and Opportunities in Latin America and the Caribbean: On Teenage Fertility Decisions, Poverty and Economic Achievement. World Bank, Washington, DC.

710

Azevedo, Joao Pedro, Gabriela Inchauste, and Vviviane Sanfelice. (2013a). “Decomposing

711

the Recent Inequality Decline in Latin America”, Policy Research Working Paper 6715,

712

The World Bank.

713

_________________, Maria Eugenia Dávalos, Carolina Diaz-Bonilla, Bernardo Atuesta, and

714

Raul Andres Castañeda. (2013b). “Fifteen Years of Inequality in Latin America: How

715

Have Labor Markets Helped?”Policy Research Working Paper 6384, The World Bank.

716

Badaracco, N. (2014). “Fecundidad y cambios distributivos en América Latina.”

717

Working Paper 173, CEDLAS-Universidad Nacional de La Plata, November.

718

Baddeley, A. D., & Hitch, G. (1974). Working memory. The psychology of learning and

719

motivation, 8, 47-89.

720

Baird, S., Francisco H. G., Ferreira, B. Ö., and M. Woolcock. 2013. “Relative Effectiveness of Conditional and Unconditional Cash Transfers for Schooling Outcomes in Developing Countries: A Systematic Review.” Campbell Systematic Reviews 9 (8), Campbell Collaboration, Oslo.

721

Balcázar, C.P., (2015). Lower bounds on inequality of opportunity and measurement error. Economics Letters 137, pp 102-105.

722

Baldwin, Peter (2003), The Politics of Social Solidarity. Class Bases of the European Welfare State 1875- 1975, Cambridge: Cambridge University Press.

723

Bandiera, O. and A. Natraj, 2013. Does gender inequality hinder development and economic growth?  Evidence and policy implications.  World Bank Research Observer 28:2-21.

724

Banerjee, A.V., and Duflo, E. (2003). Inequality and growth: what can the data Say? Journal of Economic Growth, 8, pp. 267-299.

725

Banerjee, A. and E. Duflo (2008). What is middle class about the middle classes around the world.Journal of Economic Perspectives 22: 3-28.

726

Banerjee, Abhijit, Esther Duflo, Raghabendra Chattopadhyay, and Jeremy Shapiro. 2016. “The Long-Term Impacts of a ‘Graduation’ Program: Evidence from West Bengal.” Unpublished paper, Massachusetts Institute of Technology, Cambridge, MA.

727

Bardhan, Ghatak and Karaivanov (2007), “Wealth Inequality and Collective Action”, Journal of Public Economics, Vol. 91, Issue 9.

728

Barham, T., K. Macours, and J. A. Maluccio. 2014. “Schooling, Learning and Earnings: Effects of a Three-Year Conditional Cash Transfer Program in Nicaragua after 10 Years.” Presentation at Impact Evaluation Network–Latin American and Caribbean Economic Association meeting, Washington, DC, September 11.

729

Baron, J. N., & Pfeffer, J. (1994). The social psychology of organizations and

730

inequality. Social Psychology Quarterly, 190-209.

731

Barr, M. S. (2012). No Slack: The Financial Lives of Low-income Americans. Brookings

732

Institution Press.

733

Barro, R. 1991. Economic Growth in a Cross Section of Countries.  Quarterly Journal of Economics.

734

Barro, R. J. (2000). Inequality and Growth in a Panel of Countries. Journal of economic growth5(1), pp.5-32.

735

Barros, R., M. De Carvalho, S. Franco, and R. Mendonca (2010) “Markets, the State and

736

the Dynamics of Inequality in Brazil” Chapter 6 in Lopez-Calva, Luis F. and Nora Lustig,

737

editors, Declining Inequality in Latin America: A Decade of Progress? Brookings

738

Institution Press and UNDP.

739

Basu, Kaushik. 1999. “Child labor: cause, consequence, and cure, with remarks on international labor standards.” Journal of Economic Literature 37 (3): 1083-1119

740

Basu,. K. 2013. “Shared prosperity and the mitigation of poverty: in practice and in precept,” World Bank Policy Research Working Paper No. 6700. Washington D.C. The World Bank. 

741

Bates, R. (1981), Markets and States in Tropical Africa, expanded edition 2014, Berkeley: University of California Press.

742

Beaman L., Chattopadhyay R., Duflo E., Pande R., Topalova P. 2009. “Powerful women: does exposure reduce bias?” The Quarterly Journal of Economics, 124 (4): 1497-1540.

743

Beaman L., Duflo E., Pande R., Topalova P. 2012. Female leadership raises aspirations and educational attainment for girls: A policy experiment in India. Sciencexpress, 12 January 2012.

744

Beck, T., Demirgüc-Kunt, A., and Levine, R. (2004). Finance, Inequality, and Poverty: Cross-Country Evidence. NBER Working Paper 10979.

745

Beck, T., Demirgüc, A., and Levine, R. (2007). "Finance, Inequality and the Poor". Journal

746

of Economic Growth 12(1): 27-49.

747

Beck, T., Levine, R. and Levko, A. (2010), Big Bad Banks? The Winners and Losers from

748

Bank Deregulation in the United States. The Journal of Finance, 65: 1637–1667.

749

Becker, G., S. (1960). An economic analysis of fertility. Demographic and Economic Change in Developed Countries. Princeton University Press, pp.209-231.

750

Behrman, J., N. Birdsall and M. Székely (2000) “Economic Reform, and Wage Differentials

751

in Latin America”, Working Paper of the Research Department  n. 435, Inter-American

752

Development Bank, Washington, D.C.

753

Behrman Jere R., P. Sengupta and P. Todd. 2005. “Progressing through Progresa: An Impact Assessment of a School Subsidy Experiment in Mexico.” Economic Development and Cultural Change 54 (1): 237–75.

754

Benabou, R. 1996. “Inequality and growth,” NBER Macroeconomics Annual 1996 11: 11–92. Boston: MIT Press.

755

Bénabou, R. (2000). Unequal Societies: Income Distribution and the Social Contract. American Economic Review, 90.1: 96-129.

756

Benabou, R. and E. Ok. 2001. “Social mobility and the demand for redistribution: the POUM hypothesis,” The Quarterly Journal of Economics 116: 447-487.

757

Benabou, R. (2002). Tax and Education Policy in a Heterogeneous‐Agent Economy: What Levels of Redistribution Maximize Growth and Efficiency?.Econometrica70(2), pp.481-517.

758

Berg, Ostry and Loungani (forthcoming), Inequality: Causes, Conseqences and Cures, New York: Columbia University Press.

759

Bertola, G. (1993). Factor Shares and Savings in Endogenous Growth. American Economic Review, 83, pp.1184-98.

760

Bertrand, M., Mullainathan, S., & Shafir, E. (2004). A behavioral-economics view of

761

poverty. The American Economic Review94(2), 419-423.

762

Bertrand, M., Mullainathan, S., & Shafir, E. (2006). Behavioral economics and marketing

763

in aid of decision making among the poor. Journal of Public Policy & Marketing25(1), 8

764

23.

765

Bertranou, Fabio, Esteban Calvo y Evelina Bertranou (2009), Is Latin America retreating from individual retirement accounts?, Center for Retirement Research at Boston College, Julio, No. 9-14.

766

Besley, Timothy, and Torsten Persson. 2011. Pillars of Prosperity: The Political Economics of Development Clusters. Princeton, NJ: Princeton University Press.

767

Bhalla, S. (2002): Imagine there's no country: Poverty, inequality, and growth in the era globalization. Peterson Institute

768

Bharadwaj, Prashant & Leah K. Lakdawala. 2013. “Perverse Consequences of Well-Intentioned Regulation: Evidence from India’s Child Labor Ban,” NBER Working Paper No. 19602. Cambridge, MA: National Bureau of Economic Research.

769

Binswanger, H., Deininger, K., and Feder, G., (1995), “Power, Distortions, Revolt and Reform in Agricultural Land Relations,” In Behrman, J., and Srinivasan, T.N., (eds), Handbook of Development Economics, Vol. 3. Elsevier Science B.V., Amsterdam.

770

Birdsall, N. (2006). Rising inequality in the new global economy. International Journal of Development Issues, 5(1), pp. 1 – 9.

771

Bisset, L., & Coussins, J. (1982). Badge of poverty: A new look at the stigma attached to

772

free school meals. London, England: Child Poverty Action Group.

773

Bjorklund, A. and M. Jantti. 2000. “Intergenerational Mobility of Socioeconomic Status in Comparative Perspective,” Nordic Journal of Political Economy 26 (1): 3-32.

774

Björklund, A., Jäntti, M. and Roemer, J.E. (2012). Equality of opportunity and the distribution of long-run income in Sweden. Social Choice and Welfare 39(2-3), pp 675-696.

775

Blacker, J. 2002. Kenya’s Fertility Transition. How Low will it Go?  Center for Population Studies. 

776

Blanchflower, D. & Oswald, A. (2003) Does inequality reduce happiness? Evidence from the states of the USA from the 1970s to the 1990s (unpublished manuscript, Warwick University).

777

Blanden, J., R. Haveman, T. Smeeding and K. Wilson. 2014. “Intergenerational Mobility in the United States and Great Britain: A Comparative Study of Parent–Child Pathways,”Review of Income and Wealth 60 (3): 425-449.

778

Bleaney, NM. F. Underconsumption theories: A history and critical analysis.  New York: International Publishers.

779

Bleaney, M. and A. Nishiyama. 2003. “Convergence in income inequality: differences between advanced and developing countries,” Economics Bulletin 4(22):1–10.

780

Block, S. (2010), “The decline and rise of agricultural productivity in Sub-Saharan Africa

781

since 1961”, Working Paper 16481, NBER, http://www.nber.org/papers/w16481.

782

Bloom, D. and J. Williamson. 1998. Demographic Change and Economic Miracles in Emerging Asia.  World Bank Economic Review.

783

Bordo, M. and C. Meissner (2012).  Does inequality lead to a financial crisis?  Journal of International Money and Finance 31(8): 2147-2161.

784

Boserup, E. 1981.  Population and Technological Change.  Chicago: The University of Chicago Press.

785

Bosmans, K., K. Decancq, and E. Ooghe. 2015. “What do normative indices of multidimensional inequality really measure?,”  Journal of Public Economics 130: 1-120.

786

Bosmans, K., K. Decancq and A. Decoster. 2011. “The evolution of global inequality: absolute, relative and intermediate views,” Discussion Paper, Centre for Economic Studies, Catholic University, Leuven. Available at SSRN: http://ssrn.com/abstract=1806875 or http://dx.doi.org/10.2139/ssrn.1806875.

787

Bound, K. and I. Thornton (2012) Our frugal future. Lessons from India’s Innovation System, London: NESTA.

788

Bourguignon, F. and C. Morisson (1989), “Income Distribution, Development and Foreign Trade”, European Economic Review, 34 (1990).

789

Bourguignon, F., & Morrisson, C. (2002). Inequality among world citizens: 1820–1992. The American Economic Review92(4), 727-744.

790

Bourguignon, F. (2003). The Growth Elasticity of Poverty Reduction: Explaining Heterogeneity across Countries and Time Periods. In: T. Eicher and S. Turnovsky, ed., Inequality and Growth: Theory and Policy Implications, 1st ed. Cambridge, Massachusetts: MIT Press, pp.3-26.

791

Bourguignon, F. (2012): Seuil et La République des Idées.

792

Bourguignon, F. 2015. “Appraising income inequality databases in Latin America,” in F. Ferreira and N. Lustig (eds.), Special Issue, Appraising Cross-National Income Inequality Databases. The Journal of Economic Inequality 13 (4): 557-578.

793

Bourguignon, F. 2015. The Globalization of Inequality.  Princeton: Princeton University Press.

794

S. Bowles, H. Gintis and M. Osbourne-Groves (ed.) 2005. Unequal Chances: Family Background and Economic Success, Princeton, NJ: Princeton University Press.

795

Bowles, S. and A. Jayadev (2006), “Guard Labor”, Journal of Development Economics

796

Boyce, C.J., G.D.A. Brown and S.C. Moore. 2010. “Money and happiness: rank of income, not income, affects life satisfaction,” Psychological Science 21(4): 471-475.

797

Branisa, B. S. Klasen, and M. Ziegler. 2013.  Gender inequality in social institutions and gendered development outcomes. World Development 45: 252–268, 2013

798

Bresson, F. 2009. On the estimation of growth and inequality elasticities of poverty with grouped data.  Review of Income and Wealth 55(2): 266-302

799

Brown, G.K. 2010. “Inequality, ethnicity and violent secessionism: A mixed-level and survival analysis approach to disaggregating the study of civil war,” Centre for Research on Inequality, Human Security and Ethnicity, Oxford. Working Paper No. 80.

800

Bruckner, J. (2011) Lectures on Urban Economics, Cambridge, MA: MIT Press.

801

Bueno de Mesquita, B. (1978), “Systemic Polarization and the Occurrence and Duration of War,” Journal of Conflict Resolution 22/2: 241-267.

802

Buhmann, B., L. Rainwater, G. Schmaus, T.M. Smeeding. 1988. “Equivalence scales, well-being, inequality, and poverty: sensitivity estimates across ten countries using the Luxembourg Income Study (LIS) database,” Review of Income and Wealth 34(2): 115-142.

803

Bullard, R. (1993) T Press.ics4) “Distributjustice in the United States,” Yale Journal of International Law, 18 (1), pp. 319ntern

804

Burger, R., S. Klasen, and A. Zoch. 2016. “Estimating income mobility when income is measured with error: The case of South Africa,”  Economic Research Southern Africa Working Paper No. 607.

805

Burgess R, Pande R. 2005. Can rural banks reduce poverty? Evidence from the Indian

806

Social Banking Experiment. Am. Econ. Rev. 95(3):780–95

807

Burtless, G. (1999). “Effects of Growing Wage Disparities and Changing Family

808

Composition on the U.S. Income Distribution.” European Economic Review, 43 (4-6), 853-

809

865.

810

Burtless, G. (2009). “Demographic Transformation and Economic Inequality.” In: W.

811

Salverda, B. Nolan, and T. M. Smeeding (Eds.), The Oxford Handbook of Economic

812

Inequality, Oxford University Press, Oxford, 435-454.

813

Bussolo, Maurizio, and Luis F. Lopez-Calva. 2014. Shared Prosperity: Paving the Way in Europe and Central Asia. Washington, DC: World Bank.

814

Calderón, C., & Chong, A. (2009). “Labor market institutions and income inequality: an

815

empirical exploration.” Public Choice, 138(1-2), 65-81.

816

Camerer, C., & Thaler, R. H. (1995). Anomalies: Ultimatums, dictators and manners. The

817

Journal of Economic Perspectives9(2), 209-219.

818

Campos, Raymundo, Gerardo Esquivel, and Nora Lustig. (2014). “The Rise and Fall of

819

Income Inequality in Mexico, 1989–2010.” In Cornia, Giovanni Andrea, editor, Falling

820

Inequality in Latin America: Policy Changes and Lessons, UNU-WIDER Studies in

821

Development Economics, Oxford: Oxford University Press.

