Why is inequality so unequal across the world?

Could it be that every nation

gets the inequality it deserves?

José Gabriel Palma1

Faculty of Economics, Cambridge University

Paper to be presented at the Plenary Session “Shared Prosperity and Growth”

of the 17th World Congress of the International Economic Association, June 2014

1 Also, Facultad de Administración y Economía, Universidad de Santiago. Tony Atkinson, Stephanie Blankenburg, Ha-Joon Chang, Mariana Chudnovsky, Alex Cobham, Jonathan DiJohn, Jorge Fiori, Juliano Fiori, Samer Frangie, Jorge Friedman, Daniel Hahn, Geoff Harcourt, Javier Nuñez, Emily Hogan, Pamela Jervis, Jorge Katz, Mushtaq Khan, Juan Carlos Moreno-Brid, Cristóbal Palma, Guillermo Paraje, Carlota Pérez, Ashwani Saith, Claudia Sanhueza, Paul Segal, Ignês Sodré, Andy Sumner, Bob Sutcliffe, Lance Taylor and Robert Wade (among others) have made valuable contributions to my work in this area. Carlos Díaz-Alejandro and Andrew Glyn had significant influence on my thinking on this subject. Participants at many conferences and seminars, and current and former PhD students also made very helpful suggestions. The usual caveats apply.

2

“It’s becoming so outrageously expensive to be rich nowadays!!”

Quino (Argentinian cartoonist).

“Of all classes the rich are the most noticed

and the least studied”

John Kenneth Galbraith

“Inequality is a choice”

Joseph Stiglitz

“I am my choices”

Jean-Paul Sartre

The main aim of this paper is to take another look at differences in within-nation income distribution in the current era of neo-liberal globalisation. The emphasis will be on the study of middle-income countries with high degrees of inequality, especially those that have implemented full-blown economic reforms, such as those in Latin America and Southern Africa. I first examine how unequal is inequality across the world both in terms of overall inequality and that of different groups within each country; and then why there is so much diversity in terms of distributional outcomes across the world. From this perspective, as an index of income distribution, it’s becoming clear that the Gini has already served its purpose.

The key question I try to address is why is it that while political oligarchies all over the world would be only too happy to appropriate a huge share of the national income, only some (especially those in Latin America, middle-income Southern Africa, and recently by an increasing number in higher-income Sub-Saharan countries) seem able to get away with it? Finally, I discuss some elements of a new post-1980 distributive phenomenon that could be characterised by a trend in which it is the middle-income countries, such as those in Latin America, that now seem to be showing the more advanced ones “the image of their own future”. That is, many political institutions and distributive outcomes are indeed converging in this era of globalisation, but somehow unexpectedly they are doing so towards features that characterised more those in middle-income countries than those in developed countries during the period from FDR to the ‘70s — when income and wealth became more equally distributed. Then a now what really matters seems to be the share of the rich — and what they do with it!

1.- How unequal is inequality across the world?

Let’s start with a rather obvious point: using any measurement of inequality, what one finds across the world is that different political settlements and economic structures provide a remarkably wide variety of distributional outcomes (see Figure 1). Although this is a rather well-known fact, once you really start thinking about it, it becomes difficult to think about anything else.2 Not surprisingly, Ricardo said that the study of the distribution of income among the classes that contribute to the process of production (workers, capitalists and rentiers) is what economics is really about!

2 However, as Robert Wade remarks, there is still a strange neglect of income inequality in actual public policy in many countries in the world (Wade, 2014; more on this below).

3

FIGURE 1

● Highlighted countries are those of Latin America; the two countries with the highest Gini are South Africa and Namibia.

● In the case of regions, the statistic used to measure centrality is the median. Br=Brazil; Ch=Chile; Cn=China; EA1=Korea and Taiwan; EA1*=Hong Kong and Singapore; EA2=Indonesia, Malaysia and Thailand; EE= Eastern Europe; EU*=Mediterranean EU; EU=rest of Continental Europe; In=India; Is = Israel; LA= Latin America; NA=North Africa; No=Nordic countries; OECD-1=Anglophone OECD (excluding the US); Ru=Russia; SS-A=Sub-Saharan Africa; Tr=Turkey; US=United States; VN=Vietnam; and ZA*=South Africa (the actual value is 65.4).3 Unless otherwise stated, these acronyms will be used throughout the paper.

● For the sources of the data, see Appendix 1. Unless otherwise stated, these will be the sources of all figures in this paper.

Among the countless issues arising from this graph, there are two that stand out. One is that it confirms the huge range of inequality across countries c. 2012 — in this case, from two countries with a Gini below 25 to one close to 65 and another above that level. And oddly enough, the two countries at each end of the distribution, Slovenia and South Africa, are not that dissimilar from each other in terms of their degree of development. And the other is that middle-income (mineral-rich) Southern Africa and Latin America are clearly grouped at the wrong end of the inequality ranking.

3 If one uses the World Bank-WDI dataset (instead of the OECD’s one), South Africa’s Gini falls to (the still astonishing level of) 63.1.

4

2.- Inequality and income per capita

2.1.- Evidence from the Gini

If one analyses income distribution across countries in the traditional way — i.e., vis-à-vis income per capita — it becomes immediately evident that there is a huge distributional diversity among middle and high-income countries.4 See Figure 2.

FIGURE 2

● As in Figure 1, and BGD=Bangladesh; EE*= Eastern Europe with an income per capita below US$15,000; EE=those above that level; FSU*=Former Soviet Union with an income per capita below US$10,000; FSU=those above that level (excluding Russia); LA*= Latin America with an income per capita below US$8,000; LA= those above that level; SS-A***=Sub-Saharan Africa with an income per capita below US$650; SS-A**=those between US$650 and US$1,000; SS-A*= those between US$1,000 and US$2,000; SS-A=those above that level.5

● GDP pc=Expenditure-side real GDP per capita (PPPs) in 2011, based on PWT8.0. Unless otherwise stated, throughout the paper ‘US$’ will refer to this dollar. In this and following graphs, the range of the horizontal axis corresponds to the actual range of income per capita in the sample.

Although the horizontal ellipse of Figure 2 seems to indicate that the great majority of the regions/countries of the world have today, on average, a relatively similar income-

4 When I analyse income distribution across countries from the perspective of their income per capita I do so simply as one of many mechanism to visualise the geometry of within-country inequality across the world; i.e., it is just a cross-sectional description of cross-country differences in inequality, when categorised by income per capita. 5 In this and following graphs, ‘middle-income (mineral-rich) Southern Africa’ is proxied by South Africa, as this region only includes this country and Namibia. This is so both because the last reported data for Botswana only refers to 1994 (so it is not included in my sample; see Appendix 1), and because their close relatives in Angola and Zambia still do not qualify properly as ‘middle income’.

5

distribution (around, or just below, a Gini of 40), there are still large disparities in distributional outcome among middle and high income countries (see vertical ellipses). Moreover, the latest data coming out of the World Bank (WDI) also indicates an increasing distributional diversity among low-income Sub-Saharan African countries (from Mali and Burundi, with a Gini of 33, to Zambia with one of 58).

Also, since among middle-income countries one finds these days almost every possible distributional outcome, this Figure also casts serious doubt on the contemporary relevance of the traditional Kuznets’ hypothesis. Basically, the logic of the “Inverted-U”, if it ever existed, now seems to have evaporated — and with it, the phoney excuse used by many academics, politicians and business people in middle-income countries for their high inequality.

Among middle-income countries, there are two clear outliers. One is Eastern Europe (both EE and EE*), with lower levels of inequality — although in many of them their oligarchs are doing their best to change this. The other is Latin America (both LA and LA*) and (mineral-rich) middle-income Southern Africa, with extreme levels of inequality. Moreover, in all probability, many countries of the oil-producing Middle East for which there are no data would share the inequality heights of these two regions.6

In this respect, the key question is always the same: why is it that while political oligarchies all over the world would be only too happy to appropriate such a high share of the national income, apparently — i.e., at least in terms of the information found in household surveys — only those in Latin America and middle-income Southern Africa seem able to get away with it (followed these days by an increasing number in higher-income Sub-Saharan countries)?

Moreover, and perhaps ironically, this takes place especially in Latin America and South Africa, countries characterised by the consolidation of democracy — a process that has often been led by ‘centre-left’ political coalitions. From this perspective, the common thread in most them is that many economic and political institutions have changed in the recent past — some significantly — but the narrow interests of the élite clearly have not. In the case of Latin America, for example, the unique comparative advantage of its oligarchies seems to lie precisely in being able to use different institutions (sometimes quite astutely), and in being flexible enough to enlarge its membership to individuals coming from the above-mentioned ‘centre-left’ coalitions, in order to keep achieving their fairly immutable goals. In other words, few oligarchies in the world have shown such skills in their struggle for the “persistence of élites”’ despite significant institutional change. This brings us to the complex issue of “persistence and change in institutions”, and in particular to the so-called “iron law of oligarchies” — i.e., how dysfunctional institutions are sometimes so effective in creating incentives for their own re-creation (Acemoglu and Robinson, 2006).

In the case of Chile, for example, a recent study on tax returns (López, Figueroa and Gutiérrez, 2013) shows that the top 1% is able to appropriate — and in a ‘centre-left’

6 In 2003, I met by chance in Geneva a salesperson for one of the most exclusive watchmakers in Switzerland; in the conversation he mentioned that his wristwatches cost at least more than ten times an equivalent Cartier. When I asked who would buy such an expensive item (and one that was not particularly in good taste), he replied (somehow surprised at my question) “mostly people from your part of the world [Latin America] and the Middle East, of course”. And then he added that he was just back from a very successful trip to Argentina (even though this conversation took place a year after Argentina’s worst financial crisis in modern times). As the best Argentinian cartoonist said around that time, the problem for the oligarchies today is that “[i]t’s becoming so outrageously expensive to be rich nowadays!!” (“¡¡Es una vergüenza lo caro que se está poniendo ser rico!!”), Quino (2000; see epigraph to this paper). And regarding the lack of data for the Middle East, although Qatar is the exception, it is unlikely that its household surveys include the fate of the relatively large numbers of temporary migrant workers in activities such as construction or domestic services.

6

democracy — about one third of all income (32.8%); with the top 0.1% getting one-fifth (19.9%), and the top 0.01%, corresponding to individuals belonging to about only 300 families, getting more that one-tenth (11.5%) of the total. Some would argue that perhaps there is nothing that unexpected in these findings; however, it would be difficult to argue the same for high-income countries, where income distribution data from tax returns indicate that, in contrast to Marx’s prediction, it is the middle-income countries, such as those in Latin America, that now show the more advanced ones ‘the image of their own future’. As part of this new phenomenon, for example, it seems that now it is not the Latin American labour markets or tax structures that are attempting to catch-up with those of developed countries, but the other way round!7 As mentioned above, it seems that many political institutions and distributive outcomes are indeed converging in this era of globalisation, but somehow unexpectedly they are doing so towards features that characterised more those in middle-income countries than those of developed countries during the period from FDR to the ‘70s — when income and wealth became more equally distributed.

Thus, this type of tax-return-data not only shows that now in Chile the top 1% gets one-third of national income, but also that the income-share of its (now) close relatives in the US jumped from less than 10% before Reagan, to 24% at the beginning of the current financial crisis — and now is practically back to that level. In fact, the top 1% in the US captured just over two-thirds of the overall economic growth of real incomes per family over the period 1993-2012. Furthermore, “the share of the top decile in 2012 was equal to 50.4% of overall income, a level that was higher than in any other year since 1917 — even surpassing 1928, the peak of stock market bubble in the ‘roaring’ 1920s” (Sáez, 2013).8

2.2.- Peering into the Gini

The key issue I addressed in my 2011 paper is that the Gini, as a summary inequality statistic, is particularly obscure regarding some of the dynamics that are happening ‘inside’ each country’s distribution. And there are obviously important benefits to focusing on dynamics within those distributions. See Figures 3 to 7.

7 Surely Bush had an advisor from across the Rio Grande for the electoral fraud he engineered in Florida during the 2000 presidential election... And it was just the sign of things to come, as in the 2004 presidential election ‘the Latin-contagion’ went on as one-third of all votes were unverifiable, unauditable and unrecountable due to the paperless, direct-recording-electronic voting systems. In Alabama, for example, it was ruled that under state law anyone recounting the ballots would be subject to arrest (see http://www.thenation.com/article/how-they-could-steal-election-time). So, again practices from across the border, as in Mexico, by law, there is no recount, and votes have to be destroyed after each election (to make sure that recounts can never happen). 8 In the current one-sided scenario, where capital clearly has the total upper hand, a good deal of the civilisation brought to us by the vigorous economic, social and political struggles since the London Dock Strike of 1889, the Ford-T, the fear of contagion from the utopian ideals of the first ‘soviets’, the New Deal, the British National Health Service and the Welfare State seems to have ‘gone with the wind’.

7

FIGURE 3

Starting with D10, Figure 3 indicates that there is a particularly close correlation between the distributional geographies of the Ginis and that of the income-shares of the top decile. In turn, Figure 4 shows the same phenomenon for the regional distributional structures of the income-share of the bottom 40% (D1–D4) and that of the Ginis, but one in which the former is the mirror image of the latter (and of D10 above) — with Latin America and middle-income Southern Africa in a similarly iniquitous distributional world of their own, followed these days by Singapore and Hong Kong (EA1*) on the high-income side, Russia (Ru) on the middle-income one, and by some Sub-Saharan countries with an income per capita above US$ 2,000 (SS-A) on the low-income side.

8

FIGURE 4

It is therefore clear that the Gini scene for regional inequality (Figure 2) reflects accurately the distributional disparities at both ends of the distribution. But what about the rest, the other half of the population? Figure 5 shows one of the key contributions made in my 2011 paper: that the distributional picture changes completely when one looks at the 50% of the world’s population located within D5–D9 — the ‘middle and upper-middle classes’, sometimes called the “administrative classes” in institutional economics. Now the distributional geometry suddenly changes from one of huge diversity to one of a surprisingly similarity — and, basically, some of my critics would like us to believe that this is just a fluke!9

9 See, for example, Hazledine (2014); for my reply see Palma (2014).

9

FIGURE 5

● H-K=Hong-Kong; otherwise, all countries and regions as above.

Evidence from Figure 5 indicates two striking phenomena. One is the remarkable degree of homogeneity across regions/countries regarding the share of income that the middle and upper-middle classes are currently able to appropriate. This homogeneity in the middle and upper-middle is most prominent among rich countries — i.e., no more diversity here (Hong-Kong apart), as compared to the huge diversity we did find in the Gini and in the top and bottom deciles. Moreover, Eastern Europe (both EE and EE*) is no longer an outlier; and Latin America’s median country in both sub-groups (LA and LA*, currently at 49.4% and 49.3%, respectively) are about to join the ‘50/50 club’ — 50% of the population getting at least 50% of the national income. In fact, even Hong Kong is not that far away (48.8%), and Singapore (its partner in EA-1*) already qualifies (51.8%). Only South Africa, with Namibia, Zambia and most likely Botswana are living entirely in a world of their own.10

The case of South Africa is even more remarkable, as it has simultaneously the second lowest aggregate share for D5-D9 in the whole sample (Namibia is ‘top 1’) and the highest share in the world for one of its components: the ‘civil-service-crowded’ D9.11

10 The only other countries with a low share for D5-D9 in the whole sample (say, one below 47%), are Rwanda, Central African Republic, Guatemala, Honduras and Chile — for the (somewhat surprising) low share in Chile, see below. 11 As a result, after 20 years of democracy, the top quintal gets no less than 75% of overall income! In this country the drop in income-shares below D9 is so sharp that D8 already gets less than half the share of D9. For an analysis of this remarkable phenomenon, see Appendix 3 in my 2011 paper. Basically, democracy has not only opened up opportunities for new business people and professionals, but the ‘black empowerment’ policy has succeeded in both bringing a few new entrants into the top 1%, and a particularly large number of new entrants into a much enlarged administrative class — and in bringing them to the same high level of wages (and benefits) held previously by white bureaucrats when these administrative jobs were reserved for whites, mainly

10

As is evident in Figure 5, the current high degree of homogeneity in this group (D5-D9) is reflected in the fact that its measures of central tendency are almost identical: the harmonic mean is 51.8%, the average is 52.1%, the median is 52.5% and the mode is 52.6%. In turn, the coefficients of variation of the top 10% and bottom 40% are almost four times higher than the one of this group (see Table 1 below).

Furthermore, the current similarity in the income-shares of D5–D9 across countries is even more remarkable in the ‘upper middle’ 30% of the population (D7–D9). See Figure 6.

FIGURE 6

In this case, the homogeneity is even more extreme — Latin America has already made it (both LA and LA*), and also Hong Kong.12 Now the harmonic mean is 36.7%, its average 36.8% and its median is 36.9% — and the coefficient of variation is just one-fifth those of the other two groups at the tails of the distribution (difficult to believe that this is just another twist of fate).13

As I mention in the 2011 paper — and confirmed here with data for c. 2012 — it seems that a schoolteacher, a junior or mid-level civil servant, a young professional

Afrikaners (at the time of Mandela’s release from prison, one third of the economically active white population was employed in the public sector, with Afrikaners constituting the largest number of public employees). 12 The only countries with a share for D7-D9 below one-third, other than the three already mentioned (South Africa, Namibia and Zambia) are Rwanda and the Central African Republic. And the only ones with one that reaches 40% are China (40.3%, see graph), and Israel (40.4%). 13 Among the many issues that emerge from this homogeneity in the middle and upper-middle vs. the heterogeneity in the tails is the absurdity of reporting income distribution data only in quintiles, as the top one is the blend of two very different components — while D9 has the most homogeneous share of all deciles, D10 has one of the most heterogeneous one. Therefore, there is a major (and easily avoidable) loss of information if these tow deciles are reported together.

11

(other than economics graduates working in financial markets), a skilled worker, a middle-manager, or a taxi driver who owns his or her own car (London apart), all tend to earn at the moment the same income across the world — as long as their incomes are normalised by the income per capita of the respective country.

If such homogeneity across the world would also take place among the top 10%, at least for the Latin American oligarchies this would be analogous to communism!

Basically, in those middle income countries with huge inequality, such as those in Latin America, what is really happening is that while the top 10% has succeeded in a appropriating a level of income which in absolute terms is similar to those of their counterparts in rich nations (see, for example, Sutcliffe, 2001, and Milanovic, 2010), and the middle and upper-middle have done so (or are just about to do so) in relative terms (shares in national income), the bottom 40% has a very long way to go — as currently their income per capita is more akin to those of low-income Sub-Saharan Africa. The key issue here is that in so-called ‘middle income countries’ only those in the middle and upper middle could really be considered as being ‘middle-income’, as those at the top have already reached a ‘premature catching-up’ with their counterparts in rich nations, and those at the bottom 40% are still facing the challenge of a massive ‘catching-up’ — just to get to a ‘middle-income’ levels. “Convergence”, therefore, seems to be a far more complex phenomenon than is implicit in neo-classical models.

Another issue that is important to clarify immediately, as this aspect of my ideas has led to misunderstandings, is whether this homogeneity in the middle and upper middle implies that distributive outcomes are basically the end-result of a battle of the tails, from which the middle and upper-middle are somehow able to shield themselves. In other words, the key question in this respect is whether those in D5-D9 are simply spectators of a rough game played by the top and bottom struggling for their share of the other half — like Roman plebs enjoying the (distributional) circus from the safety of their seats —, or whether they are very much part of the distributional struggle (down in the arena, rather than up in their seats). As I discuss in detail in my 2011 paper, there is no ‘lack of history’ in how this group got there, and in how they have to struggle to keep it. That is, their current position is not the result of some sort of prophecy of the type ‘Thou shalt keep only half of the product of the sweat of thy brow; no more, no less’, but a somehow surprising phenomenon in an otherwise very diverse set of distributional outcomes — the roots of which have some components which are still a mystery.