822

Cancian, M., and D. Reed. 1998. “Assessing the Effects of Wives' Earnings on Family

823

Income Inequality.” Review of Economics and Statistics, 80 (1), 73-79.

824

Cancian, M., and D. Reed. 1999. “The Impact of Wives' Earnings on Income Inequality:

825

Issues and Estimates.” Demography, 36 (2), 173-184.

826

Canning, D, S. Raja, and A. S. Yazbeck (eds).’ Africa’s Demographic Transition: Dividend

827

or Disaster?’ Africa Development Forum. The World Bank, Washington DC.

828

Card, D. & DiNardo J. E. (2002) “

829

inequality: some problems and puzzles.” No. w8769. National Bureau of Economic

830
  • .
831

Cardoso, E., and A. Souza. 2003. “The Impact of Cash Transfers on Child Labor and School Attendance in Brazil.” Instituto de Estudios do Trabalho e Sociedade, Rio de Janeiro.

832

Carlsson, F., Daruvala, D. & Johansson-Stenman, O. (2005) Are people inequality-averse, or just risk-averse? Economica, 72, pp. 375–396.

833

Caskey, J. P. (1994). Fringe banking: Check-cashing outlets, pawnshops, and the poor. New

834

York, NY: Russell Sage Foundation.

835

Castello Climent, A., & Domenech, R. (2014). Capital and Income Inequality: Some Facts

836

and Some Puzzles (Update of WP 12/28 published in October 2012) (No. 1228). BBVA

837

Bank, Economic Research Department.

838

Castells-Quintana, D. and Royuela, V., (2011), “Agglomeration, Inequality and Economic Growth”,IREA-WP series, no. 2011/14.

839

Castells Quintana, D., & Royuela Mora, V. (2014). Tracking positive and negative effects of inequality on long-run growth. IREA–Working Papers, 2014, IR14/001.

840

Chakravarty, S. 2009. Inequality, Polarization and Poverty: Advances in Distributional Analysis. New York: Springer.

841

Chambers, D., Krause, A. (2010). Is the relationship between inequality and growth affected by physical and human capital accumulation? Journal of Economic Inequality, 8(2), pp.153-172.

842

Chattopadhyay, R. and E. Duflo (2004) “Women as Policy Makers: Evidence from a Randomized Policy Experiment in India,” Econometrica, 72 (5), pp. 1409-1443. Moore, B. (1993) Social Origins of Dictatorship and Democracy: Lord and Peasant in the Making of the Modern World, Boston: Beacon Press.

843

Checchi, D. (2001). Education, inequality, and income inequality.  Sticerd Discussion Paper No. 52, London: LSE

844

Checchi, D. and V. Perragine. 2010. “Inequality of Opportunity in Italy,” Journal of Economic Inequality 8 (4): 429-450.

845

Checchi, D., Peragine, V., and Serlenga, L. (2010). Fair and unfair income inequalities in Europe. IZA Discussion Paper No. 5025.

846

Checchi, D., Peragine, V., and Serlenga, L. (2015). Income Inequality and Opportunity Inequality in Europe: Recent Trends and Explaining Factors. Mimeo.

847

Chenery, et al. 1974. Growth and Redistribution, The World Bank.

848

Cherry, T. L., Frykblom, P., & Shogren, J. F. (2002). Hardnose the dictator. The American

849

Economic Review92(4), 1218-1221.

850

Chiu, W. H., and Madden, P. (1998). Burglary and income inequality. Journal of Public Economics, 69(1), pp. 123-141.

851

Chong, A, M. Gradstein, and C. Calderon (2009). Can Foreign Aid Reduce Income Inequality and Poverty? Public Choice 140: 59–84.

852

Chotikapanich, D., Griffiths, W. E., Prasada Rao, D. S., & Valencia, V. (2012). Global income distributions and inequality, 1993 and 2000: Incorporating country-level inequality modeled with beta distributions. Review of Economics and Statistics94(1), 52-73.

853

Cingano, F. (2014), “Trends in Income Inequality and its Impact on Economic Growth”, OECD Social, Employment and Migration Working Papers, No. 163, OECD Publishing. http://dx.doi.org/10.1787/5jxrjncwxv6j-en

854

Citro, C. and R. T. Michael. 1995. Measuring poverty: A new approach, Washington, D.C.: The National Academies Press.

855

Clarke, G., R., G. (1995). More evidence on income distribution and growth. Journal of Development Economics, 47, pp.403–427.

856

Clarke, G. R., Xu, L. C., and Zou, H. F. (2006). "Finance and Income Inequality: What Do

857

the Data Tell Us? Southern Economic Journal 72(3): 578-596.

858

Claus, I., J. Martinez-Vasquez and V. Vulovic (2014), ‘Government fiscal policies and

859

redistribution in Asia’, in Kanbur et al (2014), op.cit.

860

Cogneau, D., T. Bossuroy,  Ph. De Vreyer,  C. Guénard, V. Hiller, P. Leite, S. Masplé-Somps,

861

L.  “Pasquier-Doumer and C. Torelli (2007),  “Inequalities and Equity in Africa” Agence

862

Française de Développement. Paris.

863

Cogneau, D., and Mesplé-Somps, S. (2008). Inequality of opportunity for income in five countries of Africa. Research on Economic Inequality 16: 99-128.

864

Coleman, James. 1974. Inequality, Sociology, and Moral Philosophy. American Journal of Sociology, vol. 80, no. 3: pp. 739–764.

865

Collier P., and Hoeffler, A., (2004), “Greed and Grievance in Civil War,” Oxford Economic Peace Research, 45 (2), 143-62.

866

Collins, D., Morduch, J., Rutherford, S., & Ruthven, O. (2009). Portfolios of the Poor: How

867

the World's Poor Live on $2 a Day. Princeton, NJ: Princeton University Press.

868

Corak, Miles. 2013. “Income Inequality, Equality of Opportunity and Intergenerational Mobility,” Journal of Economic Perspectives 27 (3): 79-102.

869

Corak, M, L. Curtis, and S. Phipps (2011): “Economic Mobility, Family Background, and

870

the Well-being of Children in the United States and Canada.” In: T. M. Smeeding, R.

871

Erikson, and M. Jäntti (Eds.), Persistence, Privilege, and Parenting. The Comparative Study of Intergenerational Mobility, Russell Sage Foundation, New York, 73-108.

872

Cornia, Giovanni Andrea (2004), “Inequality, growth and Poverty in an Era of Liberalization and Globalization”, Oxford University Press, Oxford.

873

Cornia, G.A. and B. Martorano (2012), ‘Development policies and income inequality in selected developing regions, 1980-2010’ UNCTAD Discussion Paper 210, November 2012

874

Cornia, G.A., J.C. Gomez Sabaini, and B. Martorano (2014), “Tax Policy and Income

875

Distribution During the Last Decade”, in G.A. Cornia (ed.) (2014), “Falling Inequality in

876

Latin America: Policy Changes and Lessons”, Oxford, Oxford University Press.

877

Cornia, G.A., editor. (2014a). Falling Inequality in Latin America: Policy

878

Changes and Lessons, UNU-WIDER Studies in Development Economics, Oxford: Oxford

879

University Press.

880

Cornia, G.A.  (2014b) “Income Inequality in Latin America: Recent Decline and

881

Prospects for its Further Reduction,” Number 149, July, CEPAL Working Papers Series

882

Macroeconomia del Desarrollo.

883

Cornia G.A. (2015), “The political economy of falling (& now rising) inequality in Latin America, 2002-2013”, powerpoint presentation at the conference on Political and Economic Inequality: Concepts, Causes and Consequences, ETH, Zurich 29-1-2016, http://www.cis.ethz.ch/news-and-events/CISEvents/Workshops/inequalityworkshop/speakers0.html

884

Cornia, G.A. (2016) “An econometric analysis of the bifurcation of within-country inequality

885

trends in Sub- Saharan Africa, 1990-2011”, UNDP, Regional Bureau for Africa, Working

886

Paper Series on Inequality N.4, New York. 

887

Cornia G. A. (2016), “An econometric investigation An econometric

888

investigation of the causes of the bifurcation of within-country inequality trends over

889

1991-2011 in Sub-Saharan Africa”, UNDP RBA Working Papers, (forthcoming).

890

Cornia, G. A . (2017). “Is Latin America’s Recent Inequality Decline Permanent or

891

Temporary?”, Development and Change. Forthcoming.

892

Conia G.A. and B. Martorano (2016),  “Building the IID-SSA inequality dataset and the 

893

seven sins of inequality measurement in  Sub-Saharan Africa” , UNDP, Regional Bureau

894

for Africa, Working Paper Series on Inequality N.2, New York.

895

Cotula, L., C. Toulmin and C. Hesse (2004), ‘ Land Tenure and Administration in Africa: Lessons of Experience  and Emerging Issues’ February 2004, International Institute for Environment and Development, London

896

Croizet, J. C., & Claire, T. (1998). Extending the concept of stereotype threat to social

897

class: The intellectual underperformance of students from low socioeconomic

898

backgrounds. Personality and Social Psychology Bulletin, 24(6), 588-594.

899

Cruces, G., R. Perez-Truglia, M. Tetaz. 2013. “Biased perceptions of income distribution and preferences for redistribution: Evidence from a survey experiment.”  Journal of Public Economics 9: 100-112

900

Cruces, G. C. Garcia Domench, and L. Gasparini (2014), “Inequality in Education:

901

Evidence from Latin America“, in G.A.Cornia (2014), “Falling Inequality in Latin America:

902

Policy Changes and Lessons“, Oxford, Oxford University Press.

903

Currie, J. (2006). The take-up of social benefits. In A. Auerbach, D. Card & J. M. Quigley

904

(Eds.), Public Policy and the Income Distribution (pp. 88-108, 121-126). New York:

905

Russell Sage.

906

Dabla-Norris, Era, & Mark Gradstein (2004). "The Distributional Bias of public education:

907

causes and consequences." IMF Working Paper 04/214

908

Dagum, C. (1990) On the relationship between income inequality measures and social

909

Welfare functions, Journal of Econometrics, 43(1–2), pp. 91–102.

910

Daly, M. C., and R. G. Valletta (2006). “Inequality and Poverty in United States: The

911

Effects of Rising Dispersion of Men's Earnings and Changing Family Behaviour.”

912
  • 73 (289), 75-98.
913

Daly, M., Boyce, C., & Wood, A. (2015). A social rank explanation of how money

914

influences health. Health Psychology34(3), 222-230.

915

Datt, Gaurav and Martin Ravallion (2002): Is India’s Growth leaving the poor behind? Journal of Economic Perspectives. Vol.16(3): pp. 89-108.

916

Datt, Gaurav and Martin Ravallion (2009): Has India’s Economic Growth Become More Pro-Poor in the Wake of Economic Reforms? World Bank Policy Research Working Paper 5103.

917

Davidai, S., & Gilovich, T. (2015). Building a more mobile America—One income quintile

918

at a time. Perspectives on Psychological Science10(1), 60-71.

919

Davis, M. (1999) The Planet of Slums, London: Verso.

920

Davies, J. A. Sandström, A. Shorrocks, and E. Wolff. 2008. “The World Distribution of Household Wealth,” Helsinki: UNU Wider.

921

Davies, J.B. and A.F. Shorrocks. 2000. “The distribution of wealth,” in A. Atkinson and F. Bourguignon (eds.), Handbook of Income Distribution Vol 1. Amsterdam: Elsevier.

922

De, R., and C. Becker (2015). The Foreign Aid Effectiveness Debate: Evidence from Malawi. AidData Working Paper 6, March 2015.

923

Deaton, A. (2003). Health, income and inequality.  NBER Research Summary. 

924

Deaton, A. 2005. “Measuring Poverty in a Growing World (or Measuring Growth in a Poor World),” The Review of Economics and Statistics 87(1): 1-19.

925

Deaton, A. and S. Zaidi. 2002. “Guidelines for creating consumption aggregates for Welfare Analysis,” Living Standards Measurement Study Working Paper No. 135.

926

Deaton, A. and R. Kahnemann. (2010) High income improves evaluation of life but not emotional well-being.  PNAS 107 (38): 16489-16493.

927

de Barros, R., F. H. G. Ferreira, J.R.M. Vega, and J. Chanduvi. 2009. Measuring Inequality of Opportunities in Latin America and the Caribbean. Washington DC: The World Bank.

928

De La Croix, D., and Doepke, M. (2003). Inequality and growth: why differential fertility matters. The American Economic Review, 93(4), pp.1091-1113.

929

de la Motte and Paquet, G. (Ed) (1996). Evolutionary Economics and the New International Political Economy. London: Pinter.

930

De la Torre, Augusto, Eduardo Levy-Yeyati, and Samuel Pienknagura. (2013). “Latin

931

America and the Caribbean as Tailwinds Recede: In Search of Higher Growth”,

932

Semiannual Report, Regional Chief Economist Office, Latin America and the Caribbean,

933

The World Bank.

934

Deere, C. D, L. Boakye-Yiadom, C. Doss, A. D. Oduro, et al. 2013. “Women’s Land Ownership and Participation in Agricultural Decision-Making: Evidence from Ecuador, Ghana and Karnataka, India,” The Gender Asset Gap Project, Research Brief Series: No. 2.

935

Deininger K. and L. Squire (1996), ‘A New Dataset Measuring Income Inequaity’, The

936

World Bank Economic Review, Volume 10,, Issue 3, p. 565-591

937

Deininger, K., and Squire, L. (1998). New ways of looking at old issues: inequality and growth. Journal of development economics, 57(2), pp.259-287.

938

Deininger, K.. and D. Byerlee (2011) ,’Rising Global Interest in Farmland: Can it Yield

939

Sustainable and Equitable Benefits?’, The World Bank, Washington DC.

940

Demirguc-Kunt, A. and R.Levine. 2009. Finance and inequality: Theory and evidence.

941

National Bureau of Economic Research.

942

Department for International Development (DFID)-UKAID. 2011. Cash Transfers Literature Review. Policy Division.

943

Dercon, S. (2011), Social Protection, Efficiency and Growth. CSAE Working Paper WPS/2011-17. Oxford: Centre for the Study of African Economies, University of Oxford.

944

de Rijcke, S. and Rushforth, A. (2015)Researcher meets Indicator: Effects of evaluative metrics on research groups”, paper presented at the Atlanta Conference on Science and Innovation Policy, September, 2015.

945

Deutsch, M. (1975). Equity, equality, and need: What determines which value will be

946

used as the basis of distributive justice?. Journal of social issues31(3), 137-149.

947

Diamond, P. (1998) Optimal income taxation: An example with a U-shaped Pattern of Optimal Marginal Tax Rates.  American Economic Review 88 (1): 83-95.

948

Diamond, P. and E. Saez. (2011). The case for a progressive tax: From basic research to policy recommendations.  CESifo working paper No. 3548.

949

Dikhanov, Y., and Ward, M. (2002): Evolution of the Global Distribution of Income 1970–99. , pp 22-29.