In order to get their relatively similar shares, the middle and upper-middle — precisely due to their lack of distributional “insulation” — have had to form different types of political alliances to help them get (and defend) their half. And when these politico-economic alliances have broken-down, the share of D5-D9 has changed significantly. For example, while in Latin America the middle classes seek to defend their share of income with different forms of alliances with the élite, in India the “administrative classes” do so mostly via clientelist alliances with the poor (which gives them the political power to mediate in the different conflicts between the capitalist élite and the state).14 In turn, in South Africa the fortunes of the middle classes appear to be uniquely different as the dominant (redistributive) political alliance has turned out to be that between the new ‘empowered’ élite, the upper stratum of the new administrative classes and the bottom 40%. In political terms the bottom 40% may have gained a lot from this alliance, but in distributional ones this alliance has only succeeded in increasing the income-share of the top (both D10 and D9) at the expense of the middle.

What seems remarkable is that currently, and with very few exceptions as in South Africa, the middle and upper middle have been quite capable of getting their half in the distributional struggle, while at the same time being apparently powerless to increase

14 See, for example, Khan (2000).

12

it much further — more on this below in Appendix 2.15

In sum, what is crucial to remember is that the regional distributional structure currently suggested by the Gini only reflects the income disparities among half the world’s population — those at the very top and at the bottom of each country’s distribution. But it tells us little about the remarkable distributional homogeneity of the other half.16 This raises serious questions regarding how useful the Gini index is as an indicator of overall income inequality, especially because (from a statistical point of view) the Gini is more responsive to changes in the middle of the distribution. That is, the most commonly used statistic for inequality is one that is best at reflecting distributional changes where changes are least likely to occur! As a result, the overall geometry of inequality as shown by the Gini is likely to underestimate the level — and somehow distort the nature — of income disparities across countries.

The problem is that the most common alternative inequality statistics, those that have the advantage of being more responsive to changes at the top and bottom of the distribution — such as the Theil — tend to have the huge disadvantage of being extremely vulnerable to measurement errors precisely at the tails of the distribution (and, above all, at the top; see Paraje, 2004).

As a result, and given the homogeneity in the middle and upper middle, I suggested in my 2011 paper a new inequality statistic — one that simply indicates the ratio of the income-share of the top 10% over that of the bottom 40%. The obvious advantage of this inequality-indicator is that it measures inequality where inequality exists; it is also simple, intuitive, transparent and particularly useful for policy purposes — i.e., especially helpful for policy-targeting, as for anyone aiming at lowering inequality the implications of this ratio are as crucial as they are straightforward. In other words, the mere fact of its simplicity — i.e., one that purposely avoids all the unnecessary (and often counterproductive) algebraic sophistication of alternative inequality statistics — becomes its main strength.17 As it happened, in many quarters this simplicity also proved to be its main attraction. It is also better at highlighting the unique voracity of some oligarchies and, especially, their capacity to get away with it. It was later christened “the Palma Ratio” by Alex Cobham and Andy Sumner.18

Table 1 presents a set of statistics for the whole sample (131 countries), confirming what Bob Sutcliffe now likes to call “Palma’s Law”: the contrast between the homogeneous middle and the heterogeneous tails in the distribution of income.

TABLE 1 Measures of Centrality and Spread for Income Groups, c. 2012

H Mean Median Average st dev c of varD10 29.1 29.4 30.5 6.9 0.227

D1-D4 16.4 18.0 17.5 4.0 0.230

D5-D10 51.8 52.5 52.1 3.4 0.066

D7-D9 36.7 36.9 36.8 1.7 0.047 ● H Mean=harmonic mean; st dev=standard deviation; and c of var=coefficient of variation.

15 The next task should be to check whether this finding is corroborated by datasets that measure inequality using sources other than household surveys, such as tax returns or industrial pay (like the UTIP-UNIDO data set). 16 For a criticism of the Gini, a ‘19th Century’ statistics, see Cobham and Sumner (2013c). See also OECD (2013, especially Shepherd, 2013). 17 For example, according to Sen, the Theil “... is an arbitrary formula, and the average of the logarithms of the reciprocals of income shares weighted by income shares is not a measure that is exactly overflowing with intuitive sense.” (Sen, 1973: 36). 18 See Cobham and Sumner (2013a and b); see also Fisher (2013, especially the brilliant animation at the end of the article), Fisher (2014, map 7; and Green (2012). See also, Chang (2014).

13

Of all the statistics in Table 1, the coefficient of variation is the one that matters: it best shows the current distributional contrast across countries between the homogeneous middles and the heterogeneous tails — as already mentioned above, the coefficients of variation for both D10 and D1–D4 are nearly four times greater than that for D5–D9. Furthermore, they are five times larger than that for D7–D9. This suggests that regardless of the per capita income level of the country, the characteristics of the political regimes, the quality of their institutions, the economic policies implemented, the structure of property rights, or whether or not they belong to countries that managed to get their prices ‘right’, their institutions ‘right’, or their social capital ‘right’, the 50% of the population located in ‘D5–D9’ in most countries seems to have currently the capacity to appropriate in the distributional struggle about just over half the national income. This is even clearer in the case of those in D7-D9 regarding their just over one third of the pie.19

In turn, for the bottom 40% characteristics such as those mentioned above can make the difference between getting as much as one-quarter of national income (as in the Nordic countries and in some countries of Eastern Europe), or as little as a tenth — or even less, as in South Africa, Namibia and Honduras (6.4%, 8.2%, and 9.3%, respectively).

Finally, for D10 the sky is (almost) the limit, with some oligarchies managing to appropriate a share above 50% of national income (as in Namibia and South Africa), or close to it (as Zambia, Central African Republic and Honduras).

But as almost any summary statistics can hide some diversity among its members, let’s look at this contrast between the homogeneity in the middle and upper-middle, and the heterogeneity in the tails in the whole sample; see Figure 7.

19 Note that as far as the ex-communist countries are concerned, this became so only when in ‘full transition’ (see Appendix 2). Also, this homogeneity seems to be a group characteristic, as individuals within the group — as evidenced in household surveys — can easily be upwardly or downwardly mobile.

14

FIGURE 7

● Countries are ranked according to the income share of D1-D4.

It seems patently obvious that the huge diversity of distributional outcomes across the world is almost entirely due to different shares for the top 10% and bottom 40%.

3.- Income distribution and education: a more multifaceted relationship?

Among the many analytical issues that need to be re-studied following this stylised fact, this regional distributional homogeneity in the middle and upper-middle also seems to indicate that the relationship between ‘human capital’ and income distribution is probably more complex than the one usually assumed in several quarters of mainstream economics and UN reports.20 According to this theory, education — both in terms of equality of opportunities, and of better overall quality — is not just one variable in the determination of income inequality, but the crucial one. However, in all regions of the world (developed and developing; Latin American and non-Latin American), the top income decile is made up of individuals with relatively high levels of education, while those in the bottom four deciles have either relatively little schooling, or (in the more advanced countries), schooling of a very doubtful quality. So why do these two relatively homogeneously ‘educated’ groups (one homogenously ‘highly-educated’, the other homogenously ‘little-educated’) have the greatest distributional diversity across countries? In turn, if most of the world’s educational-diversity (both in terms of quantity and quality) is found among the population in D5-D9 — e.g. in terms of the share of the population with secondary and (especially) tertiary education — why does one find extraordinary similarity across countries in the shares of national income appropriated by this educationally highly diverse group?

Chile, for example, with a gross tertiary enrolment of no less than 71% (World

20 See, for example, Neal and Rosen (2000). See also ECLAC (2010a and b).

15

Bank-WDI, 2014), together with Argentina and Venezuela, has the largest tertiary education enrolment among all non-communist and non-ex-communist developing countries in the world. However, in 2011 the 30% of its population between D7-D9 were only able to appropriate the 7th lowest income share in the whole sample (33.7%) — with only Namibia, South Africa, Zambia, Rwanda, Central African Republic and Guatemala posting an even lower share. In fact, some countries with shares similar to Chile for D7-D9 have tertiary enrolment rates in single digits — e.g., Central African Republic with an enrolment of just 3.1%, and Rwanda with one of 7.1%.

And despite the fall in Chile’s overall inequality between 2003 and 2011 (the Gini fell from 0.55 to 0.51, while the “Palma Ratio” did so from 4.1 to 3.3), the share of D7-D9 remained invariant in its particularly low share. So, in terms of the rôle of education in the distribution of income, it is important not to lose sight of the multifaceted nature of the relationship between increased ‘equality of opportunities’, and of better quality in education, and increased distributional-equality in terms of income — and of the fact that education (or any other factor that may be influencing the distribution of income for that matter) can only operate within specific institutional dynamics. Needless to say, a lot more work is needed in this challenging area.

4.- Is the share of the rich what it’s all about? “The Palma Ratio”

Obviously, more research needs to be done on the forces shaping the income shares of different groups along such different paths — particularly in such opposite ‘centrifugal’ and ‘centripetal’ directions. Remarkably, this simple observation does not seem to have been emphasised before my 2011 paper. Moreover, it seems odd that most of the recent literature on income ‘polarisation’ has produced indices that emphasise distributional changes around the middle of the distribution, exactly where there is greater income-homogeneity. Wolfson, for example, started the whole ‘polarisation’ literature by developing an index that cuts the Lorenz curve right in the middle!21 In fact, the higher the degree of homogeneity in the middle and upper-middle of the income distribution, and the higher the degree of heterogeneity at the very top and bottom, the more statistically-meaningful simple income ratios (like that suggested by the ‘Palma Ratio’) become as indicators of distributional disparities across the world. In fact, following this logic, it could be argued that as the sum of all shares has to be equal to 100, the share of decile 10 could suffice as an inequality statistic for the whole distribution!

The same is true for the Gini, because as Tony Atkinson remarked in his comments on a draft of my 2011 paper, one interesting result of this homogeneity in the middle and upper-middle is that if D5-D9 gets half the income, then the Gini coefficient (in percentage points) is 1.5 times the share of the top 10% (in percentage points) minus 15. In this case the Gini has a maximum of 60% (although it may be larger on account of inequality within the groups, since this calculation linearises the Lorenz curve).

For this reason the subtitle in my 2011 paper stated that “the share of the rich is what it’s all about”. However, I still believe that the ratio of the two components that creates diversity in inequality (decile 10 and deciles 1 to 4) is a more informative statistic of inequality, as it highlights better the rôle of both ‘offending parties’. Figure 8 shows the inequality-ranking of the 131 countries in the sample according to the ‘Palma Ratio’.

21 See, for example, Wolfson (1997).

16

FIGURE 8

● Sources: as in Figure 1. Highlighted countries are those of Latin America and (mineral rich) middle-income Southern Africa. The last two, Namibia and South Africa, are — literally — off the chart!22

The most important stylised fact revealed by Figure 8 — and a phenomenon that was not that evident in the Gini-inequality-ranking in Figure 1 — is that inequality across the world, as measured by this ratio, increases first relatively slowly, and almost linearly, only to switch gear when Latin American countries enter en masse (around ranking 100); to increase then rapidly and geometrically.23 In fact, had the ‘steady pace’ found in the first 100 countries continued in the last quarter of the sample, the most unequal country in the world today would have posted a ‘Palma Ratio’ of about 3 — rather than one that is nearly three times as much! Inevitably, this lead to the same question yet again: why is it that only some can get away with such rapidly increasing levels of inequality?

And what about the historical trend of the shares of D5-D9? Has this ‘homogeneity in the middle and upper middle’ — the foundation of the Palma Ratio — always been there? Or have some ‘centripetal forces’ been at work within this group more recently? As this is a rather important matter, I look at this in detail in Appendix 2, where I conclude that there seems to be clear evidence in the data available so far for some sort of centripetal force homogenising the share of the middle and upper-middle; i.e., working towards a world where the ‘50-50 rule’ dominates in all regions.

22 Again, if one uses the World Bank-WDI dataset instead of the OECD’s one, South Africa’s ‘Palma Ratio’ falls to (the still dismal level of) 7.1 — in fact, since the Fall of Apartheid in 1994 and the beginning of democracy, inequality in South Africa has increased among all races and geotypes (see Leibbrandt, et al, 2010; and Palma, 2011). 23 The three Latin American countries ranked below 100 are Venezuela (72), Argentina (89) and Uruguay (94); however, in the first two, due to high and repressed inflation, it must be a great challenge for statisticians to process household surveys properly, and also to estimate implicit values for subsidies, and so on.

17

5.- Why so much emphasis in previous analyses on supposed changes in the middle?

With the above evidence, it is somehow perplexing to find that the early Washington Consensus’ explanation of high inequality, and the one it developed in the ‘90s, is somehow obsessed with looking at the middle of the distribution at the expense of the top 10%. As John Kenneth Galbraith once remarked in a different context, “of all classes the rich are the most noticed and the least studied” (see also epigraph) — a phenomenon that characterised most distributional theories until very recently (something that only changed thanks to the new data produced by Sáez, Piketty and Atkinson, among others).

For example, those of my generation will remember the ‘60s and ‘70s hypothesis concerning the supposed inequalising rôle of import-substituting industrialisation (ISI), due to the creation of “labour aristocracies” in high inequality regions such as Latin America. This hypothesis was recycled in the ‘90s in an attempt to explain the import-liberalisation-related “skill-biased technical change” proposition. But the myopic focus was always the same: in terms of diversities in inequality, what matters is the capacity to explain the supposed diversity among the share of the middle of the distribution across the world.

The first (and rather influential at the time) hypothesis aiming to explain high inequality via diversities in the middle (due to “labour aristocracies”), which was widely invoked by the emerging ‘Washington Consensus’ of the time, argued that one of the main causes of inequality in Latin America was the price distortions associated with ISI. These were supposed to have distorted the values of sectoral marginal productivities, allowing for artificially high wages in manufacturing (à la Stolper and Samuelson). That is, wage differentials were larger than if free trade had predominated.24 However, there was little then (as now) to differentiate Latin America from the rest of the world — developing and developed, ISI and non-ISI — in terms of the income shares among groups that would include ‘aristocratic’ and ‘non-aristocratic’ labour. Also, as the case of Chile indicates (Figure 9), there is little difference among these groups between the ISI and the post-ISI periods, even though both development strategies took particularly extreme forms in this country. See, for example, the remarkable stability of the ratio of ‘Q3/Q2’ — and even of Q4/Q2 — in the two periods (the pre-ISI 1957-1973, and post-ISI 1973-2010) in Figure 9.

24 See, for example, World Bank (1987) and Krueger (1983).

18

FIGURE 9

● 1=election of Allende; 2=Pinochet’s coup d’état; 3=the year Pinochet called a plebiscite seeking a mandate to remain in power for another eight years; 4=first democratic government (centre-left coalition) that took office in 1990 after Pinochet lost his 1988-plebiscite (and was forced to call presidential elections at the end of 1989); 5=second democratic government (same centre-left coalition, but a return to more ‘free-market’ distributional policies); 6-7 and 7-8=next two governments by the same coalition. 3-year moving averages. ● Source: calculations done by Pamela Jervis and myself using the FACEA (2012) database. Chile is one of the very few countries in the developing world (if not the only one) that has a relatively robust set of historical data for such a long period of time — at least for the ‘Greater Santiago’, where almost 40% of Chile’s population live.

In fact, both ratios, Q3/Q2 and Q4/Q2, end up in 2010 exactly where they started in 1957, even though trade policies, especially tariffs — an almost everything else, including technological sophistication, demand for skilled workers, financial ‘deepening’, labour market ‘flexibility’, and so on — could not have been more different. So, unless the first distortion (the ‘ISI-labour aristocracy’), was perfectly and immediately matched by the second (the ‘post-ISI-shortages of skilled labour’), the remarkable stability contradicts the above mentioned hypotheses.

The second proposition, constructed for the post-1980 liberalisation-cum-globalisation era, was basically a recycled version of the previous (distortions-due-to-ISI vs. an ideal world without ISI) approach. It tried to explain away the (supposedly) unexpected increase in inequality in many developing countries after the implementation of policies aiming at trade and financial opening (and ‘deepening’), and labour market ‘flexibility’ — and neo-liberal economic and political reforms in general. These increases in inequality, following greater integration into the world economy by ‘liberalised’ and ‘flexible’ economies, were the exact opposite of the unambiguous predictions made by of those circling around the ‘Washington Consensus’ at the time.25 Hence, it was argued

25 See, for example, Lall (1983).

19

that these (previously unforeseen) neo-liberal-reform-related increases in inequality took place because trade liberalisation had allowed for a substantial increase in the importation of modern capital goods with latest technologies embodied in them, which were intensive in the use of skilled workers (a scarce factor in most LDCs). As a result, this would have led (yet again) to an increase in wage differentials. However, as is becoming increasingly obvious, and as Figure 9 indicates for Chile (and had already happened in the ISI-period), what really differentiates Latin America’s inequality was not located where skilled and unskilled workers of the formal sector are likely to be located, but more towards the tails of the distribution of income. Then, even if import liberalisation did allow for the importation of capital goods which introduce new production techniques that made intensive use of skilled labour, evidence suggests that this does not alone account for much of the region’s increased inequality after economic reforms.26

In fact, as Figure 9 indicates (assuming that skilled workers — i.e., those able to handle new technologies — are located as high as Q4, while unskilled ones in the formal sector are probably relegated to Q2), during the initial post-ISI period in Chile the Q4/Q2 ratio did actually increase. This was the highly inequalising period between 2 to 3 in Figure 9, which was characterised by radical liberalisation-cum-flexibilisation policies. However, this period (to the surprise of many, particularly the Chicago-Boys) was actually characterised by low levels of investment (on average, just 16.9% of GDP for overall investment, both private and public), low levels of imports of new capital goods, little technological change, productivity stagnation (average growth of just 0.3% p.a.), and stagnant or even negative TFP growth (according to how one measures it; see Fuentes, Larraín and Schmidt-Hebbel, 2006). In turn, during the post-fundamentalistic-reform period — the “roaring” ‘90s (from 3 to 6 in Figure 9) — wage differentials between skilled and unskilled workers (proxied here by Q4/Q2) actually decreased from 3.1 to 2.5. And this was a period of high investment rates (high at least for Latin American standards — 24% of GDP), high levels of imports of capital equipment, rapid technological change, rapid productivity growth (4.8% p.a.), and particularly high TFP growth (see Ibid.). Even the ratio of the upper-middle (D7-D9) vis-à-vis that of the bottom 40% (not included in the graph) fell during this period of dynamic technological change and particularly high demand for skilled workers (from 4.1 to 3.1).

Therefore, if one wants to understand distributional diversities across the world, rather than continue to be distracted by ‘the middle’ of the distribution of income, and increased wage differentials (other than between those in financial markets and the rest), it would be more productive to focus on the tails — as the work of Sáez, Piketty, Atkinson and Stiglitz (among others) has done, as my 2011 paper tries to do — and the resulting “Palma Ratio” highlights — and as Piketty’s latest book (2014) does brilliantly again (with the exception of the already mentioned analytical factor). So, paraphrasing John Kenneth Galbraith, let’s not just notice the rich, let’s get on with the study of their behaviour!

Obviously, the ‘usual suspects’ won’t like it, but as long as the share of the middle and upper-middle remain relatively homogenous across countries, we have little choice but to keep reminding ourselves yet again of this basic distributional fact by sticking a note on our notice-boards saying: “the share of the rich is what it’s all about”.

6.- Why is inequality still so extreme in Latin America despite recent developments? And why is it getting even more extreme in South Africa?

When one wants to study extreme forms of inequality, there is only one place to start: Latin America. Why is it so extreme? is the million-dollar question — one which I can

26 Among those that favour this hypothesis, see for example Cline (1997, this book has a very useful survey of the literature). For critiques of this literature, see Krugman and Lawrence (1993), Atkinson (1997), and Paraje (2004).

20

only begin to address in this paper. However, there are some issues that may help scratch the surface of this phenomenon.

6.1.- Is there a ‘ratchet effect’ at play?

In Latin America there seem to be a tendency towards a ‘ratchet effect’ following distributional shocks; see Figure 10.

FIGURE 10

● From 1 to 8 and source, as in Figure 9.

● Black lines are harmonic means between the pre- and post-Pinochet periods (i.e., between 1957 and the coup d’état in 1973; and the return to democracy and 2010; this was the last year for which I was able to get these data).27

The evidence of the dataset available for this period of over half a century points in this direction. Chile’s ‘distributional ratchet’ seems to result from the fact that, at least in Latin America, improvements in inequality have tended to be temporal (e.g., from 1 to 2, and from 3 to 5), while deteriorations have tended to have more permanent effects (from 2 to 3, and from 5 to 6). That is, the well-known restrained ability of human processes to be reversed once certain things have happened seems to apply only to increases in inequality.