950

Dollar, D. and Kraay, A. (2002). Growth is Good for the Poor. Journal of Economic Growth, 7, pp.195-225.

951

Dowrick, S., & Akmal, M. (2005). Contradictory trends in global income inequality: A tale of two biases. Review of Income and Wealth51(2), 201-229.

952

Doyle, J. M., Ahmed, E., and Horn, R. N. (1999). The effects of labor markets and income inequality on crime: evidence from panel data. Southern Economic Journal, pp.717-738.

953

Dreher, A., P. Nunnenkamp, and R. Thiele (2008). Does Aid for Education Educate Children: Evidence from Panel Data. World Bank Economic Review 22(2): 291-314.

954

Drèze, J. and A. Sen. 1995. India: Economic Development and Social Opportunity.  New York: Clarendon Press. 

955

Dreze, J. and A. Sen. 1989.  Hunger and Public Action.  Oxford: Clarendon Press.

956

Ducos, J-Y., J-M. Esteban, and D. Ray. 2004. “Polarization: Concepts, Measurement and Estimation,” Econometrica 72 (6): 1737-1772.

957

Duflo, E. 2012. Women empowerment and economic development.  Journal of Economic Literature 50(4): 1051-1079.

958

Dworkin, R., 1981. “What is Equality? Part 1: Equality of Welfare,” Philosophy and Public Affairs 10 (3): 185-246.

959

Dworkin, R., 1981. “What is Equality? Part 2: Equality of Resources,” Philosophy and Public Affairs 10 (4): 283-345.

960

Easterly, W. (2001). The middle class consensus and economic development.Journal of economic growth6(4), pp.317-335.

961

Easterly, W. (2007). Inequality does cause underdevelopment: Insights from a new instrument. Journal of Development Economics, 84(2), pp.755-776.

962

ECLAC (Economic Comission for Latin America and the Caribbean), 2014. Social Panorama of Latin America 2013, Santiago.

963

ECLAC (Economic Comission for Latin America and the Caribbean), 2015, Inclusive social development. The next generation of policies for overcoming poverty and reducing inequality in Latin America and the Caribbean, Santiago.

964

ECLAC (1990) Changing production patterns with social equity: the prime task of Latin American and Caribbean development in the 1990s, Santiago de Chile.

965

Edin, K., & Shaefer, H. L. (2015). $2.00 a Day: Living on Almost Nothing in America. New York, NY: Houghton Mifflin Harcourt.

966

Ehrhart, C. (2009). The effects of inequality on growth: a survey of the theoretical and empirical literature. ECINEQ WP, 107.

967

Eligon, J. (2016) “A Question of Environmental Racism in Flint,” New York Times, Jan 21.

968

Ellis, F. 2008. We are all poor here: economic difference, social divisiveness, and targeting cash transfers in sub-Saharan Africa. Social Protection for the Poorest in Africa: Learning from Experience, Entebbe, Uganda. 

969

Elster, J. (1985), Sour Grapes: Studies in the Subversion of Rationality, Cambridge: Cambridge University Press.

970

Engelhardt, C. and A. Wagener. 2014. “Biased perceptions of income inequality and redistribution,”  Hannover Discussion Paper No. 526

971

Engerman, S. L., and Sokoloff, K. L. (2002). Factor endowments, inequality, and paths of development among new world economics. National Bureau of Economic Research, No. w9259.

972

Engerman, Stanley L. and Kenneth, L. Sokoloff (2005), “Colonialism, Inequality and Long Run Paths of Development”, Working Paper 11057, NBER, Cambridge, MA.

973

Erten, B. and J.A. Ocampo (2012), “Super-cycles of commodity prices since the mid-nineteenth century”, DESA Working Paper No. 110ST/ESA/2012/DWP/110, February 2012.

974

Esquivel, G., N. Lustig, and J. Scott. (2010). “Mexico: A Decade of Falling

975

Inequality: Market Forces or State Action?,” Chapter 7 in Lopez-Calva, Luis F. and Nora

976

Lustig, editors, Declining Inequality in Latin America: A Decade of Progress? Brookings

977

Institution Press and UNDP.

978

Esteban, J-M. and D. Ray. 1993. “On the Measurement of Polarization,” Econometrica 62 (4): 819-851.

979

Esteban, J., and Ray, D. (1999). Conflict and distribution. Journal of Economic Theory, 87(2), pp.379-415.

980

Esteban, J., and Ray, D. (2006). Inequality, lobbying, and resource allocation. The American Economic Review, pp.257-279.

981

Esteban, J. M. and D. Ray. 2010. Comparing polarization measures, available at: http://www.nyu.edu/econ/user/debraj/Papers/EstRayChapter.pdf.

982

Esteban, J. and D. Ray. (2011) Linking Conflict to Inequality and Polarization.  American Economic Review 101: 1345-1374.

983

Esteva-Volart, B. 2009. gender discrimination and growth: Theory and evidence from India.  Mimeographed, University of York, Canada.

984

Evans, A. 1989. Women: Rural Development Gender Issues in Rural Household Economics. Institute of Development Studies, Discussion Paper 254.

985

Falkinger, J., and Grossmann, V. (2005). Institutions and development: The interaction between trade regime and political system. Journal of Economic Growth, 10(3), pp.231-272.

986

Falkinger, J., and Zweimüller, J. (1997). The impact of income inequality on product diversity and economic growth. Metroeconomica, 48(3), 211-237.

987

Fallah, B. and Partridge, M., (2007), “The elusive inequality-economic growth relationship: are there differences between cities and the countryside?”Annals of Regional Science 41, 375-400.

988

Fan, S., Kanbur, R., & Zhang, X. (2011). „China's regional disparities: Experience and

989

policy.” Review of Development Finance, 1(1), 47-56.

990

Fearon J.D. and D.D. Laitin (2003), “Ethnicity, Insurgency, and Civil War,” American Political Science Review, 97 (1), 75-90.

991

Feenstra, R., and G. Hanson (1997). Foreign Investment, Outsourcing and Relative Wages: Evidence from Mexico’s Maquiladoras. Journal of International Economics 42(3-4): 371-393.

992

Fehr, E., Bernhard, H., & Rockenbach, B. (2008). Egalitarianism in young

993

children. Nature454(7208), 1079-1083.

994

Fehr, E., & Fischbacher, U. (2004). Third-party punishment and social norms. Evolution

995

and human behavior25(2), 63-87.

996

Fehr, E., & Schmidt, K. M. (1999). A theory of fairness, competition, and cooperation. The

997

quarterly journal of economics114(3), 817-868.

999

Ferreira, Francisco H. G., Sergio Firpo, and Julian Messina. (2014). “A More Level Playing

1000

Field? Explaining the Decline in Earnings Inequality in Brazil, 1995-2012”, IRIBA

1001

Working Paper: 12, The University of Manchester.

1002

Ferreira, F.H.G., and J. Gignoux. 2011. “The Measurement of Educational Inequality: Achievement and Opportunity,” Discussion Paper No. 6161, Bonn: Institute for the Study of Labor.

1003

Ferreira, F.H.G. and J. Gignoux. 2011. “The Measurement of Inequality of Opportunity: Theory and an Application to Latin America,” Review of Income and Wealth 57 (4): 622-657.

1004

Ferreira, F. H., & Gignoux, J. (2014). The measurement of educational inequality: Achievement and opportunity.The World Bank Economic Review,28(2), 210-246.

1005

Ferreira, F. and A. Lugo. 2012. “Multidimensional poverty analysis: Looking for a middle ground,”  World Bank Research Observer 28(2): 220-35.

1006

Ferreira, F. H. G. et al. 2016. “A global count of the extreme poor in 2012: data issues, methodology and initial results,”  Journal of Economic Inequality 14: 141-172.

1007

Ferreira, F. H. G., N. Lustig, and D. Teles. 2015. “Appraising Cross-National Income Inequality Databases: An Introduction,” in F. Ferreira and N. Lustig, Special Issue, Appraising Cross-National Income Inequality Databases Journal of Economic Inequality 13 (4): 497-526.

1008

Ferreira, Francisco H.G.; Messina, Julian; Rigolini, Jamele; López-Calva, Luis-Felipe; Lugo, Maria Ana; Vakis, Renos. 2013. “Economic Mobility and the Rise of the Latin American Middle Class,” Latin America and Caribbean Studies; Washington, DC: World Bank.

1009

Ferreira, F. H. G., and Peragine, V. (2015). Equality of opportunity: Theory and evidence. IZA Discussion Paper No. 8994.

1010

Ferrer-i-Carbonell, A. and X. Ramos (2012). Inequality and happiness: a survey,

1011

Ferroni, Maeto, and Payne, J. M. (2007). Development under conditions of inequality and distrust: Social cohesion in Latin America. IFPRI discussion paper, 777.

1012

Fields, G and E. Ok. 1999. The measurement of income mobility: An introduction to the literature in J. Silber (ed.), Handbook on Income Inequality Measurement, Kluwer Academic Publishers.

1013

Fiske, S. T. (2011). Envy up, scorn down: How status divides us. New York: Russell Sage

1014

Foundation.

1015

Fiszbein, A., and N. Schady. 2009. “Conditional Cash Transfers: Reducing Present and Future Poverty.” World Bank, Washington, DC.

1016

                                                                                     

1017

Fleurbaey, M and V. Peragine. 2009. “Ex ante versus ex post equality of Opportunity,” Working Paper No. 2009-141, Society for the Study of Economic Inequality.

1018

Foellmi, R., and Zweimüller, J. (2006). Income distribution and demand-induced innovations. The Review of Economic Studies, 73(4), pp.941-960.

1019

Förster, M. (2016) Rising inequalities in OECD countries: Trends, consequences, drivers,

1020

and policies how to curb them.  Paper presented at 2016 APPAM Conference, London

1021

Förster, M. (2016a) Trends in income inequality in OECD countries.  Mimeographed,

1022

OECD.

1023

Foguel, M.N. and Veloso, F.A. (2014). Inequality of opportunity in daycare and preschool services in Brazil. The Journal of Economic Inequality 12, pp 191-220.

1024

Fong, C. (2001). Social preferences, self-interest, and the demand for

1025

redistribution. Journal of Public Economics82(2), 225-246.

1026

Forbes, Kristin J. (2000), “A reassessment of the relationship between inequality and growth,” American Economic Review, 90(4): 869-887.

1027

Formby, J, J. Smith and B. Zheng. 2004. “Mobility Measurement, Transition Matrices and Statistical Inference,” Journal of Econometrics, 120 (1): 181-205.

1028

Foster, J.E., and M.C. Wolfson. 1994. “Polarization and the Decline of the Middle Class: Canada and the US,” Working Paper No. 31, Oxford Poverty & Human Development Initiative.

1029

Foster, J., Lopez-Calva, L., & Szekely, M. (2005). ‘Measuring the distribution of human development: methodology and an application to Mexico’, Journal of Human Development, 6(1), 5-29.

1030

Fosu, A. K. (2009). Inequality and the impact of growth on poverty: comparative evidence for sub-Saharan Africa, Journal of Development Studies, 45(5), pp. 726–745.

1031

Fosu, A. K. (2011). Growth, Inequality, and Poverty Reduction in Developing Countries: Recent Global Evidence. CSAE Working Paper WPS/2011-07, Centre for the Study of African Economies, Department of Economics, University of Oxford.

1032

Fosu. A K. (2015). Growth, Inequality and Poverty in Sub-Saharan Africa: Recent Progress in a Global Context, Oxford Development Studies, 43:1, 44-59.

1033

Frank, R.H., 1985. Choosing the right pond: Human behavior and the quest for status. Oxford: Oxford University Press.

1034

Frankema  E. (2009), “The colonial roots of land inequality: geography, factor endowments, or institutions?”, The Economic History Review, Volume 63, Issue 2, pages 418–451, May 2010

1035

Freeman, C. (1992) “Science and economy at the national level”, Chapter prepared for the OECD Experimental Working Session on Science Policy, reprinted in Freeman, Ch. The Economics of Hope, Pinter Publishers, 31-49.

1036

Fuest, C., Niehues, J., Peichl, A. (2010). „The Redistributive Effects of Tax Benefit Systems

1037

in the Enlarged EU.” Public Finance Review (38), 473-500.

1038

Galbraith, J. and J. Lu (2001) Inequality and financial crises: some early findings.  In

1039

Galbraith, J. and Berner, M. (eds.), Inequality and industrial change: a global view.  Cambridge, Cambridge University Press. 

1040

Gallup, J. J. Sache, and A. Mellinger. 1998. Geography and Economic Development.  In Stiglitz and Yusuf (eds.).  Proceedings from the 1998 Annual Bank Conference on Development Economics.  New York: Oxford University Press. 

1041

Galor, O., and Zeira, J. (1993). Income distribution and macroeconomics. The review of economic studies, 60(1), pp.35-52.

1042

Galor, O. and D. Weil. 1996. “The Gender Gap, Fertility, and Growth.” American Economic Review 86: 374-387.

1043

Galor, O., and Moav, O. (2004). From physical to human capital accumulation: Inequality and the process of development. The Review of Economic Studies, 71(4), pp.1001-1026.

1044

Gasparini, Leonardo and Guillermo Cruces. (2010). “A Distribution in Motion: The Case of Argentina,” Chapter 5 in Lopez-Calva, Luis F. and Nora Lustig, editors, Declining Inequality in Latin America: A Decade of Progress? Brookings Institution Press and UNDP.

1045

Gasparini, Leonardo, Guillermo Cruces, and Leopoldo Tornarolli. (2016). "Chronicle of

1046

Deceleration Foretold: Income inequality in Latin America in the 2010s". Working Paper

1047

198, CEDLAS-Universidad Nacional de La Plata, May.

1048

____________, Sebastian Galiani, Guillermo Cruces, and Pablo Acosta. (2011). “Educational

1049

Upgrading and Returns to Skills in Latin America. Evidence from a Supply-Demand

1050

Framework, 1990–2010”. Policy Research Working Paper 5921, The World Bank.

1051

____________ and Nora Lustig. (2011). “The Rise and Fall of Income Inequality in Latin America,” Chapter 28 in Jose Antonio Ocampo and Jaime Ros, eds., Oxford Handbook of Latin American Economics, Oxford University Press.

1052

Gasparini, L. and L. Tornarolli. 2015. “A review of the OECD Income Distribution Database,” in Ferreira, F. and N. Lustig, Special Issue, Appraising Cross-National Income Inequality Databases Journal of Economic Inequality 13 (4):

1053

Gennetian, L., & Shafir, E. (2015). The persistence of poverty in the context of economic

1054

instability: A behavioral perspective. Journal of Policy Analysis and Management, 34(4),

1055

904-936.

1056

Gertler, P. (2005), “The Impact of Conditional Cash Transfers on Human Development Outcomes: A Review of Evidence from PROGRESA in Mexico and Some Implications for Policy Debates in South and Southern Africa”, Paper presented at the South African Regional Poverty Network Meeting, Pretoria, 27 January 2005

1057

Ghani, Ejaz, William R. Kerr, Stephen D. O'Connell. 2014. “Political reservations and women's entrepreneurship in India.” Journal of Development Economics 108 (2014) 138–153.