In the case of Brazil, although there are no data as systematic as for Chile, the evidence that exists (mostly thanks to Albert Fishlow; see for example Fishlow, 1972) also indicates that the huge deterioration of income that took place immediately after the 1964 coup remained in place almost unchanged for the next forty years, and has only marginally improved since — and it is still very far away from the levels of inequality

27 Remember that we have to stick to Chile owing to a lack of systematic historical data for other developing countries.

21

found in that country pre-1964. So, although Brazil is moving in the right direction (unlike South Africa), it has a long way to go as it still ranks today according to its “Palma Ratio” as the 122nd most unequal country in my sample of 131 countries (123rd according to the Gini). Too many bottles of champagne have been opened prematurely; and from a political point of view, as a result of the small improvements that have taken place, inequality has lost what for Wittgenstein was the key requirement needed for success in policy matters: a sense of urgency! As a result, inequality has dropped massively in the list of priorities — in fact, so far in the current presidential campaign (and even in the street demonstrations), it has been hardly mentioned directly as an issue...

The key issue here is that what has happened in Chile and Brazil in terms of their ‘ratchet effects’ makes me wonder how sustainable the recent relatively minor (although significant at the traditional, but somehow unambitious ‘α’) improvements in inequality in some countries of the region really are — as evidence suggests that those that benefit from reduced inequality are nowhere near as capable of retaining their gains as those that benefit from increased inequality are when the pendulum (almost invariably — at least in my part of the world) swings in the opposite direction.28

6.2.- Is the top 10% in Latin America — and in South Africa — unique?

When Tony Atkinson read an early draft of my 2011 paper, his first question was whether I thought that the top 10% in Latin America is just simply able to get a higher proportion of national income than most, or whether they are a different kettle of fish altogether. Part of the answer to that question can be found in Figure 11.

FIGURE 11

● Black crosses within the circles are D9/D1-D4 ratios — ZA# for South Africa; and Br# for Brazil. ZA*= South Africa’s actual ‘Palma Ratio’ is 8.5.

28 For a detailed analysis of this issue, see Appendix 1 in my 2011 paper.

22

Briefly, in my view, Latin America’s and middle-income Southern Africa’s distributional settlements are unique not simply because somehow the rich are able to appropriate a larger share of national income than their counterparts in other regions. They are unique because the rich there do not seem even to have proper counterparts elsewhere (except, as mentioned before, for countries in the oil-producing Middle East). So much so that if we substitute South Africa’s ‘Palma Ratio’ (D10/D1-D4) for its ‘D9/D1-D4’ ratio, this country would still rank as the 117th most unequal in the sample (instead of last); and Brazil would still rank as low as 53rd (instead of 122nd).29

Also, the already mentioned tax-return data confirm this ‘uniqueness’: in Chile the top 1% is able to appropriate 32.8% of national income, while the top 0.1% gets 19.9%, and the top 0.01% gets 11.5% of the total. One possible narrative for this phenomenon could borrow a metaphor from the Darwinian concept of “living fossils” — both in the sense that these oligarchies do not seem to have close ‘living relatives’, and that they appear to be similar to social and political “organisms” otherwise only known to us from the study of (social and political) fossils.30 In other words, these odd ‘species’ may only be in existence today because they are probably better equipped than oligarchies in other regions in the world to survive (and resist) major social and political evolutionary upheavals.31 Many economic and political institutions have changed (some significantly), but the underlying distribution of political power has not — and neither have the narrow interests of the élite. In fact, the unique comparative advantage of the Latin American oligarchies lies precisely in being able to use different institutions (sometimes quite astutely) to achieve their fairly immutable goals. As mentioned before, few oligarchies in the world seem to have such skills in their struggle for the ‘persistence of élites’ despite significant institutional change.

At the same time, Latin America’s and Southern Africa’s distributional outcomes are so extremely unequal that they seem to show little evidence of Pigou’s law of “diminishing marginal utility”, or “less intense wants”, at work, at least as far as income distribution (or status, power, or greed) is concerned — not much evidence of ‘diminishing returns’ here; not least because (as mentioned before) ‘it’s becoming so outrageously expensive to be rich nowadays!’ So, perhaps Adam Smith was closer to the mark when he said that as far as issues such as income distribution are concerned, “[...] it is the vanity, not the ease, or the pleasure, which interests us.” (1759). Vanity indeed.32

29 Note that getting rid of the top 10% in South Africa and Brazil here is just an analytical exercise, and not a policy proposal... 30 According to Darwin, “living fossils [...] like fossils, connect to [...] orders now widely separated in the natural scale. [... but] they have endured to the present day from having [...] been exposed to less severe competition”. (1859) 31 See also Arantes (2007); and Oliveira (2003). 32 The New York Times reported recently on a meeting with a Chilean businessman, describing him in the following way: “With his custom-designed Zegna suits, pink tie with matching Brioni handkerchief and colored diamond cufflinks [...] [he] boasted of having five Hummers, a private jet, a Caribbean island getaway, a wristwatch designed for him by Cartier at the request of Prince Albert of Monaco, even a Rolls-Royce Phantom Drophead convertible [for which] he paid $2.2 million [...]. [Also] he paid more than $400,000 to be the first South American to travel into space as part of Richard Branson’s Virgin Galactic tour next May. [...] He built a large home overlooking Santiago with 24-carat-gold-trimmed tiles in the swimming pool. He threw outlandish parties, including a 15th wedding anniversary celebration for 200 guests last November that cost $4 million and involved 600 entertainers, including Brazilian carnival dancers, and the musical acts Donna Summer and Air Supply. [...] he was now considering offers from companies to buy a majority of his mining assets [because] “I am not so happy working so much, it’s very stressful,” he said.” (http://www.nytimes.com/2010/11/20/world/ americas/20chile.html). As Nietzsche said, "vanity is the fear of appearing original: it is thus a lack of pride, but not necessarily a lack of originality”.

23

Nearly a century ago, José Ortega y Gasset found in his visit to the region that “[many in Latin America] had a narcissistic tendency to use reality as a mirror for self-contemplation” (1918). He was struck to find “so many self-satisfied individuals” — a phenomenon that for him was a major obstacle for progress, as “[...] human history is the product of discontent” (Ibid.). Perhaps there is no better way to summarise what is wrong with Latin America’s élites, current political settlements and distributive outcome than Ortega’s observations, as (for reasons beyond the scope of this paper) with the new ideological, political and economic paradigm these regional features have been revitalised with a vengeance.33

The uniqueness of Latin America’s and Southern Africa’s political settlement and distributional outcomes also becomes evident in Figure 12, when the ‘Palma Ratio’ is tested as the dependent variables against income per capita.

FIGURE 12

● The regression has 4 intercept dummies: Southern Africa (South Africa and Namibia), Eastern Europe (EE and EE* , line 5), Qatar, and the EA1* (Hong-Kong and Singapore). It also has three slope dummies (LA and LA*, lines 1 and 2, and Sub-Saharan Africa with and income per capita above US$2,000 — SS-A, line 3). ZA*= South Africa’s actual ‘Palma Ratio’ is 8.5. Line 4=base regression. All parameters are statistically significant at the 1% level (‘t’ statistics are based on ‘White’s heteroscedasticity adjusted standard errors’); i.e., pretty unlikely to have occurred by chance. The R2 of the regression is 67%. Regional dummies are reported only within the income per capita range of its members.

Figure 12 shows that despite the usual structural instability of this type of cross-country regressions, the result of such exercise produces a regression with highly significant

Also, the idea of ‘work being too stressful’ for many in the Latin American élite reminds me of Groucho Marx’s idea: "Money frees you from doing things you dislike. Since I dislike doing nearly everything, money is handy." 33 For an analysis of this issue, and of the main features of what I call the “Anglo-Iberian” neo-liberal paradigm, see Palma (2010).

24

parameters. Figure 12 also indicates that there is clearly no statistical evidence anymore for the ‘upwards’ part of the traditional “Inverted-U” hypothesis. That is, for the idea that posits that (for whatever reason) ‘things have to get worse before being able to get better’.

However, it is important to emphasise in this respect that this regression is simply meant to be a cross-sectional description of cross-country inequality differences, categorised by income per capita. That is, they should not be interpreted in a ‘predicting’ way, because there are a number of difficulties with a curve estimated from a single cross-section — especially regarding the homogeneity restrictions that are required to hold.34 This is one reason why the use of regional dummies is so important, as they can provide crucial information regarding the required homogeneity restrictions — and their evidence points in a different (heterogeneous) direction. Hence, regional dummies are reported only within the income per capita range of its members.

In the regional dummies there are two opposite paths. In one, for Latin America, but especially LA*, inequality gets, on average, slightly worse as countries have higher income per capita (lines 1 and 2), even though some countries in Latin America have already reached high middle-income level status. And increasingly unequal Sub-Saharan African countries (those with an income per capita above US$ 2,000; line 3) seem to be copying not only Latin American style football, but also other less pious features of the region… In turn, on the other path inequality gets, on average, systematically better in the base regression (line 4), and in Eastern Europe (EE and EE*, line 5). However, it is important to emphasise that these downwards slopes do not necessarily mean that the distribution of income within individual countries among these groups is currently improving as they get richer; it only means that although the distribution of income within many of these countries is currently deteriorating, it does so still in a way that does not change the fact that the richer the country the lower the level of inequality (as a group).

As the relationship between inequality and income per capita is not homogenous across regions and countries, the homogeneity restrictions that are required to hold for ‘prediction’ are visibly not fulfilled. For example, not only analytically but also statistically there is no reason to expect Latin America and Southern Africa to improve their remarkable inequality as their income per capita continues to increase simply because countries in other regions have done so in the past.

And, as is often the case, when work of this nature produces such statistically interesting results, this “[...] involves the evolution of knowledge as well as ignorance” (Krugman, 2000). That is, as mentioned before, the question that keeps recurring is why is it that only political oligarchies of middle-income Latin America and Southern Africa — and increasingly so in higher-income Sub-Saharan Africa — are able to get away with such degrees of inequality?