1058

Ghatak, M., and Jiang, N. N. H. (2002). A simple model of inequality, occupational choice, and development. Journal of development Economics,69(1), pp.205-226.

1059

Ghosh, J. (2015),’ Growth, industrialization and inequality in India’, Journal of the Asia and Pacific Economy, Vol. 20, 2015, Issue 1.

1060

Giesbert, L. and S. Schotte (2016). Africa's new middle class: Fact and fiction of its transformative power.GIGA Focus Africa, Number 1.

1061

Giovarelli, Renee, Elizabeth Katz, Susan Lastarria-Cornhiel, and Sue Nichols. 2005. Gender Issues and Best Practices in Land Administration Projects: A Synthesis Report. Report 32571-GLB, Agriculture and Rural Development Department. Washington, DC: World Bank.

1062

Girma,S. and Shortland, A. (2007), ‘The Political Economy of Financial Development’, Oxford Economic Papers, 60(4), pp.567–96.

1063

Glaeser, E., Scheinkman, J., and Shleifer, A. (2003). The injustice of inequality. Journal of Monetary Economics, 50(1), pp.199-222.

1064

Glewwe, P., and P. Olinto. 2004. “Evaluating the Impact of Conditional Cash Transfers on Schooling in Honduras: An Experimental Approach.” Unpublished paper, University of Minnesota, Minneapolis.

1065

Goesling, B., & Firebaugh, G. (2004). The trend in international health inequality. Population and Development Review30(1), 131-146.

1066

Goldberg, P., and N. Pavcnik (2007). Distributional Effects of Globalization in Developing Countries. Journal of Economic Literature 45(1): 39-82.

1067

Goldstein, S. 1972. The Influence of Labour Force Participation and Education on Fertility in Thailand. Population Studies 26: 419-436.

1068

Goldstein, M. and C. Udry. 2008. The profits of power: Land rights and agricultural investment in Ghana.  Journal of Political Economy 116(6): 982-1022.

1069

Goldthorpe, J. H. Social mobility and class structure in Modern Britain. Oxford: Clarendon Press (1987).

1070

Goos, M., Manning, A., & Salomons, A.(2009). “Job polarization in Europe.” The American

1071

Economic Review, 99(2), 58-63.

1072

Gottschalk, P. and S. Danziger: Inequality of wage rates, earnings, and family income in the United States, 1975-2002.  Review of Income and Wealth

1073

Greenwood J, Jovanovic B. 1990. Financial development, growth, and the distribution of

1074

income. J. Polit. Econ.98(5):1076–107.

1075

Griffin, K., A. R. Khan, and A. Ickowitz (2002) “Poverty and the Distribution of Land,” Journal of Agrarian Change, 2 (3), pp. 279-330.

1076

Grimm, M. 2007. Removing the anonymity axiom in the measurement of pro-poor growth. Journal of Economic Inequality.

1077

Grimm, M., K. Harttgen, and M. Misselhorn. (2008) „A Human Development Index by Income Groups”.   World Development 12: 2527-2546 (2008).

1078

Grimm, M. et al. 2010. “Inequality in Human Development: An Empirical Assessment of 32 Countries,” Social Indicators Research 97: 191-211.

1079

Grosse, M., K. Harttgen and S. Klasen. 2008. “Measuring Pro-Poor Growth using Non-Income Indicators,” World Development 36 (6): 1021-1047.

1080

Grossmann, V. (2008). Risky human capital investment, income distribution, and macroeconomic dynamics. Journal of Macroeconomics30(1), 19-42.

1081

Gruen, C. & Klasen, S. (2001) Growth, income distribution and well-being in transition countries, Economics of Transition, 9, pp. 359–394.

1082

Gruen, C. & Klasen, S. (2003) Growth, inequality, and well-being: intertemporal and global comparisons, CESifo Economic Studies, 49, pp. 617–659.

1083

Gruen, C. & Klasen, S. (2008) Growth, inequality, and welfare: comparisons across space and time, Oxford Economic Papers 60, 212-236.

1084

Gruen, C. and Klasen, S. (2012) Has transition improved well-being?  Economic Systems 36, 11-30

1085

Gruen, C. and Klasen, S. (2013) Income, inequality, and subjective well-being: an international and intertemporal perspective using panel data.  Jahrbuch für Wirtschaftsgeschichte/ Economic History Yearbook 2013/1: 15-36 (2013).

1086

Grüner, H. P. (2001). Wealth distribution and interest rates: empirical evidence for the US. CEPR Discussion Paper 3106, Review of Economic Studies, 71, pp. 1001–1026.

1087

Grüner, H. P. (2003). Redistribution as a selection device. Journal of Economic Theory, 108(2), pp.194-216.

1088

Gummerum, M., Hanoch, Y., Keller, M., Parsons, K., & Hummel, A. (2010). Preschoolers’

1089

allocations in the dictator game: The role of moral emotions. Journal of Economic

1090

Psychology31(1), 25-34.

1091

Gurr, T. R., and Moore, W.H., (1997), “Ethno-political Rebellion: A Cross-Sectional Analysis of the 1980s with Risk Assessment for the 1990s,” American Journal of Political Science, 41: 4 1079-1103.

1092

Gwatkin, D, S. Rutstein, K. Jonhson, E. Suliman, A Wagstaff, and A. Amouzou (2007),

1093

‘Socio-economic differences in health, nutrition and population within developing

1094

countries: an overview’, Country Report on PHN and Poverty. The World Bank,

1095

Washington DC.

1097

Haddad, L., & Kanbur, R. 1990. “How Serious Is the Neglect of Intra-Household Inequality?,” Economic Journal 100 (402), 866-81

1098

Hailu, D and R.Tsukada (2012), ‘Is the Distribution of Foreign Aid MDG-Sensitive?’, DESA

1099

Working Paper No. 111, February 2012, United Nations, New York.

1100

Hall, C. C., Zhao, J., & Shafir, E. (2013). Self-Affirmation Among the Poor: Cognitive and

1101

Behavioral Implications. Psychological Science, 25(2), 619-625.

1102

Hallegatte, Stephane; Bangalore, Mook; Bonzanigo, Laura; Fay, Marianne; Kane, Tamaro; Narloch, Ulf; Rozenberg, Julie; Treguer, David; Vogt-Schilb, Adrien. 2016. Shock Waves: Managing the Impacts of Climate Change on Poverty. Climate Change and Development; Washington, DC: World Bank.

1103

Hanson,G. (2003). What Has Happened to Wages in Mexico since NAFTA? NBER Working Paper No. 9563.

1104

Harttgen, K. and S. Klasen. (2012) A Household-based Human Development Index.  World Development 40(5): 878-899.

1105

Haushofer, J., & Fehr, E. (2014). On the psychology of poverty. Science, 344(6186), 862

1106

867.

1107

Henrich, J., McElreath, R., Barr, A., Ensminger, J., Barrett, C., Bolyanatz, A., Cardenas, J.C.,

1108

Gurven, M., Gwako, E., Henrich, N, Lesorogol, C., Marlowe, F., Tracer, D., & Ziker, J. (2006).

1109

Costly punishment across human societies. Science312(5781), 1767-1770.

1110

Herzer, D., and P. Nunnenkamp (2012). The Effect of Foreign Aid on Income Inequality: Evidence from Panel Cointegration. Structural Change and Economic Dynamics 23(3): 245-255.

1111

Herzer, D., and Vollmer, S. (2012). Inequality and growth: evidence from panel cointegration. The Journal of Economic Inequality, 10(4), pp.489-503.

1112

Herzer, D., and Vollmer, S. (2013). Rising top incomes do not raise the tide. Journal of Policy Modeling, 35(4), pp.504-519.

1113

Hess, D. (2007) Alternative Pathways in Science and Industry. Activism, Innovation, and the Environment in an Era of Globalization. Cambridge: The MIT Press.

1114

Hicks, D. (2004) The Four Literatures of Social Science. In Handbook of Quantitative Science and Technology Research, ed. H. Moed, W.  Glänzel and U. Schmoch, 473-496. Dordrecht: Kluwer Academic Publishers.

1115

Himanshu and P. Lanjouw (2015), ‘Some aspects of inequality in India’ Powerpoint

1116

presentation, UNU-WIDER 30th Anniversary Conference on “Mapping the Future of

1117

Development Economics”September 17-19, 2015 Helsinki, Finland

1118

Hirano, Y., and S. Otsubo (2014). Aid is Good for the Poor. World Bank Policy Research Working Paper 6998, August 2014.

1119

Hnatkovska, V., A. Lahiri and S.B. Paul. 2011. “Breaking the Caste Barrier: Intergenerational Mobility in India,” Journal of Human Resources 48 (2): 435-473.

1120

Ho, E. A., Sanbonmatsu, D. M., & Akimoto, S. A. (2002). The effects of comparative status

1121

on social stereotypes: How the perceived success of some persons affects the

1122

stereotypes of others. Social Cognition20(1), 36-57.

1123

Hoeffler, A., and V. Outram (2011). Need, Merit, or Self-Interest: What Determines the Allocation of Aid. Review of Development Economics 15(2): 237-250.

1124

Hoffman, E., McCabe, K., Shachat, K., & Smith, V. (1994). Preferences, property rights, and

1125

anonymity in bargaining games. Games and Economic Behavior7(3), 346-380.

1126

Holmes, R. and Jones, N., 2010, Rethinking social protection from a gender lens. ODI Working Paper No. 320, London http://www.odi.org.uk/resources/details.asp?id=5099&title=social-protection-gender

1127

Honohan, P. (2005) "Banking sector crises and inequality," Policy Research Working Paper Series 3659, The World Bank.

1128

Horan, P. M., & Austin, P. L. (1974). The social bases of welfare stigma. Social Problems,

1129

648-657.

1130

Hsieh, C., and K. T. Woo (2005).The Impact of Outsourcing to China on Hong Kong's Labor Market. American Economic Review 95(5): 1673-1687.

1131

Hufe, P. and Peichl. A (2016). Beyond Equal Rights: Equality of Opportunity in Political Participation. Mimeo.

1132

Hufe, P., Peichl, A., Roemer, J. and Ungerer, M. (2015). Inequality of Income Acquisition: The Role of Childhood Circumstances. ZEW Discussion Paper No. 15-084.

1133

Humphries, R. (2008). Periodical Review of the Cash Transfers for Development Project. Technical Report, Oxfam Great Britain in Viet Nam.

1134

Huntington, S. (1968), “Political Order in Changing Societies” Yale University Press, New Haven, C.T.

1135

Hyslop, D. R., and D. C. Mare (2005). “Understanding New Zealand's Changing Income

1136

Distribution, 1983-1998: A Semi-Parametric Analysis.” Economica, 72 (3), 469-495.

1137

ILO (2014), “World Social Protection Report 2014-5”, Geneva, ILO

1138

IMF (2007), ‘World Economic Outlook’ IMF, Washington DC.

1139

International Monetary Fund (IMF). (2014). “Fiscal Policy and Income Inequality.” IMF

1140

Policy Paper, Washington.

1141

Immervoll, H., Richardson, L. (2011). “Redistribution Policy and Inequality Reduction in

1142

OECD Countries: What Has Changed in Two Decades?.” LIS Working Paper No. 571.

1143

Ivaschenko, O. (2003). Inequality and growth: Evidence from Transitional Economies in the 1990s. In Eicher and Turnovsky (2003), Inequality and Growth, MIT Press.

1144

Jayadev, A., S. Motiram and V. Vakulabharanam. 2007. “Patterns of Wealth Disparities in India During the Era of Liberalization,” Economic and Political Weekly 42 (39): 3853-63.

1145

Jayadev, A., Lahoti, R., & Reddy, S. G. (2015). Who got what, then and now? A Fifty Year Overview from the Global Consumption and Income Project. A Fifty Year Overview from the Global Consumption and Income Project (May 8, 2015).

1146

Jayaraj, D. and S. Subramanian. 2012. “On the “Inclusiveness” of India’s Consumption Expenditure,” UNU-WIDER Working Paper, No. 2012/57, Helsinki, Finland.

1147

Jenkins, S. P. (1995): “Accounting for Inequality Trends: Decomposition Analysis for the

1148

UK, 1971-86,” Economica, 62, 29–63.

1149

Jenkins, S. (1997) Trends in real income in Britain: a microeconomic analysis, Empirical Economics, 22, pp. 483–500.

1150

Jenkins, S. 2015. “World income inequality databases: an assessment of WIID and SWIID,” in F. Ferreira and N. Lustig, Special Issue, Appraising Cross-National Income Inequality Databases Journal of Economic Inequality, 13 (4):

1151

Joerg, B., Marco, M. D. E., Auke, R., & Marcel, T. (Eds.). (2014). How Was Life? Global Well-being since 1820: Global Well-being since 1820. OECD Publishing.

1152

Johnson, T., Dawes, C. T., Fowler, J. H., McElreath, R., & Smirnov, O. (2009). The role of

1153

egalitarian motives in altruistic punishment. Economics Letters,102(3), 192-194.

1154

Jost, J. T., & Banaji, M. R. (1994). The role of stereotyping in system‐justification and the

1155

production of false consciousness. British journal of social psychology33(1), 1-27.

1156

Josten, S. D. (2003). Inequality, crime and economic growth. A classical argument for distributional equality. International Tax and Public Finance10(4), pp.435-452.

1157

Kakwani, N. (1981) Welfare measures: an international comparison, Journal of Development Economics, 8, pp. 21–45.

1158

Kaldor, N., (1955), “Alternative Theories of Distribution,” The Review of Economic Studies, 23(2), 83-100.

1159

Kalmijn W. and RE. Veenhoven. 2005. Measuring inequality of happiness in nations: In search for proper statistics. Journal of Happiness Studies 6: 357-396.

1160

Kalwij, A. & Verschoor, A. (2007). Not by growth alone: the role of the distribution of income in regional diversity in poverty reduction, European Economic Review, 51, pp. 805–829.

1161

Kanbur, R., C. Rhee and J.Zhuong (2014), ‘Inequality in Asia and the Pacific:Trends,

1162

Drivers and Policy Implications’, Co-published by the Asian development Bank and

1163

Routledge. 

1164

Kanbur. R., and A. Wagstaff. 2009. “How Useful is Inequality of Opportunity as a Policy Construct?,” Policy Research Working Paper No. 6980, Washington DC: The World Bank.

1165

Karter, A. J., Parker, M. M., Moffet, H. H., Ahmed, A. T., Ferrara, A., Liu, J. Y., & Selby, J. V.

1166

(2004). Missed appointments and poor glycemic control: an opportunity to identify

1167

high-risk diabetic patients. Medical care42(2), 110-115.

1168

Katz, S. J., & Hofer, T. P. (1994). Socioeconomic disparities in preventive care persist

1169

despite universal coverage: breast and cervical cancer screening in Ontario and the

1170

United States. Jama, 272(7), 530-534.

1171

Kayser, J. (2004) Glycobiology. Synthetic Vaccine Is a Sweet Victory for Cuban Science. Science July 23, 305 (5683).

1172

Keefer, P., and Knack, S. (2002). Polarization, politics and property rights: Links between inequality and growth. Public choice, 111(1-2), pp.127-154.

1173

Keifman, S. and R. Maurizio (2014), “Changes in Labour Market Conditions and Policies,

1174

and Their Impact on Wage Inequality During the Last Decade”, in G.A. Cornia, “Falling

1175

Inequality in Latin America: Policy Changes and Lessons”, Oxford, Oxford University

1176

Press. 

1177

Kelly, M. (2000). Inequality and crime. Review of economics and Statistics, 82(4), pp.530-539.

1178

Kerbo, H.R. (1976). The stigma of welfare and a passive poor. Sociology & Social

1179

Research, 60, 173-187.

1180

Kiatpongsan, S., & Norton, M. I. (2014). How much (more) should CEOs make? A

1181

universal desire for more equal pay. Perspectives on Psychological Science9(6), 587-593.

1182

King, E., S. Klasen, and M. Porter. 2009. Women and Development. In Lomborg, B. (ed.) Global Crises, Global Solutions, 2nd Edition. Cambridge: Cambridge University Press. 

1183

Kissane, R. J. (2003). What's Need Got to Do With It? Barriers to Use of Nonprofit Social

1184

Services. Journal of Sociology & Social Welfare, 30, 127-148.

1185

Klasen, S. 2002. Low Schooling for Girls, Slower Growth for All? Cross‐Country Evidence on the Effect of Gender Inequality in Education on Economic Development World Bank Econ Rev (2002) 16 (3): 345-37

1186

Klasen, S. 2003. “Malnourished and Surviving in South Asia, better Nourished and Dying Young in Africa:  What can Explain this Puzzle?”  In FAO (eds.) Measurement and Assessment of Food Deprivation and Undernutrition, pp. 283-286.  Rome: FAO (2003).

1187

Klasen, S. 2004. In Search of the Holy Grail: How to achieve Pro-Poor Growth?  In

1188

Tungodden, B., N. Stern, and I. Kolstad (eds).  Toward Pro Poor Policies-Aid, Institutions, and Globalization.  New York: Oxford University Press (forthcoming 2004). 

1189

Klasen, S. 2007. Gender-related indicators of well-being,” in M. McGillivray (ed.), Human Well-Being: Concept and Measurement. London, 167-192.

1190

Klasen, S. 2008. The efficiency of equity. Review of Political Economy 20(2):257-274 (2008). 

1191

Klasen, S. 2008. “Economic Growth and Poverty Reduction: Measurement Issues in Income and Non-Income Dimensions,” World Development 36 (3): 420-45.

1192

Klasen, S. and F. Lamanna. 2009. “The Impact of Gender Bias in Education and Employment on Economic Growth: New Evidence for a Panel of Countries,” Feminist Economics 15.3: 91-132.

1193

Klasen, S. and R. Lahoti. 2016. “How Serious is the Neglect of Intra-Household Inequality in Multi-dimensional Poverty Indices?,” Courant Research Center Discussion Paper. Available at SSRN.

1194

Klasen, S. and R. Lahoti. 2016. “How serious is the neglect of intrahousehold inequality in multidimensional poverty measures in India?,” Courant Research Centre: Poverty, Equity and Growth Discussion Paper. Available at SSRN.

1195

Klasen, S. and S. Lange. 2016. “How narrowly should anti-poverty programs be targeted?,” Courant Research Center: Poverty, Equity and Growth - Discussion Paper No 213.

1196

Klasen, S. and Misselhorn, M. (2008). Determinants of the Growth Semi-Elasticity of Poverty Reduction. Ibero America Institute for Ecn. Research (IAI) Discussion Papers 176, Ibero-America Institute for Economic Research.

1197

Klasen, S. and C. Wink. 2003. Missing Women: Revisiting the Debate.  Feminist Economics 9.2-3:263-299.

1198

Klasen, S. et al. (2016). “ (No. 209). Courant Research Centre: Poverty, Equity and Growth-Discussion Papers.

1199

Klasen, S. et al. (2016). “International income poverty measurement: which way now?,”  Journal of Economic Inequality 14.2: 199-225.

1200

Kluegel, J. R., & Smith, E. R. (1986). Beliefs about inequality: Americans' views of what is

1201

and what ought to be. Hawthrone, NY: Aldine de Gruyter.

1202

Knowles, S. (2005). Inequality and economic growth: the empirical relationship reconsidered in the light of comparable data. The Journal of Development Studies, 41(1), pp.135-159.

1203

Kogut, T. (2011b). Someone to blame: When identifying a victim decreases

1204

helping. Journal of Experimental Social Psychology47(4), 748-755.

1205

Kolm, Sergeu-Christophe. 1976. “Unequal Inequalities I,” Journal of Economic Theory 12 (3): 416-442.

1206

Kolm, Sergeu-Christophe 1976. “Unequal Inequalities II,” Journal of Economic Theory 13 (1): 82-111.

1207

Kornai, J. (1992) The Socialist System. Cambridge: Harvard University Press. 

1208

Koujianou-Goldberg, P., and N. Pavcnik (2007). ‘Distributional Effects of Globalization in

1209

Developing Countries’. NBER Working Paper 12885. Cambridge, MA: National Bureau of

1210

Economic Research.

1211

Kraus, M. W. (2015). Americans Still Overestimate Social Class Mobility: A Pre-

1212

Registered Self-Replication. Frontiers in Psychology, 6.

1213

Kraus, M. W., & Tan, J. J. (2015). Americans overestimate social class mobility. Journal of

1214

Experimental Social Psychology 58: 101-111.

1215

Kremer, M. 1993. Population Growth and Technical Change: 10 Million B.C. to 1990.  Quarterly Journal of Economics.

1216

Kremer, M., and Chen, D. L. (2002). Income distribution dynamics with endogenous fertility. Journal of Economic growth, 7(3), pp.227-258.

1217

Krueger, Alan. 2012. “The Rise and Consequences of Inequality.” Presentation made to the Center for American Progress, January 12th.  http://www.americanprogress.org/events/2012 /01/12/17181/the-rise-and-consequences-of-inequality

1218

Kumhof, M.  and R. Rancière, (2011) Inequality, Leverage and Crises (2011) IMF Working Paper 10/268.

1219

Lagerlöf, N. 2003. ‘‘Gender Equality and Long-Run Growth.’’ Journal of Economic Growth 8(4): 403–26.

1220

Lahoti, R., Jayadev, A., & Reddy, S. G. (2016). The Global Consumption and Income Project (GCIP): An Overview. Journal of Globalization and Development. Volume 7, Issue 1, Pages 61–108

1221

Lakner, C., & Milanovic, B. (2015). Global Income Distribution: From the Fall of the Berlin Wall to the Great Recession. The World Bank Economic Review, lhv039.

1222

Lazonick, W. and M. Mazzucato (2012) ‘The Risk-Reward Nexus. Innovation, Finance and Inclusive Growth’, Policy Network Paper, New York.

1223

Lee, J. S., & Bowen, N. K. (2006). Parent involvement, cultural capital, and the

1224

achievement gap among elementary school children. American Educational Research

1225

Journal43(2), 193-218.

1226

Lefranc, A., Pistolesi, N. and Trannoy, A. (2008), Inequality of Opportunities vs. Inequality of Outcomes: Are Western Societies all alike?. Review of Income and Wealth 54, pp 513–546.

1227

Levine R, Levkov A, Rubinstein Y. 2009. “Racial Discrimination and Competition.”

1228

National Bureau of Economic Research Working Paper No. 14273.

1229

Li Shi, H. Sato and T. Sicular (eds.). 2013.‘Rising Inequality in China:  Challenges to a

1230

Harmonious Society’.  New York:  Cambridge University Press.  

1231

Li, H., and Zou, H., (1998), “Income Inequality Is not Harmful for Growth: Theory and Evidence,” Review of Development Economics, 2, 318-334.

1232

Lindert, Peter H. 2004. Growing Public: Social Spending and Economic Growth since the Eighteenth Century. Volume I: The Story. New York: Cambridge University Press.

1233

Liska, A. E., and Warner, B. D. (1991). Functions of crime: A paradoxical process. American Journal of Sociology, pp.1441-1463.

1234

Loayza, N., Rigolini, J., and Llorente, G. (2012). Do middle classes bring about institutional reforms? Economics letters, 116(3), pp.440-444.

1235

Loewenstein, G. F., Thompson, L., & Bazerman, M. H. (1989). Social utility and decision

1236

making in interpersonal contexts. Journal of Personality and Social psychology57(3),

1237

426-441.

1238

Lohmann, S. 2015. “Information technologies and subjective well-being: does the internet raise material aspirations?,” Oxford Economic Papers.

1239

Lopez-Calva, Luis F. and Nora Lustig, editors. (2010). Declining Inequality in Latin

1240

America: A Decade of Progress? Brookings Institution Press and UNDP.

1241

Lopez-Calva, L.F. and E, Ortiz-Juarez. (2014). A vulnerability approach to the definition of the middle class. Journal of Economic Inequality 12: 23-47.

1242

López-Calva, Luis Felipe and Carlos Rodríguez-Castelán. 2016. “Pro-Growth Equity: A Policy Framework for the Twin Goals.” Policy Research Working Paper No. 7897, The World Bank.

1243

Luck, S. J., & Vogel, E. K. (1997). The capacity of visual working memory for features and

1244

conjunctions. Nature390(6657), 279-281.

1245

Lustig, N., 2016. “The Sustainable Development Goals, Domestic Resource Mobilization and the Poor,” Background Document for WDR 2017.

1246

Lustig, N. (forthcoming),’Inequality and Fiscal Redistribution in Middle-income

1247

Countries: Brazil, Chile, Colombia, Indonesia, Mexico, Peru and South Africa’, Journal of

1248

Globalization and Development, forthcoming.

1249

Lustig, Nora.  forthcoming. “Fiscal Policy, Income Redistribution and Poverty Reduction in Low and Middle Income Countries,” Chapter 9 in Lustig, Nora, editor, Commitment to Equity Handbook A Guide to Estimating the Impact of Fiscal Policy on Inequality and Poverty, Brookings Institution and CEQ Institute (available online).

1250

Lustig, Nora and Ruthanne Deutsch. 1998. The Inter-American Development Bank and Poverty Reduction: An Overview. Inter-American Development Bank, Washington, DC, May.

1251

Lustig, N., Luis F. Lopez-Calva and Eduardo Ortiz-Juarez. (2016). “Deconstructing the

1252

Decline in Inequality in Latin America, ” Chapter 7 in Basu, Kaushik and Joseph Stiglitz

1253

Proceedings of IEA roundtable on Shared Prosperity and Growth, Palgrave-Macmillan.

1254

Lustig, N. and C. Pessino. 2014. “Social Spending and Income Redistribution in

1255

Argentina in the 2000s:  the Rising Role of Noncontributory Pensions,” in Public Finance

1256

Review, May, Volume 42, Issue 3.

1257

Lustig, N. and D. Teles. 2016. “Inequality Convergence: How Sensitive Are Results to the Choice of Underlying Data?,” Tulane University Economics Working Paper XXX, XXX[MONTH].

1258

Lysandrou, P. (2011) Globan inequality as one of the root causes of the financial crisis: a suggested explanation.  Economy and Society 40(3) : 323-344.

1259

Maddison, A. (2001). The World Economy: A Millenial Perspective, Development Centre of the Organization for Economic Cooperation and Development.

1260

Maldonado, J. (2011) ‘Políticas de Inovação no Complexo Industrial da Saúde’, paper presented at the international seminar ‘Innovation Policies and Structural Change in a context of Growth and Crisis’, Rio de Janeiro.

1261

Mancini, L. 2008. “Horizontal Inequality and Communal Violence: Evidence from Indonesian Districts,” in F. Stewart (ed.), Horizontal Inequalities and Conflict: Understanding Group Violence in Multiethnic Societies, Basingstoke: Palgrave Macmillan.

1262

Mandisodza, A. N., Jost, J. T., & Unzueta, M. M. (2006). “Tall Poppies” and “American

1263

Dreams” Reactions to Rich and Poor in Australia and the United States. Journal of Cross-

1264

Cultural Psychology37(6), 659-668.

1265

Mani, A., Mullainathan, S., Shafir, E., & Zhao, J. (2013). Poverty impedes cognitive

1266

function. Science341(6149), 976-980.

1267

Mankiw, G. D. Roemer, and P.Weil. 1992. A Contribution to the Empirics of Economic Growth.  Quarterly Journal of Economics. 

1268

Mann, Michael (1986-2012, three volumes), The Sources of Social Power , Cambridge:

1269

Cambridge University Press

1270

Martin, B. and Whitley, R. (2010) The UK Research Assessment Exercise: A Case of Regulatory Capture? In Reconfiguring Knowledge Production: Changing Authority Relationships in the Sciences and their Consequences for Intellectual Innovation, ed. R.

1271

Whitley, J. Gläser and L. Engwall. New York: Oxford University Press Inc.

1272

Martin, M. A. (2006). “Family Structure and Income Inequality in Families with Children,

1273

1976 to 2000.” Demography, 43 (3), 421-445.

1274

Muller, J. (2010) Befit for Change: Social Construction of Endogenous Technology in the South. FAU Conference. Gjerrild, Grenaa, Djursland, Denmark.

1275

Meltzer, Allan H., and Scott F. Richard, "A Rational Theory of the Size of Government," Journal of Political Economy, 89 (1981), 914-927

1276

Messner, S. F., Baumer, E. P., and Rosenfeld, R. (2004). Dimensions of social capital and rates of criminal homicide. American Sociological Review, 69(6), pp.882-903.

1277

Midlarsky, M.I. and Roberts, K. (1985) “Class, state, and revolution in Central America: Nicaragua and El Salvador compared.” Journal of Conflict Resolution, Vol. 29, pp. 163–193.

1278

Milanovic, B. (2000). The median-voter hypothesis, income inequality, and income redistribution: an empirical test with the required data. European Journal of Political Economy, Vol. 16, pp. 367–410.

1279

Milanovic, B. (2002). True world income distribution, 1988 and 1993: First calculation based on household surveys alone. The economic journal,112(476), 51-92.

1280

Milanovic, B. (2005). Can we discern the effect of globalization on income distribution? Evidence from household surveys. The World Bank Economic Review19(1), 21-44.

1281

Milanovic, B. 2005. Worlds Apart: Measuring International and Global Inequality. Princeton: Princeton University Press.

1282

Milanovic, B. (2012). Global inequality recalculated and updated: the effect of new PPP estimates on global inequality and 2005 estimates. The Journal of Economic Inequality10(1), 1-18.

1283

Milanovic, B. (2011). A short history of global inequality: The past two centuries. Explorations in Economic History48(4), 494-506.

1284

Milanovic, B (2016), Global inequality: A new approach for the age of globalization, Harvard University Press.