In particular, why is it that their shares in Latin America’s national income are so much higher than those of other middle-income countries — such as those found in North Africa, the former Soviet Union, Eastern Europe, Russia and the second-tier NICs (among others) — although the latter often have more markets rigidities; often have prices, institutions and social capital that are even less ‘right’; where property rights are often less well-defined and less well-enforced; where there is often even more educational segmentation; and where the educational systems for the poor are even more dismal; where there is even more gender discrimination; and even more shortages of skilled labour; and where there are even more ‘low-intensity’ democracies; and even more problems of ‘governance’; where success or failure in business depends even more on political connections and corruption, and so on?

In fact, the monotonous insistence of so many economists and politicians on blaming Latin America’s huge inequality on ‘exogenous’ factors is very much like using a

34 See Pesaran, Haque and Sharma (2000).

25

pair of scissors to cut an (analytical) knot that cannot be unravelled. Among the most recurrent ‘exogenous’ factors appearing in most of the relevant literature, we find the nature of new technology, the abundance of natural resources, market distortions resulting from erroneous policies or regulation (how could it possibly not be the fault of governments!), and the unfortunate institutions created at the start of the colonial past, half a millennium ago, such as the ‘mita’ and the ‘encomienda’ (an institution that was already pretty much gone by the end of the 16th century).35

As in so many areas, it is analytically far more productive to reject mechanical determinisms and the blaming ‘exogenous’ factors. Instead, let’s concentrate on individuals and societies’ ultimate freedom and responsibility. If we have the income distribution we do in Latin America it’s (basically) our choice.36 In Chile, for example, using Solt’s database (Solt, 2014), the Gini only improves 1.4 percentage points, or 3%, after taxes and transfers. In Finland, however, the Gini improves by 20 percentage points, or by 43% — 14 times more than in Chile. ‘Structuralist’, ‘exogenous’ or path-dependency explanation may help contribute to the understanding of a really complex (and probably ‘over-determined’) whole, but ultimately, as Stiglitz has stressed (see also epigraph), inequality is a choice (2013b).

Furthermore, according to the same author, the current choice of high levels of inequality is holding back the recovery (Stiglitz, 2013a).37 Also, as evident today particularly in middle-income countries, low wages (and the resulting inequality) is a subsidy for inefficient producers; as Churchill explained so well many years ago:

It is a national evil that any class of Her Majesty’s subjects should receive less than a living wage in return for their utmost exertions… Where you have what we call sweated trades, you have no organisation, no parity of bargaining, the good employer is undercut by the bad and the bad by the worst; the worker, whose whole livelihood depends upon the industry, is undersold by the worker who only takes up the trade as a second string… Where these conditions prevail you have not a condition of progress, but a condition of progressive degeneration. (Hansard HC, vol 155, col 1888; 24 April 1906).

In other words, inequality is not only a choice, but one that matters for more reasons than one…

35 Few phenomena have had so many explanations of the ‘exogenous’-type as inequality; some have even blamed Latin America’s huge inequality on the lack of major wars in the region — as supposedly in OECD countries (especially Europe and Japan), and in some of the first-tier NICs (namely Korea and Taiwan) income distribution is supposed to have improved only due to the horror and anxieties of major conflicts. Others, instead, keep insisting on looking at what happened in Latin America’s distant colonial past — as if ‘path-dependency’ had to be the inevitable recourse for every complex social and ideological process too complex to analyse (for an investigation that stretches the concept of path dependency well beyond its breaking point, see Sokoloff and Engerman, 2000; see also Cornia, 2012. For a view, which I endorse, that attributing Latin America’s current inequality to historical persistence is just a myth, see Williamson, 2009). 36 As a Cepal study shows (2010), in 6 countries in Latin America (Argentina, Brazil, Chile, Costa Rica, Panama and Uruguay), the total cost of a monetary transfer equivalent to ‘one poverty line’ (the cost per capita of two baskets of basic foods) to all the unemployed, all people over 64 years of age, and all children under 15 years living in vulnerable households, is only equivalent to between 1.8% and 2.7% of GDP. If this subsidy is given only to each child and adolescent between 5 and 14 years of age, the total cost is 1% of GDP or below for those six countries; and if the subsidy is only given to each unemployed person, the total cost is below 1% of GDP for thirteen of the sixteen countries studied — and just one-third of a percentage point of GDP or less, in eight countries — i.e., not such an insurmountable task! For an analysis of the ample scope that middle-income countries have to eradicate poverty, see Ravillion (2010); and Tregenna (2012). 37 See also http://www.ft.com/cms/s/0/753839f0-e75b-11e3-88be-00144feabdc0.html?siteedition=uk#axzz33TZAOSYR.

26

6.3.- On multiple stable equilibria. Why the note in our notice-boards should really read: “it’s the share of the rich — and what they do with it”

Income polarisation in Latin America and Southern Africa — no matter how extreme — only tells us half of the story. The other half is that (despite the huge share of national income appropriated by the top earners, abundant finance, fairly well-defined and enforced property rights, and ‘pro-market’ reforms) every time private investment in Latin America or South Africa manages to rise much above 15% of GDP, the capitalist élite starts experiencing feelings of vertigo.38 From this perspective, the most striking difference between these countries and fast-growing Asia is found in their contrasting relationships between private investment and income distribution (see Figure 13).

FIGURE 13

● LA=Latin America (a=Argentina; b=Brazil; cl=Chile; c=Colombia; cr=Costa Rica; d=Dominican Republic; e=Ecuador; mx=Mexico; p=Paraguay; pe=Peru; s=El Salvador; u=Uruguay; and ve=Venezuela); n-1=first tier NICs (k=Korea; and sg=Singapore); n-2=second-tier NICs (m=Malaysia and th=Thailand); n-3=third-tier NICs (cn=China; in=India; and v=Vietnam); US=United States (in 1980 and in 2010); P=Philippines; and za=South Africa.

● Sources: for the share of the top 10% as in Appendix 1 (except for the US, which is Alvaredo, Atkinson, Piketty and Saez, 2014 — this different source complicates the comparison of the US ratio with that of other countries in the graph). And for private investment data, the IMF-databank.

It is often acknowledged that the only historical legitimacy of capitalism — i.e., the legitimacy of a small élite to appropriate such a large proportion of the social product — rests on its capacity to use it productively (i.e., it’s capacity to develop society’s productive forces). And it can only do so by reinvesting — out of market ‘compulsions’

38 Kaldor (1959) was the first to discuss the contrast between Chile’s high share of profits in national income and the country’s low levels of savings and investment (see Marcel and Palma, 1989).

27

rather than Samaritan tendencies — most of that huge share.39 So, no other statistic seems to reflect so neatly the difference between Latin America’s ‘sub-prime’ capitalism and fast-growing Asia’s capitalism (one that, despite all its problems and contradictions, has a remarkable capacity for productivity-growth) than that in Figure 13. While in Latin America this ratio of private investment vis-à-vis the income share of the top 10% currently hovers around one third, in most of Asia it has a value of at least double that (e.g., Thailand), or even higher (with Korea’s ratio around 1). In turn, as Figure 13 indicates for the US, no other statistic seems to reflect so neatly how today it is the middle-income countries, as those in Latin America, that now show the more advanced ones ‘the image of their own future’.

From the perspective of Keynes, today Emerging Asia would be the equivalent to what was happening in continental Europe, especially Germany, between about 1870 and the First World War — i.e., during the ‘Third Technological Revolution’, or third great surge of development and industrialisation (Pérez, 2004); that of the ‘Age of Steel, Electricity and Heavy Engineering’, when the US and Germany overtook Britain:

Europe was so organised socially and economically as to secure the maximum accumulation of capital. While there was some continuous improvement in the daily conditions of life of the mass of the population, society was so framed as to throw a great part of the increased income into the control of the class least likely to consume it. The new rich of the nineteenth century were not brought up to large expenditures, and preferred the power which investment gave them to the pleasures of immediate consumption. [...] Herein lay, in fact, the main justification of the capitalist system. If the rich had spent their new wealth on their own enjoyments, the world would long ago have found such a régime intolerable. [...] Thus this remarkable system depended for its growth on a double bluff or deception. On the one hand the labouring classes [...] were compelled [...] into accepting, a situation in which they could call their own very little of the cake that they and nature and the capitalists were co-operating to produce. And on the other hand the capitalist classes were allowed to call the best part of the cake theirs and were theoretically free to consume it, on the tacit underlying condition that they consumed very little of it in practice. (Keynes, 1919).

Not much danger of finding these ‘puritanical-cum-Schumpeterian’ tendencies in capitalist élites in Latin America today, or in their re-emerging close relatives in the US or most of Europe either. The common factor among them is their tendency to spend the new wealth on their own enjoyments (including the casino-economy). In South Africa, in turn — in so many respects, Latin America’s honorary middle-income country in Africa — and in The Philippines (the honorary one in Asia), a similar low ratio for private investment as a proportion of the income share of the top decile indicates that their capitalist élites have that same ‘discreet charm’ of the Latin American bourgeoisie too — that preference for having their cake and eating it.

Clearly, in Figure 13 there are two different steady states: a transparent case of multiple stable equilibria — and the obvious question for Latin America and Southern Africa is how to get from their sub-optimal and fairly stable equilibria to the more dynamic one found in Asia.

It is fairly obvious that Latin America’s capitalist élites — and increasingly so their

39 As discussed in Foucault (2004); Khan (2005); Woods (1999); and Palma (2009), classical capitalism is characterised not just by the presence of market opportunities but by market ‘compulsions’, which ensure that both capitalists (of all sizes) and workers (of all skills) continuously have to strive to improve their performance in order to remain in the market. In no other economic system does continued existence depend on market competition, and therefore on the systematic improvement of labour productivity. Only in capitalism are there continuous pressures from competitive struggles, which lead to the constant improvement of the forces of production. Therefore, like in Alice in Wonderland, only in capitalism (when there is still proper competition!) is it necessary to run just to remain in the same place.