1285

Milazzo, Annamaria. 2016. “Governance and Women’s Economic and Political Participation: Formal Constraints, Norms and Power.” Background paper for the World Development Report 2017, World Bank, Washington, DC.

1286

Miller, G. A. (1956). The magical number seven, plus or minus two: some limits on our

1287

capacity for processing information. Psychological Review, 63(2), 81.

1288

Miller, C.M. (2011), Cash Transfers and Economic Growth: A Mixed Methods Analysis of Transfer Recipients and Business Owners in Malawi. Poverty & Public Policy, Vol. 3, Iss. 3.

1289

Miranti, Riyana. et al. (2013): Trends in Poverty and Inequality in Decentralising

1290

Indonesia. OECD Social, Employment and Migration Working Papers No. 148. OECD

1291

Publishing.

1292

Mirrlees, J.A., (1971), “An Exploration in the Theory of Optimum Income Taxation,” Review of Economic Studies, Vol. 38, No 2, 175-208.

1293

Mishra, P., and D. Newhouse (2009). Does Health Aid Matter? Journal of Health Economics 28: 855-872.

1294

Mogues, T., and Carter, M. R. (2005). Social capital and the reproduction of economic inequality in polarized societies. The Journal of Economic Inequality, 3(3), pp.193-219.

1295

Molina-Millan, Teresa, Tania Barham, Karen Macours, John A. Maluccio, and Marco Stampini.2016. “Long-Term Impacts of Conditional Cash Transfers in Latin America”. TECHNICAL NOTE Nº IDB-TN-923. Inter-American Development Bank,

1296

Montalvo, Jose and Marta Reynal-Querol (2005), “Ethnic Polarization, Potential Conflict, and Civil Wars,” American Economic Review, 95(3), 796-816.

1297

Motiram, S. and A. Singh. 2012. “How Close does the Apple Fall to the Tree? Some Evidence on Intergenerational Occupational Mobility from India,” Economic and Political Weekly 47(40): 56-65.

1298

Motiram, S. and K. Naraparaju. 2014. “Growth and Deprivation in India: What does Recent Evidence Suggest on “Inclusiveness”?,” Oxford Development Studies 43(2):145-164.

1299

Motiram, S. and N. Sarma. 2014. “The Tragedy of Identity: Reflections on Violent Social Conflict in Western Assam,” Economic and Political Weekly 49 (11): 45-53.

1300

Motiram, S. and N. Sarma. 2014. “Polarization, Inequality and Growth: The Indian Experience,” Oxford Development Studies 42 (3): 297-318.

1301

Moyos, S. and P. Yeros (2007), ‘Reclaiming the land: the Resurgence of Rural Movements

1302

in Africa, Asia and Latin America’, Zed Books, London.

1303

Mueller, C. and A. Trannoy. 2012. “Multidimensional inequality comparisons: A compensation perspective,” Journal of Economic Theory 147(4): 1427-1449.

1304

Mullainathan, S., & Shafir, E. (2009). Savings policy and decision-making in low-income

1305

households. In M. Barr & R. Blank, Insufficient funds: Savings, assets, credit, and banking

1306

among low-income households121, 140-142. New York, NY: Russell Sage Foundation

1307

Press.

1308

Mullainathan, S., & Shafir, E. (2013). Scarcity: Why Having Too Little Means So Much. New

1309

York, NY: Henry Holt Times Books.

1310

Muller, E. N. (1985), “Income Inequality, Regime Repressiveness, and Political Violence,” American Sociological Review, 50(1), 47–61.

1311

Muller, E.N., and Seligson, M.A., (1987), “Inequality and Insurgency,” American Political Science Review, Vol. 81, No. 2, pp. 425–452.

1312

Murphy, K. M., Shleifer, A., and Vishny, R. (1989): Income Distribution, Market Size, and Industrialization. Quarterly Journal of Economics, Vol. 104(3), pp. 537-564.

1313

Murray, C., A. Lopez, and M. Alvarado. 2013. “The State of US Health, 1990–2010: Burden

1314

of Diseases, Injuries, and Risk Factors.” The Journal of the American Medical Association

1315

310 (6): 591–606.

1316

Murthi, M. and J. Drèze. 2001.  “Fertility, Education, and Development: Evidence from India.” Population and Development Review 27: 33-63.

1317

Ndikumana, L. (2014) Capital flight and tax havens: Impact on investment and growth in Africa.  Revue d'Economie du developpement 2014(2), 99-124.

1318

Neal, R. D., Lawlor, D. A., Allgar, V., Colledge, M., Ali, S., Hassey, A., Portz, C., & Wilson, A.

1319

(2001). Missed appointments in general practice: retrospective data analysis from four

1320

practices. British Journal of General Practice51(471), 830-832.

1321

Neisser, U. (1976). Cognition and reality: Principles and implications of cognitive

1322

psychology. WH Freeman/Times Books/Henry Holt & Co.

1323

Neves, P. C., and Silva, S. M. T. (2014). Inequality and growth: Uncovering the main conclusions from the empirics. Journal of Development Studies, 50(1), pp.1-21.

1324

Niehues, J. and Peichl, A. (2014). Upper bounds of inequality of opportunity: theory and evidence for Germany and the US. Social Choice and Welfare 43(1), pp 73-99.

1325

Nielsen, J. (2005). Lower-Literacy Users: Writing for a Broad Consumer Audience.

1326

Nielsen Norman Group. Retrieved from https://www.nngroup.com/articles/writing-for

1328

Niño-Zarazúa, M, A. Barrientos, D. Hulme and S. Hickey (n.d.) ‘Social protection in sub-

1329

Saharan Africa:Getting the politics right’ Brooks World Poverty Institute, University of

1330

Manchester.

1331

Nolte, K., C. Wytske and M. Giger (2016), “International Land Deals for Agriculture. Fresh insights from the Land Matrix: Analytical Report II“. Bern, Montpellier, Hamburg, Pretoria: Centre for Development and Environment, University of Bern; Centre de coopération internationale en recherche agronomique pour le développement; German Institute of Global and Area Studies; University of Pretoria; Bern Open Publishing.

1332

Norton, M. I. (2013). All ranks are local: why humans are both (painfully) aware and

1333

(surprisingly) unaware of their lot in life. Psychological Inquiry,24(2), 124-125.

1334

Norton, M. I., & Ariely, D. (2011). Building a better America—One wealth quintile at a

1335

time. Perspectives on Psychological Science6(1), 9-12.

1336

Norton, M. I., Neal, D. T., Govan, C. L., Ariely, D., & Holland, E. (2014). The Not‐So‐

1337

Common‐Wealth of Australia: Evidence for a Cross‐Cultural Desire for a More Equal

1338

Distribution of Wealth. Analyses of Social Issues and Public Policy14(1), 339-351.

1339

Nozick, R. (1974), Anarchy, State and Utopia, New York: Basic Books.

1340

Ocampo, J. A. 2012. The Development Implications of External Integration in Latin America. UNU/WIDER Working Papers 48/2012. Helsinki, UNU/WIDER.

1341

Ocké-Reis, Carlos Octavio. 2012. Sus. O desafio de ser único, Editora Fiocruz, Rio de Janeiro.

1342

Odokonyero, T., A. Ijjo, R. Marty, T. Muhumuza, G.O. Moses (2015). Sub-National Perspectives on Aid Effectiveness: Impact of Aid on Health Outcomes in Uganda. AidData Working Paper 18, December 2015.

1343

O’Donnell, O, E. van Doorslaer, A. Wagstaff, and A. Lindelow. 2008. “Analyzing Health Equity Using Household Survey Data: A Guide to Techniques and Their Implementation,” Washington DC: The World Bank.

1344

OECD (2008). Growing Unequal?: Income Distribution and Poverty in OECD Countries,

1345

OECD Publishing, Paris.

1346

OECD (2011). Divided We Stand: Why Inequality Keeps Rising, OECD Publishing, Paris.

1347

OECD (2011). “Growing Income Inequality in OECD Countries: What Drives it and How

1348

Can Policy Tackle it?” OECD Forum on Tackling Inequality, Paris 2 May.

1349

OECD (2015). In It Together: Why Less Inequality Benefits All, OECD Publishing, Paris.

1350

Oezler, Berk, Sarah Baird, and Craig McIntosh. 2016. “When the Money Runs Out: Do Cash Transfers Have Sustained Effects on Human Capital Accumulation?” World Bank Policy Research Working Paper 7901, Washington, DC.

1351

Olson, M. (1965), The Logic of Collective Action: Public Goods and the Theory of Groups, Cambridge, MA: Harvard University Press.

1352

Osberg, L., Smeeding, T. M., and Schwabish, J. (2004). Income distribution and public social expenditure: Theories, effects, and evidence. Social inequality, pp.821-859.

1353

Østby, G. 2008. “Inequalities, the Political Environment and Civil Conflict: Evidence from 55 Developing Countries,” in F. Stewart (ed.), Horizontal Inequalities and Conflict: Understanding Group Violence in Multiethnic Societies. Basingstoke: Palgrave Macmillan.

1354

Ostry, J. A. Berg, and C. Tsangarides. (2014) Redistribution, inequality, and growth.  IMF Staff Discussion Note 14/02.

1355

Ostrom, Elinor (1990). Governing the commons: the evolution of institutions for collective action. Cambridge University Press.

1356

Ott, J. 2005. “Level and inequality of happiness in nations: Does greater happiness of a greater number imply greater inequality in happiness?,” Journal of Happiness Studies 6(4): 397-420.

1357

Oxoby, R. J., & Spraggon, J. (2008). Mine and yours: Property rights in dictator

1358

games. Journal of Economic Behavior & Organization65(3), 703-713.

1359

Pande, R. (2003) “Can Mandated Political Representation Increase Policy Influence for Disadvantaged Minorities? Theory and Evidence from India,” American Economic Review, 93 (4), pp. 1132-1151.

1360

Parfit, D. (1997), “Equality and Priority”, Ratio, Vol. 10, No. 3.

1361

Pavcnik, N. (2011). Globalization and Within-Country Income Inequality. In: M.

1362

Bacchetta and M. Jansen (eds), Making Globalization Socially Sustainable. WTO/ILO: Geneva.

1363

Peichl, A., N. Pestel, and H. Schneider (2012). „Does Size Matter? The Impact of Changes

1364

in Household Structure on Income Distribution in Germany.” Review of Income and

1365

Wealth, 58 (1), 118-141.

1366

Peichl, A. and Ungerer, M. 2015. Equality of Opportunity: East vs. West Germany. SOEPpapers No. 798.

1367

Perotti, R. 1993. Political equilibrium, income distribution, and growth. The Review of Economic Studies60(4), pp.755-776.

1368

Perotti, R. 1996. Growth, income distribution, and democracy: what the data say. Journal of Economic growth, 1(2), pp.149-187.

1369

Persson, T. and Tabellini, G. 1994. Is inequality harmful for growth?. American Economic Review, 84, pp.600-621.

1370

Perugini, C. J. Hölscher, and S. Collie. 2015. Inequality, credit, and financial crises.  Cambridge Journal of Economics doi:10.1093/cje/beu075.

1371

Petri, P.A., and M. Plummer. 2016. Economics of the Trans-Pacific Partnership: Distributional Impact. VoxEU, 30 April 2016.

1372

Pfeffer, J., & Langton, N. 1988. Wage inequality and the organization of work: The case

1373

of academic departments. Administrative Science Quarterly, 588-606.

1374

Pickett, K. and R. Wilkinson. 2015. Income inequality and health: a causal review.  Social Science and Medicine 128: 316-326.

1375

Piketty, T 1997. The dynamics of the wealth distribution and the interest rate with credit rationing. The Review of Economic Studies, 64(2), pp. 173-189

1376

Piketty, T. 2014. Capital in the Twenty-First Century. Cambridge: Harvard University Press.

1377

Pistolesi, N. 2009. Inequality of opportunity in the land of opportunities, 1968–2001. The Journal of Economic Inequality 7(4), pp 411-433.

1378

Pogge, T. 2002. “Can the Capability Approach be Justified?”, Philosophical Topics, Vol. 30.

1379

Poulantzas, N. 1974. Fascism and Dictatorship, London: Verso.

1380

Pradhan, M. and M. Ravallion. 2000. “Measuring poverty using qualitative perceptions of consumption adequacy,” Review of Economic and Statistics 82(3): 462-471.

1381

Pradhan, M., Sahn, D. E., & Younger, S. D. 2003. Decomposing world health inequality. Journal of health economics22(2), 271-293.

1382

Prados de la Escosura, L. 2007. “Inequality and Poverty ion Latin America: A Long Run Exploration”, in K.H.Hatton, K.O’Rourkeand A.M. Taylor (eds) ‘The New Coparative Economic History: Essays in Honour of Jeffrey Williamson’, MIT Press, Cambridge.

1383

Prados de la Escosura, L. 2015. World human development: 1870–2007. Review of Income and Wealth61(2), 220-247.

1384

Prasad, E., Rogoff, K., Wei, S-J. and M. Kose. 2003. “Effects of Financial Globalisation on

1385

Developing Countries: Some Empirical Evidence”, IMF Occasional Paper 220, IMF,

1386

Washington D.C. 

1387

Prete, A. L. 2013. "Economic literacy, inequality, and financial development." Economics

1388
  • 118.1: 74-76.
1389

Pritchett. L. 1995. Desired Fertility and the Impact of Population Policy.  Population and Development Review.

1390

Pritchett, L. 1997. Divergence, big time.The Journal of Economic Perspectives,11(3), 3-17.

1391

Rajan, R. 2010. Fault Lines. Princeton University Press, Princeton, NJ (2010).

1392

Ramos, X., and Van de Gaer, D. 2016. Empirical Approaches to Inequality of Opportunity: Principles, Measures, and Evidence. Journal of Economic Surveys (forthcoming).

1393

Ravallion, M. 1998. Does aggregation hide the harmful effects of inequality on growth?. Economics Letters, 61(1), pp.73-77.

1394

Ravallion, M. 2001. Growth, Inequality and Poverty: Looking Beyond Averages. World Development, 29(11), pp.1803-1815.

1395

Ravallion, M. 2003. “Inequality convergence,” Economics Letters 80(3): 351–356.

1396

Ravallion, M. 2011. “Mashup Indices of Development,” The World Bank Research Observer.

1397

Ravallion, M. 2015. “The Luxembourg Income Study,” in F. Ferreira and N. Lustig, Special Issue, Appraising Cross-National Income Inequality Databases Journal of Economic Inequality 13(4): 

1398

Ravallion, M. S. Chen, and P. Sangraula. 2007. “New Evidence on the urbanization of global poverty,” Population and Development Review 33(4): 667-701.

1399

Ravallion, M., and S. Chen. 1997. “What Can New Survey Data Tell Us about Recent Changes in Distribution and Poverty,” World Bank Economic Review 11(2): 357-82.

1400

Rawls, J. 1971.  A Theory of Justice, Cambridge, MA: Harvard University Press.

1401

Rawls, J. 1999. A Theory of Justice (Revised Edition), Cambridge, MA: Harvard University Press.

1402

Ray, D. 1998. Development Economics, Princeton, NJ: Princeton University Press.

1403

Ray, Debraj. 2016. “Aspirations and the Development Treadmill.” Journal of Human Development and Capabilities. 17, 309-323, 2016.

1404

Reddy, S. G., & Pogge, T. W. 2002. How not to count the poor. World Bank.

1406

Reinikka, R., and J. Svensson. 2004. Local Capture: Evidence from a Central Government Transfer Program in Uganda. Quarterly Journal of Economics 119: 679-705.

1407

Rodriguez, F. 2004. “Inequality, Redistribution and Rent-Seeking”, Economics and Politics,Col. 16, No. 3.

1408

Rodríguez-Castelán, C. 2016. “The Distributional Effects of Cash Transfers”, unpublished paper, The World Bank.

1409

Rodríguez-Oreggia, E. and S. Freije-Rodríguez. 2009. “Evaluación de impacto sobre empleo, salarios y movilidad ocupacional intergeneracional del programa oportunidades”, in SEDESOL, Evaluación Externa del Programa Oportunidades 2008. A Diez Años de Intervención en Zonas Rurales (1997-2007). Mexico City: Secretariat of Social Development.

1410

Rodrik, D. 2007. One Economics, Many Recipes: Globalization, Institutions, and Economic Growth, Princeton, NJ: Princeton University Press.

1411

Roemer, J.E. 1993. A pragmatic theory of responsibility for the egalitarian planner. Philosophy and Public Affairs 22.2: 146–166.

1412

Roemer, J. 1998. Equality of Opportunity. Cambridge, MA: Harvard University Press.

1413

Roemer, J.E. 2004. “Equal Opportunity and Intergenerational Mobility: Going Beyond Inter-generational Income Transition Matrices,” in M. Corak (ed.), Generational Income Mobility in North America and Europe, Cambridge: Cambridge University Press.

1414

Roemer, J. 2006. “Party Competition Under Private and Public Financing: A Comparison of Institutions”, Cowles Foundation Paper No. 1191, Yale University.

1415

Roemer, J., 2008. “Equality of Opportunity,” in S. Durlauf and L. Blume (eds.) The New Palgrave Dictionary of Economics, London: Macmillan.

1416

Roemer, J.E. 2015. On the importance of circumstances in explaining income inequality. Revue Economique (forthcoming).

1417

Roemer, J.E., and Trannoy, A. 2015. Equality of opportunity: Theory and measurement. The Journal of Economic Literature (forthcoming).

1418

Roemer, J.E., and Trannoy, A. 2015. “Equality of Opportunity.” In: A.B. Atkinson, and F. Bourguignon (Eds.), Handbook of Income Inequality, Elsevier, New York, 217-300.

1419

Rogers-Dillon, R. 1995. The dynamics of welfare stigma. Qualitative Sociology, 18(4),

1420

439-456.

1421

Rosenfeld, R., Baumer, E. P., and Messner, S. F. 2001. Social capital and homicide. Social Forces80(1), pp.283-310.

1422

Ruffle, B. J. 1998) More is better, but fair is fair: Tipping in dictator and ultimatum

1423

games. Games and Economic Behavior23(2), 247-265.

1424

Rufrancos  H. G., Power M., Pickett, K. E., Wilkinson, R. (2013).  Income Inequality and Crime: A Review and Explanation of the Time–series Evidence . Sociology and Crimonology  1, pp.103.

1425

Runciman, W.G. (1972), Relative Deprivation and Social Justice: A Study of Attitudes to Social Inequality in Twentieth Century England, London: Pelican.

1426

Russett, B. M. (1964), “Inequality and insurgency: The Relation of Land Tenure to Politics,” World Politics, Vol. 16, pp. 442–454.

1427

Sahn, D. E., & Younger, S. D. (2007). Decomposing world education inequality. Cornell Food and Nutrition Policy Program, (187).

1428

Sala-i-Martin, Xavier. "The world distribution of income: falling poverty and... convergence, period." The Quarterly Journal of Economics (2006): 351-397.

1429

Samson, M., van Nierkerk, I. and Mac Quene, K. (2011), Designing and Implementing Social Transfer Programs. Cape Town: EPRI Press.

1430

Sanfey, A. G., Rilling, J. K., Aronson, J. A., Nystrom, L. E., & Cohen, J. D. (2003). The neural

1431

basis of economic decision-making in the ultimatum game. Science300(5626), 1755-

1432

1758.

1433

Sanfey, P., Teksoz, U., 2007. Does transition make you happy? Economics of Transition 15, 707–731.

1434

Savani, K., Stephens, N. M., & Markus, H. R. (2011). The unanticipated interpersonal and

1435

societal consequences of choice victim blaming and reduced support for the public

1436

good. Psychological Science22(6), 795-802.

1437

Savani, K., & Rattan, A. (2012). A choice mind-set increases the acceptance and

1438

maintenance of wealth inequality. Psychological science23(7), 796-804.

1439

Schady, N., and M. Araujo. 2008. “Cash Transfers, Conditions, and School Enrollment in Ecuador.” Economía 8 (2): 43–70.

1440

Schmader, T., Johns, M., & Forbes, C. (2008). An integrated process model of stereotype

1441

threat effects on performance. Psychological Review115(2), 336.

1442

Scholl, N., Klasen, S., and Vollmer, S. (2015): Re-estimating the Relationship between Inequality and Growth. University of Goettingen, mimeo.

1443

Schubert, B. and Huijbregts, M., 2006. The Malawi Social Cash Transfer Pilot Scheme: preliminary lessons learnt. Paper presented at the conference Social protection initiatives for children, women and families: an analysis of recent experiences. UNICEF, New York, October 30-31 2006.

1444

Schultz, T. P. 1994. “Human Capital, Family Planning, and their Effects on Population Growth.”  American Economic Review 84: 255-260

1445

Schultz, T.P.. 1997. “Demand for Children in Low Income Countries” in  M. Rosenzweig and O. Stark (editors) Handbook of Population and Family Economics.  Amsterdam: Elsevier.

1446

Schultz, T. P. (1998). Inequality in the distribution of personal income in the world: how it is changing and why. Journal of Population Economics11(3), 307-344.

1447

Schwartz, C. R. (2010). “Earnings Inequality and the Changing Association between

1448

Spouses' Earnings.” American Journal of Sociology, 115 (5), 1524-1557.

1449

Selden, M. and J-M. Wu (2011), “The Chinese State, Incomplete Proletarianization and

1450

Structures of Inequality in Two Epochs”, The Asia-Pacific Journal Vol 9, Issue 5 No 1,

1451

January 31, 2011. http://japanfocus.org/-Mark-Selden/3480

1452

Sen. A. (1979), “Equality of What”, The Tanner Lectures on Human Values delivered at Stanford University

1453

Sen, A. (1982) Real national income, in: A. Sen (Ed.) Choice, Welfare and Measurement, pp. 388–415 (Oxford: Basil Blackwell).

1454

Sen, A.K. 1992. Inequality Reexamined. Cambridge, MA: Harvard University Press.

1455

Sen, A. (2002), Rationality and Freedom, Cambridge, MA: Harvard University Press.

1456

Sen, A. (2000) ‘Social Exclusion: Concept, Application and Scrutiny’, Social Development Papers N° 1, Asian Development Bank.

1457

Shafir, E. (2014). Povery and Civil Rights: A Behavioral Economics Perspective.

1458

University of Illinois Law Review, 1, 205-229.

1459

Shafiullah, M., 2011. Foreign Aid and its Impact on Income Inequality. International Review of Business Research Papers 7: 91-105.

1460

Shah, A. K., Mullainathan, S., & Shafir, E. (2012). Some consequences of having too little.

1461

Science, 338(6107), 682-685.

1462

Shah, A. K., Shafir, E., & Mullainathan, S. (2015). Scarcity frames value. Psychological

1463

Science, 26(4). 402-412.

1464

Shariff, A. F., Wiwad, D., & Aknin, L. B. (2016). Income Mobility Breeds Tolerance for

1465

Income Inequality Cross-National and Experimental Evidence. Perspectives on

1466

Psychological Science11(3), 373-380.

1467

Sherman, D. K., & Cohen, G. L. (2006). The psychology of self-defense: Self-affirmation

1468

theory. In M.P. Zanna, Advances in Experimental Social Psychology38, 183-242. San

1469

Diego, CA: Academic Press.

1470

Shorrocks, A.F. 1978. “Income Inequality and Income Mobility,” Journal of Economic Theory 19 (2): 376-93.

1471

Shorrocks, A.F., and G. Wan. 2005. “Spatial decomposition of inequality,” Journal of Economic Geography 5: 59–81.

1472

Singh, A. 2011. “Inequality of Opportunity in Earnings and Consumption Expenditure: The Case of Indian Men,” Review of Income and Wealth 58 (10): 79-106.

1473

Slater, 2011. "Cash transfers, social protection and poverty reduction." International Journal of Social Welfare 20, no. 3 (2011): 250-259.

1474

Slaughter, S. and Rhoades, G. (2004) Academic Capitalism and the New Economy: Markets, State and Higher Education. Baltimore, MD, The Johns Hopkins University Press, 2004.

1475

Smeeding, T and J. P. Latner. 2015. “PovcalNet, WDI and ‘All the Ginis’: a critical review,” in Ferreira, F. and N. Lustig, Special Issue, Appraising Cross-National Income Inequality Databases Journal of Economic Inequality 13 (4):

1476

Soares, S., R.G. Osorio, F.V. Soares, M. Medeiros and E. Zepeda (2007), Conditional Cash Transfers in Brazil, Chile and Mexico: Impacts Upon Inequality. Brasilia: International Poverty Centre.

1477

Soares, F.V, Ribas, R.P. and Osorio, G.R. (2007), “Evaluating the Impact of Brazil’s Bolsa Familia: Cash Transfer Programmes in Comparative Perspective.”, IPC Evaluation Note.

1478

Sojo, Ana. 2014. ¨The Chilean system of contributory pensions as locus of rivalry and of a new social compact¨, ECLAC, Series Social Policy 211, Santiago.

1479

Sojo, Ana. 2015. ¨Including informal economy workers in contributory social protection: Current challenges in Latin America¨, International Social Security Review Vol. 68, 4/2015.

1480

Sojo, Ana (forthcoming). Protección social en América Latina: la desigualdad en el banquillo, Siglo XX editores, Buenos Aires.

1481

Sommers, P., and J. Conlisk. 1979. “Eigen Value Immobility Measures for Markov Chains,” Journal of Mathematical Sociology 6(2): 253-76.

1482

Spencer, S. J., Steele, C. M., & Quinn, D. M. (1999). Stereotype threat and women's math

1483

performance. Journal of Experimental Social Psychology, 35(1), 4-28.

1484

Srinivas, S. and J. Sutz ( 2008) ‘Developing Countries and Innovation: Searching for a New Analytical Approach’, Technology in Society, 30, 129–140.

1485

Statistics South Africa (2008), Income and Expenditure of Households 2005/2006: Analysis of Results. Pretoria: Statistics South Africa.

1486

Steele, C. M. (1988). The psychology of self-affirmation: Sustaining the integrity of the

1487

self. In L. Berkowitz, Advances in experimental social psychology21(2), 261-302. San

1488

Diego, CA: Academic Press.

1489

Stewart, F. 2002. “Horizontal Inequality: A Neglected Dimension of Development,” 5th UNU-WIDER Annual Lecture, Helsinki.

1490

Stewart, F. 2010. “Horizontal inequalities as a cause of conflict: a review of CRISE findings,” No.1, January 2010, Centre for Research on Inequality, Human Security and Ethnicity, Oxford.

1491

Stewart, F. (ed.) 2008. Horizontal Inequalities and Conflict: Understanding Group Violence in Multiethnic Societies, Basingstoke: Palgrave Macmillan.

1492

Stiglitz, Joseph. 2004. ¨Information and the Change in the Paradigm in Economics¨, in Michael Szenberg and Lall Ramrattan (editors), New Frontiers in Economics, Cambridge University Press.

1493

Stiglitz, J. (2012). The Price of Inequality: How Today's Divided Society Endangers Our Future. W. W. Norton & Company.

1494

Stiglitz, J. (2007) “Give prizes not patents”, accessible at: http://www.project-syndicate.org/commentary/prizes--not-patents, last accessed January 4, 2016.

1495

Stockhammer, E. (2015) Rising inequality as a cause of the present crisis.  Cambridge Journal of Economics 39(3): 935-958.

1496

Stutzer, A. 2004. “The role of income aspirations in individual happiness.” Journal of Economic Behavior & Organization. 54(1): 89-109.

1497

Suarez, M. et al., 2006. The Bolsa Família Programme and the tackling of gender inequalities. Report to Brazilian Ministry of Social Development and Fight Against Hunger (MDS) and DFID, Brasilia

1498

Subramanian, S. 2013. “Assessing Inequality in the Presence of Growth: An Expository Essay,” The Review of Black Political Economy, 42(3): 179-199.

1499

Subramanian, S. and D. Jayaraj. 2013. “The Evolution of Consumption and Wealth Inequality in India: A Quantitative Assessment,” Journal of Globalization and Development 4 (2): 253-281.

1500

Sutter, M. 2007. Outcomes versus intentions: On the nature of fair behavior and its

1501

development with age. Journal of Economic Psychology28(1), 69-78.

1502

Szekely M, Mendoza P. 2016. “Declining inequality in Latin America: structural shift or temporary phenomena? Oxford Development Studies 2016;44

1503

Tabibnia, G., Satpute, A. B., & Lieberman, M. D. 2008. The Sunny Side of Fairness

1504

Preference for Fairness Activates Reward Circuitry (and Disregarding Unfairness

1505

Activates Self-Control Circuitry). Psychological Science19(4), 339-347.

1506

Tanndal, J. and D. Waldenström. 2016. "Does Financial Deregulation Boost Top

1507

Incomes? Evidence from the Big Bang."

1508

Taylor, L. 2004.  “External Liberalization,   Economic  Performance  and  Distribution  

1509

in Latin  America and Elsewhere”, in Cornia G.A. (2005) op.cit.

1510

Taylor, J.E. 2015. Cash Transfer Spillovers: A Local Economy-wide Impact Evaluation (LEWIE). Poverty in Focus, Volume 11, No. 1.

1511

Teignier, M. and D. Cuberes. 2015.  Aggregate costs of gender gaps in the labor market: A quantitative excercise.  Journal of Human Capital (forthcoming). 

1512

Tendler, J. 2000. Why are social funds so popular. Local Dynamics in an Era of Globalization: 21st Century Catalysts for Development, 114.

1513

Tertilt, M. and M. Doepke. 2014. Does female empowerment promote economic development?  NBER Working Paper No. 19888.

1514

Teulings, C. and R. Baldwin. 2014. Secular Stagnation: Facts, Causes, and Cures.  Vox EU, 

1515

Te Velde, D. W. and O. Morissey. 2002. Foreign Direct Investments, Skills and Wage

1516

Inequality, mimeo ODI, London  

1517

Therborn, G. 2006. What Does the Ruling Class Do When it Rules: State Apparatuses and State Power under Feudalism, Capitalism and Socialism, London: Verso.