28

re-emerging close relatives in advance countries — seem to have a unique preference for sumptuous consumption and for accumulation via mobile assets (financial ones and capital flight) rather than via ‘fixed’ capital formation. However, on the positive side, easy access to mobile assets has at least helped Latin American and South African oligarchies to become more democratic (as it helps give them an ‘exit’ option; see especially Boix, 2003).

Finally, how can this Latin American sub-optimal equilibrium be stable over time, particularly within a democratic environment — and despite the fact that it is highly unequal and fairly undynamic? Figure 13 indicates that in part this is so because it is a rather pleasing environment for the capitalist élite — one with huge rewards with few market ‘compulsions’. But what about the middle classes and the bottom 40%? What is in it for them?

Starting with the bottom 40%, the key component of the answer is revealed in Figure 14: Latin America’s huge capacity for employment creation.

FIGURE 14

● LA=Latin America (a=Argentina; b=Brazil; cl=Chile; c=Colombia; cr=Costa Rica; d=Dominican Republic; e=Ecuador; m=Mexico; pe=Peru; u=Uruguay; and v=Venezuela); n-1=first tier NICs (k=Korea; tw=Taiwan; and sg=Singapore); n-2=second-tier NICs (id=Indonesia; m=Malaysia; and th=Thailand); n-3=third-tier NICs (cn=China; in=India; and v=Vietnam); oecd (eu=European Union; ir=Ireland; j=Japan; and US=United States, 1980 and 2010); and za*=South Africa (since 1994, the election of the first democratic government). Sub-Saharan countries are excluded because the GGDC dataset does not provide data on employment for this region, and the ILO database only provides econometric estimates.

● Source: GGDC (2014); for South Africa Quantec (2014).

As far as (gross) employment elasticities are concerned, post-1980 LA seems to live in a world of its own. In fact, as Figure 14 indicates, Latin American countries’ post-1980 (gross) employment elasticities are about twice as high as anybody else’s. However, and going against the expectations of many — other than in the Central American and Mexican ‘maquila’ industry (an industry that exists mostly due to artificially-created trade

29

restrictions in the US, which gave these countries preferential access to its markets) — there is little evidence that Latin America’s huge capacity for employment creation relates to export expansion following trade liberalisation. This is especially true in commodities. In fact, not only did employment in the primary sector decline in most countries (Brazil lost more than 2 million jobs in them), but also, with a few exceptions, there is no evidence that the jobs created in services are associated with the commodity boom in any significant way.

A sectoral analysis indicates that LA’s high elasticities are entirely due to services. For example, between 1980 and 2013 Brazil created 55 million jobs in all, of which about 80% were in services. That is, in Brazil employment in services grew about twice as fast as output in this sector. Furthermore, whatever the so-called ‘populist’ literature may suggest, there is no evidence that in the latter category these are mainly government jobs — until data are available, the employment elasticity of services increases even further when the latter sub-sector is excluded.

In fact, if I were to follow the traditional perspective of explaining inequality via wage differentials, I could argue that the mere fact that Brazil has created employment practically only in services since 1980 would explain a good deal its declining inequality. The reason being that wage differentials in services (at least of the kind created in Latin America) are lower than in manufacturing! So, huge de-industrialisation (see Palma, 2008) and high employment creation in services per se would reduce wage differentials. More importantly, as high incomes are now made more and more in financial markets — an activity that household surveys are particularly powerless to capture — recorded inequality could decline even more. There are, of course, other factors such as the increase in the minimum wage, higher formalisation of labour, greater social expenditure, and so on; but the above factors should not be forgotten.

So, where does the huge demand for services come from? An important component of the answer is that in middle-income countries that have joined (or are about to join) the ’50-50’ club for the middle and upper-middle — and have flexible labour markets and low levels of wages — there is a highly income elastic demand for cheap services. Jobs created may be precarious, mostly at minimum wages (at least when this legislation is actually enforced — something that has actually improved in Latin America recently), and in activities with little or no potential for long-term productivity growth, but at least they are jobs and there are plenty of them. Thus, Latin America has this unique capacity to have grown at an average of just 2.8% p.a. between 1980 and 2013, but was able to generate new employment at 2.7% p.a. (GGDC, 2014; see also ILO, 2014). That is, this region can uniquely combine very disappointing output-growth rates with single-digit unemployment rates (at least in the way we measure unemployment these days). In fact, ‘dynamic Asia’ has not been able to generate employment at the same rate even though output has grown nearly three times faster than in Latin America (or twice as fast as Latin America if we exclude China).

The contrast between Brazil and South Africa is revealing; since 1994 — the start of economic reforms in Brazil and of democracy and reforms in South Africa — the former has had a gross employment elasticity more than twice as large as the latter despite the fact that GDP growth has been almost identical. While the former created 32 million jobs between 1994 and 2013 (an increase of 45%), the South African economy only created 2.5 million (an increase of 20%). So, yet again, another ‘inverted-U: in high-middle income countries with low levels of wages, inequality tends to be associated highly positively with employment creation via high income elasticities for cheap services; however, this is so only until the point where further inequality begins to lower that elasticity, as the share of the ‘middle’ (in the case of South Africa, Q3 and Q4) is squeezed to such an extent that they cannot afford even these cheap services…

High shares for Q5 does create employment — in South Africa there are almost as many security guards as all manufacturing employment (1.2 and 1.4 million jobs,

30

respectively) — but weak demand for services from Q3 and Q4 does not help.

Therefore, although both Latin America and South Africa (despite relatively high transaction costs) have a large gap between their incomes per capita measured in current dollar terms and in PPP — reflecting relatively cheap prices for non-tradables — Latin America can at least get a lot of employment creation out of that gap.

With the only exceptions being a few fiscal paradises and a couple of ex-communist countries, according to the data from the World Bank there is no country in the world that has both such a high income per capita in PPP terms as Chile, and such a large gap between this measure and the one in current dollars (about 50%). So much so that Jim Yong Kim visited Chile to join in the celebration of the fact that the country’s income per capita had gone above the level required to be classified as ‘high-income’. In fact, in terms of PPP the country had gone above the magic $20 thousand mark. However, this 50% gap between the two measures of income per capita could also be understood in a very different way: as another indicator of inequality. Leaving aside transaction costs and other factors, the higher the level of inequality, the higher the gap between the two measures of income per capita. Thus, the persistence of such a gap in such high middle-income countries could be taken as another indicator of the persistence of inequality — of the persistence of backwardness.40 In other words, in the hypothetical situation in which Chile had the same income per capita in current dollars as now, but (somehow) were to engineer an even higher level of inequality (for example, by lowering the real value of the minimum wage), its income per capita in PPP terms could actually go up even further just because of that — as the price of many services would fall due to the fact that they tend to be a mark-up on those wages. Obviously, in such a case to interpret an increase in income per capita in PPP terms as an indicator of higher levels of ‘welfare’ becomes almost surreal.41

In Chile, for example, in the early 2000s, when the market for domestic servants (an occupation that accounts for 12% of female employment) became slightly tight, and meagre wages began to increase, the government (presided over by a member of the Socialist Party) immediately opened up immigration from Peru — many things are possible in Latin America, but middle classes being unable to afford domestic servants is not one of them! Although the massive influx of domestic servants from Peru that resulted from this measure had a huge positive impact in Chile’s culinary culture, it also made sure that the gap between the country’s income per capita when measured in current and PPP dollars would stay as high as it was. Had the government not taken this measure, it would have risked an increase in wages, a reduction of the gap, and the country falling below the magic US$20 thousand (PPP) mark… (A bit of magical realism in play).

Consequently, Latin America’s sub-optimal equilibrium is more stable than one would have expected in a democracy, as:

i).- it keeps the rich in a blissfully state (huge rewards with few market ‘compulsions’);

ii).- it allows the middle and upper-middle groups to have access to a particularly large variety of cheap services; and

iii).- it does at least provide high levels of employment for the bottom 40%.

One could even use the concept of a ‘not-very-dynamic Nash equilibrium’ here, making the tackling of inequality an even more complicated issue — not only would the rich not volunteer to help in this direction, but also the middle and upper-middle would be rather anxious regarding the risk of losing their easy access to cheap services. Even for the bottom 40% there is the uncertainty of increased unemployment (at least during the

40 See Palma (2013a). 41 This phenomenon also complicates the interpretation of statistics such as the purchasing power of currencies, and so on.

31

transition).

Figure 15 distributes countries and regions according to their rates of employment creation and productivity growth since 1980; this graph confirms this peculiar equilibrium — sub-optimal and relatively stable — unique to Latin America.

FIGURE 15

● W av = average employment and productivity growth in the world economy. LA=Latin America (br=Brazil; co=Colombia; cr=Costa Rica; ec=Ecuador; mx=Mexico; and ve=Venezuela; also ch1=Chile 1974-1986; ch2=1986-1998; and ch3=1998-2011); EE*=Easter Europe; N1=First-tier NICs (countries as above); N2=Second tier NICs; N3=third tier NICs; and za*=South Africa (1994-2011).

● Source: as in Figure 14; this graph again excludes Sub-Saharan countries due to lack of reliable employment data.

Allocating countries to these four quadrants, in the second one (quadrant 2 = productivity growth below world average, but employment creation above this average) one can only find countries from Latin America. And as the case of Chile indicates, when (for whatever reason) they do manage to accelerate productivity growth (in this case between 1986 and 1998), they find it difficult to sustain it — as if some form of law of gravity were pulling then back down to their ‘rightful’ position.42

As growth in Latin America since 1980 has been mostly factor accumulation (particularly additional employment), for most countries in the region TFP growth has been negative for this period (see ibid.).

One could argue that it is possible to break this sub-optimal ‘Nash equilibrium’ via,

42 For a more detailed analysis of this phenomenon — i.e., Latin American countries being at best ‘middle distance runners’ (and often just good sprinters), while many in Asia have became top ‘marathon-runners’ — see Palma (2010), especially how Chile in 1998, after a decade of fast growth, simply ran out of oxygen.

32

for example, efficiency wage-type synergies.43 But one should never underestimate short-term obstacles to that — including those from unexpected quarters. Although the analysis of how could it be possible in this region to increase productivity growth at the same time as reducing inequality and sustaining high employment creation is a subject well beyond the scope of this paper, suffice to say here that it is unlikely that the ‘invisible hand’ would be up to the job. And the well-rehearsed answer (repeated these days ad nauseam) that all that is needed is yet more of the same neo-liberal reforms sounds increasingly hollow.