1518

Thorbecke, E., and Charumilind, C. 2002. “Economic Inequality and Its Socioeconomic Impact”, World Development Vol. 30, No. 9, pp. 1477–1495.

1519

Tilly, C. 2005. Identities, Boundaries, and Social Ties, Boulder, CO, USA: Paradigm Publishers.

1520

Torche, F. and Spilerman, S. 2006. “Household Wealth in Latin America”, UNU – WIDER, Research Paper No. 2006/114.

1521

Trapp, Measuring the labour income share of developing countries.  UNU-WIDER Working Paper 41/2015.   

1522

Tricomi, E., Rangel, A., Camerer, C. F., & O’Doherty, J. P. 2010. Neural evidence for

1523

inequality-averse social preferences. Nature463(7284), 1089-1091.

1524

Tyler, T. 2011. Procedural Justice Shapes Evaluations of Income Inequality:

1525

Commentary on Norton and Ariely (2011). Perspectives on Psychological Science6(1),

1526

15-16.

1527

Udry, C. 2006. Gender, agricultural productivity, and the theory of the household. Journal of Political Econom 104(5):1010-1046.

1528

UN. 2015. Transforming our world: the 2030 Agenda for Sustainable Development. New York: United Nations.

1529

UN. 2016. Report of the Inter-Agency and Expert Group on Sustainable Development Goal Indicators E/CN.3/2016/2/Rev.1*

1531

UN DESA. 2016. “International Migration Report 2013”, United Nations, New York.

1532

UNDP. 2010. Regional Human Development Report for Latin America and the Caribbean 2010: Acting on the future: breaking the intergenerational transmission of inequality. United Nations Development Programme. New York: UNDP.

1533

UNDP. 2013. Human Development Report for Latin America 2013-2014: Citizen Security with a human face.  New York: UNDP. 

1534

UNDP. 2015. Human Development Report 2015: Work for Human Development. Accessed online http://hdr.undp.org/en/2015-report

1535

United Nations Population Division. 2015. “World Population Prospects 2015’, New York, United Nations.

1536

United Nations Population Division. 2015. World Population Prospects Database.  New York: United Nations. 

1537

Van De Gaer, D., E. Schokkaert, and M. Martinez. 2001. “Three Meanings of Intergenerational Mobility,” Economica 68 (272): 591-537.

1538

Verhoogen, E.A. 2008. Trade, Quality Upgrading, and Wage Inequality in the Mexican Manufacturing Sector. Quarterly Journal of Economics 123(2): 489–530.

1539

Voitchovsky, S. 2005.  Does the profile of income inequality matter for economic growth? Journal of Economic Growth, 10, pp.273–296.

1540

Voitchovsky, S. 2009. Inequality and economic growth. The Oxford Handbook of Economic Inequality, Oxford University Press, pp. 549–574.

1541

Vos, R., J.A. Ocapo and L.Cortez. 2008. “Aging and Development’, Zed Books and the

1542

United Nations. London and New Yourk.

1543

Wakslak, C. J., Jost, J. T., Tyler, T. R., & Chen, E. S. 2007. Moral outrage mediates the

1544

dampening effect of system justification on support for redistributive social

1545

policies. Psychological Science18(3), 267-274.

1546

Wang, Y. Q. and K.Y. Tsui. 2000. “Polarization orderings and new class of polarization indices,” Journal of Public Economic Theory, 2 (3): 349–363.

1547

Weinhold, D. and Nair-Reichert, U. 2008. Innovation, Inequality and Intellectual Property Rights. World Development, 37(5), pp.889–901.

1548

Wimmer, A., L-E. Cederman and B. Min. 2009. “Ethnic Politics and Armed Conflict: A Configurational Analysis of a New Global Data Set,” American Sociological Review, 74 (2): 316–337.

1549

Winters, L.A., N. McCulloch, and A. McKay. 2004. Trade Liberalization and Poverty: The Evidence So Far. Journal of Economic Literature 42(1): 72-115.

1550

Winters, J. 2012.  Oligarchy, Cambridge: Cambridge University Press

1551

Wittenberg, M. 2015. “Problems with SWIID: the case of South Africa,” in F. Ferreira and N. Lustig, Special Issue, Appraising Cross-National Income Inequality Databases Journal of Economic Inequality 13 (4).

1552

Wolfson, M.C. 1994. “When inequalities diverge,” The American Economic Review 84 (2): 353–358.

1553

Wood, A. 1994. North-South Trade, Employment and Inequality, Clarendon Press, Oxford, United Kingdom.

1554

Woolard, I. and S. Klasen. 2005. “Determinants of Income Mobility and Household Poverty Dynamics in South Africa,” Journal of Development Studies 41(5): 865-897.

1555

World Bank. 1990, 2000. World Development Report.  New York: Oxford University Press.

1556

World Bank. 2001. World Development Report 2000/2001: Attacking Poverty. World Development Report;. New York: Oxford University Press.

1557

World Bank. 2002. World Development Report 2003.  New York: Oxford University

1558

Press. 

1559

World Bank.. 2003. World Development Report 2004: Making Services Work for Poor People.  Washington, DC: World Bank.

1560

World Bank. 2005. World Development Report 2006: Equity and Development. Washington, DC: World Bank.

1561

World Bank, 2011.  World Development Report 2012: Gender equality and development. Washington DC: The World Bank. 

1562

World Bank. 2016. World Development Report 2017: Governance and the Law. Washington, DC: World Bank.

1563

World Bank/FAO/IFAD (Food and Agricultural Organization/International Fund for Agricultural Development). 2009. “Module 4: Gender Issues in Land Policy and Administration.” In Gender in Agriculture Sourcebook. Washington, DC and Rome: World Bank/FAO/IFAD.

1564

World Development Report (WDR) 2006. “Equity and Development,” World Bank: Washington DC.

1565

World Health Organization. 2001. The World Health Report 2001: Mental health: new

1566

understanding, new hope. World Health Organization.

1567

WHO 2012. Environmental Inequalities in Europe, World Health Organization European Centre for Environment and Health, Bonn.

1568

Wright, E. O. 1979. Class structure and income determination.New York: Academic Press.

1569

Yamarik, S. and S. Ghosh. 2003. ‘‘Is Female Education Productive? A Reassessment.’’ Mimeograph, Tufts University, Medford, MA.

1570

Younger S.D, Myamba F., Mdadila K. 2016. “The Incidence of Fiscal Policy in Tanzania,” Commitment to Equity, Working Paper Series.

1571

Yusuf, Arief A., Sumner Andy and Irlan A. Rum. 2015. Twenty Years of Expenditure Inequality in Indonesia, 1993-2013. Bulletin of Indonesian Economic Studies. Vol. 50(2): 243-254.

1572

Zhang, X. and R. Kanbur. 2001. “What Difference Do Polarization Measures Make: An Application to China,” 37 (3): 85-98.

1573

Zanden, J. L., Baten, J., Foldvari, P., & Leeuwen, B. 2014. The Changing Shape of Global Inequality 1820–2000; Exploring a New Dataset.Review of income and wealth60(2), 279-297.

1574

Zheng, B. 2006. “Unit Consistent Decomposable Inequality Measures,” Economica 74 (293): 97-111.

1575

Zweimüller, J. 2000. Schumpeterian Entrepreneurs Meet Engel's Law: The Impact of Inequality on Innovation-Driven Growth. Journal of Economic Growth, 5, pp.185-206.

1576

[1] Affiliations : University of Florence, Italy; SEGIB, Spain; University of Göttingen, Germany; World Bank, USA; Tulane University, USA

1577

[2] Affiliations : University of Ghana; Ghana; Indira Gandhi Institute of Development Research, India; REPOA, Tanzania; ZEW Mannheim, Germany; The New School, USA; Princeton University, USA; CEPAL, Chile; University of Cape Town, South Africa

1578

[3] Affiliations : The New School, USA; OECD, Paris; University of Göttingen, Germany; IfW, Germany

1579

[4] An application of the non-parametric approach is to decompose an index belonging to the single-parameter entropy family of inequality indices based upon a circumstance variable, e.g. decompose the Theil index for income based upon father’s education. The contribution of the between-component can then be interpreted as a measure of inequality of opportunity for the particular circumstance variable. There are two different approaches: ex-ante (same circumstances, different outcomes) and ex-post (same effort, different outcomes) and there is an uneasy relationship between these two (see Fleurbaey and Peragine (2009) who describe and address this issue).

1580

[5] Common scales used are the modified OECD scales (which give a weight of 1 to the first adult, 0.5 to all other adults, and 0.3 to children below 16) or the LIS scales which use the square root of household size.  In developing countries, different scales tend to be used.  See Deaton and Zaidi (2003).

1581

[6] Based on the Income and Wealth Inequality Working Group of the High Level Group for the Measurement of Economic Performance and Social Progress.  See Alvaredo, Lustig, and Piketty (2016).

1582

[7] Other authors have studied the “missing rich” using Forbes magazine’s lists and other analogous initiatives.

1583

[8] (see e.g. Michael Mann (1986-2012, three volumes), The Sources of Social Power, Cambridge: Cambridge University Press, or Jeffrey Winters (2012), Oligarchy , Cambridge: Cambridge University Press).  

1584

[9] See e.g. Bowles, S. and A. Jayadev (2006), “Guard Labor”, Journal of Development Economics.

1585

[10] See for instance the seminal work of Meltzer, Allan H., and Scott F. Richard, "A Rational Theory of the Size of Government," Journal of Political Economy, 89 (1981), 914-927.  and subsequent work, e.g. by Rodriguez, F. (2004), “Inequality, Redistribution and Rent-Seeking”, Economics and Politics, Col. 16, No. 3.

1586

[11] Countering this view are scholars who worry about underconsumption and resulting stagnation due to low and shrinking purchasing power of the masses.  See, for example, Bleaney (1976) for a survey.  While this view as not had many adherents in recent decades, the recent debate about secular stagnation has revived this underconsumption as a potential cause of a growth slow-down.  See, e.g. Teulings and Baldwin (2014)

1587

[12]This theory is, however, at variance with Alesina and Rodrik (1994) who assume that the taxation finances mainly productive services by government.

1588

[13] David Castells-Quintana and Vicente Royuela (2014) as well as Erhart (2009) provide alternatives of structuring the different transmission channels between inequality and growth proposed in the literature.

1589

[14] See e.g. Runciman (1972) and the pursuant debate.

1590

[15] On which see e.g. Elster (1985), and Sen (2002)

1591

[16] On related issues, see e.g. Bardhan, Ghatak and Karaivanov (2007).

1592

[17] See e.g. Olson (1965) or applications such as Bates (1981).

1593

[18] On which see for instance Berg, Ostry and Loungani (forthcoming) and the underlying papers published by the authors at the IMF Research Dept.  

1594

[19] See Roemer (2006) among other relevant literature.

1595

[20] See e.g. the Pew Research Center results reported on http://www.people-press.org/2015/11/23/1-trust-in-government-1958-2015/.

1596

[21] This is due to lack of household surveys or tax records data for the period before 1950’s and more crucially lack of purchasing power parity estimates to convert incomes expressed in national currencies into single international currency prior to 1990s. 

1597

[22] Purchasing power parity (PPPs) are spatial indices that estimate the amount of local currency units that has the same purchasing power as that of one unit of the international numeraire currency.

1598

[23] This estimate is similar for all the three studies that calculate a measure of global inequality for this period (0.5 according to Bourguignon and Morrisson, 0.48 accoring to van Zanden, Baten, Foldvari and van Leeuwen (2010) and 0.54 according to Milanovic (2011)).

1599

[24] Given the large uncertainty and estimation errors from various sources in measuring global inequality this small change might likely be insignificant.

1600

[25] MLD and Theil are more sensitive to changes in bottom and top of the distribution as compared to Gini that is sensitive to changes in the middle of the distribution.

1601

[26] Data obtained from http://hdr.undp.org/en/composite/IHDI. Accessed September, 2016.

1602

[27] In fact, output collapse and increase in inequality was closely correlated in the early phase of the transition.  See Grün and Klasen (2001) and Ivashenko (2003)

1603

[28] Some of the differences can relate to different years used for the start and end of the period.

1604

[29] The welfare measure utilized here is per capita disposable income (that is, income after direct taxes and transfers) for Latin America and the Caribbean (SEDLAC), equivalized disposable income for OECD non-LAC countries (IDD), and per capita consumption for the rest (World Bank’s POVCAL). A caveat is in order.   The microdata for advanced countries distinguishes between income before and income after taxes and transfers in a clear way, this is not the case, however, for the income-based surveys in Latin America.

1605

[30] GCIP shows again the same trends, but higher levels of inequality in low-income countries, dominated by Sub Saharan African countries.

1606

[31] World inequality here is the average of unweighted Ginis. This it is different from the concept of global inequality which integrates all the individuals in the world (from all the available household surveys) under a single ranking.

1607

[32] See, for example, Bourguignon (2015) and Ravallion (2003).

1608

[33] Gasparini and Lustig, 2011; Lopez-Calva and Lustig, 2010.

1609

[34]Datt and Ravallion, (1992).

1610

[35] The decline of the Gini coefficient over 2002-2014 is robust to using household equivalized income, other inequality indicators (Theil index or the variations of the Kuznetz ratio) or data source (i.e. the Gini from CEPALStat or PovcalNet of the World Bank). The order of magnitude of the decline and the ranking of countries change, but the results are qualitatively the same.

1611

[36]Cornia (2014a).

1612

[37]See, for example, Lopez-Calva and Lustig (2010); De la Torre et al. (2012); Azevedo et al. (2013b); Cornia, (2014a); and Lustig et al. (2016). Microsimulation techniques suggest that a faster fertility decline observed among poor houseolds over 1990-2012 reduced the Gini coefficient by between 0.7 points in Chile and 2.0 points in Peru (Badaracco, 2014).

1613

[38]Azevedo et al. (2013a)

1614

[39] See, for example, Barros et al. (2010); Gasparini and Cruces (2010); Lopez-Calva and Lustig (2010); De la Torre et al. (2012); Azevedo et al., (2013b); Campos et al. (2014); Cornia,(2014a); Ferreira et al. (2014).

1615

[40] Lopez-Calva and Lustig (2010), Azevedo et al. (2013b), Cornia (2014a). Similar results were obtained for Argentina (Gasparini and Cruces, 2010), for Brazil by Barros et al. (2010), and for Mexico by Esquivel et al. (2010) and Campos et al. (2014).

1616

[41] Gasparini et al. (2011);,De la Torre (2012).

1617

[42] The role of noncontributory pensions cannot be disentangled because they are included in the total pensions (which account, on average, for 9 percent of the decline in overall per capita income inequality).

1618

[43] Atkinson, 2015. Kindle Locations 2364-2365 and 2391-2392. Harvard University Press. Kindle Edition.

1620

[45] This section draws closely from the World Bank’s World Development Report 2017 Governance and the Law.

The commenting phase has been concluded. We thank you for your thoughtful comments and responses.