7.- Why is inequality becoming so extreme in an increasing number of high-income countries?

7.1.- Piketty’s (neo-classical) explanation for increased inequality among advanced countries

It is really unfortunate that one who has made some of the most important empirical contributions to our capacity to measure income and wealth distribution more accurately, particularly the high share of the rich, is leading the pack in the wrong direction. In my opinion, this is so because he still insist in analysing this massive increase in inequality from the perspective of the neoclassical theory of factor shares (see Piketty’s otherwise excellent book, 2014) — the by-now-pretty-much-obsolete Solow-Swan 1950s analyses / an analytical tool that was developed when profits were made almost entirely in the real economy / a period pre-financial deregulation / pre-financialisation44 / pre-labour market ‘flexibilisation’ / a period when the Tobin’s ‘Q’ hovered around 1 / and when it was able to explain investment pretty well / and, crucially, a period when the share of profits was directly related to the stage at which high-income economies were in terms of their ‘technology-cycle’ (the “maturity stage” of their “Fourth Technological Revolution”, or fourth great surge of development and industrialisation— that of the Age of Oil, Automobiles and Mass Production; see especially Pérez, 2004). A theory in which what matters most — still today? — is the link between the capital intensity of production and the share of profits in total output. Furthermore, in this approach (whose ‘best before date’ is well gone) the nature of this link depends on the elasticity of substitution between capital and labour. As Piketty assumes in his analysis that this elasticity is greater than unity (against substantial evidence to the contrary), he concludes that the increase in the share of profits — and, as a result, of wealth-owners in national income — is due to an (unconvincing) increase in the capital-output ratio.

The latter results mostly because he fails to allow for factors such the role of financialisation and asset bubbles (including tax-free capital gains from housing), which have led to disproportionate increase in the market value of certain assets in the last three decades, as well as that of regressive taxation, or the increasing number of free-lunches that wealth-owners get from governments.45 As a result, Piketty concludes — not

43 On the issue of efficiency wages in LDCs, see Stiglitz (1976). 44 I understand for “financialisation” the rise in size and dominance of the financial sector relative to the non-financial sector, as well as the increasing diversification towards financial activities in non-financial corporations. 45 These include not only the well-know massive (and unconditional) bailouts when things go wrong, (following the famous “heads I win, tails you lose” new style-accumulation), but also those that result from the new logic of state-agency. The main aim of this was to reverse the post-FDR, post-war Keynesian-type state agency — one in which the key component of the interaction between political power and markets was a role for the state as a ‘constrainer’ of the rent-seeking practices of oligopolistic capital (in order to foster competition and rationalise financial markets). The de facto outcome of this transformation was a new role for the state as a facilitator of the rent-seeking practices of big business. Not only Bush asked polluters to write environmental regulation, but when Blair and Brown created a new regulatory body for the financial industry in the UK (the

33

just mistakenly, but surprisingly so — that the basic dynamic that has led to increased inequality is one of over-accumulation of capital in the real economy, while all evidence indicates that the opposite is the case: in the US and Europe since the neo-liberal reforms there has been too little real (non-residential) investment, and too much gains from financial speculation and fraud.46 On these issues, see Figures 16 and 17 for the post-1980 increasing decoupling between the real and financial worlds leading to rapidly rising inequality and (among other things) to the 2007/2008 global financial crisis.

FIGURE 16

● fin assets=stock of total financial assets (all sectors); and priv inv=private investment (excludes private inventories). Both series are expressed as percentage of GDP. 3-year moving averages.

● Sources: US Census Bureau (2014), and US Federal Reserve (2014); se also Palma (2009).

During the period of so-called “financial repression” that followed the Bretton Woods agreement in 1944 — one in which there was a deliberate attempt to create an economic environment that would help create a greater degree of linkage between financial and productive capital — total financial assets remained relatively stable as a share of GDP for about three decades (at a level of about 500%), while private investment experienced

Financial Service Authority – FSA), they set it not only as an “independent non-governmental body” (i.e., a company limited by guarantee), but one that was actually financed by the financial services industry and run by financial-industry insiders (ex-bankers became Chairperson and Chief Executive Officer). That is, they set the FSA as operationally independent of Government, funded entirely by the financial corporations it was supposed to regulate, and led by financial-industry insiders. I suppose this was an attempt to solve the problem of ‘regulatory capture’: if lobbyists inevitably succeed in capturing the regulators, why not make them the regulators in the first place? As Foucault envisaged, by projecting the logic of unregulated markets into the heart of government, the new framework resulted in “[a] state under the surveillance of the market, rather than a market under the surveillance of the state” (2004:120). 46 On this point, see especially the insightful paper by Rowthorn (2014); see also Palma (2009).

34

some acceleration (the two extreme points in the cycle were 13.8% of GDP in 1961 and 18.5% in 1979). The subsequent period of “financial liberalisation”’, a period of huge asset inflation (that more than doubled the value of total financial assets as a share of GDP) was accompanied by a slowdown of the rate of private investment (from 18.5% of GDP in 1979 to 15.5% in 2007). During this period the value of financial assets not only decoupled from the real economy, but the abundance of finance and the associated asset-price-led (not so) ‘irrational exuberance’, instead of having a positive pulling effect on private investment, had instead the effect of ‘friendly fire’.47 Therefore, there is not much evidence here to support the McKinnon and Shaw-type argument in favour of financial liberalisation — one of the most influential ideas behind the emergence of the Washington Consensus.

At the same time, the huge increase in inequality concentrated income at the top to such an extent — between 1980 and 2006 just the taxable income of the top 1% increased by nearly US$2 trillion, and that of the top 10% by US$3.5 trillion — that obviously became one of the major contributors (and one probably more important than the Asian ‘savings glut’) to the increased liquidity in the US financial markets (the abundance of which transformed financial markets into fundamentally fragile institutions, totally unable to self-correct).48 In fact, in 2007 no less than 3 million families in the US declared to the Internal Revenue Service (IRS) that they had liquid investable assets of more than US$1 million.49 So, the current global financial crisis may have many roots, but (as discussed in more detail in Palma, 2009) a crucial one relates to rapidly rising inequality. In particular, as Figure 17 indicates the current crisis has again shown that developments in financial markets are closely related to the distribution of income, so the latter is a crucial component in the understanding of the crisis and in the planning of how to get out of it.

47 According to Keynes, a greater degree of linkage between financial and productive capital was also essential for the recovery of the 1930s crisis: “[t]here cannot be a real recovery, in my judgment, until the ideas of lenders and the ideas of productive borrowers are brought together again. […] Seldom in modern history has the gap between the two been so wide and so difficult to bridge” (1931, p. 146; also quoted in Pérez 2002, 167). 48 On the recurrent nature of this factor in the lead up to financial crises, see Kindleberger (2000). On financial fragility, see especially Minsky (1992). 49 See http://www.nytimes.com/2009/08/21/business/economy/21inequality.html?_r=1&ref= business.

35

FIGURE 17

● fin assets=value of financial assets as percentage of GDP; and top 10%=income share of the top 10% (includes realised capital gains). 3-year moving averages.

● Sources: Alvaredo et. all. (2014), and US Federal Reserve (2014).

In sum, increased inequality (both income, and wealth) in high-income countries since 1980 — rather than being the outcome of too many Schumpeterian entrepreneurs rising the share of profits by investing an increasing share of GDP in real capital formation in order to create a growing number of new productive capacities, as Piketty seems to suggest — it has been mostly the result of the artificial creation of an economic and political environment in which (paraphrasing Oscar Wilde) anyone who wants to become rich by doing something socially useful simply lacks imagination.50

7.2.- Some elements of an alternative narrative regarding why is inequality is becoming so extreme

As mentioned above, and as the data coming out of “The World Top Incomes Database” indicates, these days it is the middle-income country that seems to be showing an increasing number of more advanced ones ‘the image of their own future’. Borrowing again the language of Darwinian evolution, in what could be one of the supreme political ironies of all time, Latin America’s “living fossils” may end up having the last (evolutionary) laugh, as the US’s oligarchy (and several of their other lost relatives) may be experiencing what in palaeontology is called a “Lazarus taxon”: an organism that, having disappeared from the fossil record, inexplicably reappears sometime later. Although it is true that rather than plantations now we have speculative finance, it is

50 Figure 17 also indicates that while pre-1980 the ‘American Dream’ did not just belong to the élite, it seems to have been hijacked by a rather tiny minority since — and for the rest, it has only been available on credit…

36

difficult not to feel that there is probably some resemblance between the present and what might have happened had the South won the Civil War...

The remarkable success of this 1% tends to confirm the hypothesis that neo-liberalism may well have become the most effective technology of power ever (Palma, 2009). Perhaps this ideology is just shorthand for ‘the art of getting away with such a remarkably asymmetric distributional outcomes within democracies’. Or, in the language of game theory, a technology of power capable of transforming a particularly asymmetric set of distributive strategic choices, and the corresponding payoffs, into a Nash equilibrium built around the ‘pure’ strategy of the élite. And in part this was achieved mainly by convincing the majority that there is no point in trying to challenge this Nash equilibrium while the all-too-powerful top income players keep their strategies unchanged. The latter is also true in Latin America, as despite recent minor improvements, what is particularly remarkable about neo-liberalism in Latin America is its capacity to still achieve extreme forms of inequality by means other than the ‘old-fashioned’ forms of social conflict resolution.

The key point here is that the élite in may middle- and high-income countries is now able to achieve this Nash equilibrium by far more imaginative forms of ideological conviction. So much so that one could even argue that there is no longer any need to threaten the majority in these countries credibly with the idea that they have too much to lose and little chance of winning by challenging the top player’s strategy. By ideologically convincing the majority that neo-liberalism is the only workable game in town (or, in Mrs. Thatcher’s terms, that ‘there is no alternative’ — TINA), this élite can now get away with such a remarkably asymmetric distributional outcome through a spontaneous consensus type of hegemony (in the Gramscian sense): a hegemony that is able to deliver such an unequal distribution of income through non-openly-violent means. In Latin America, for example, (with the exception of some Central American countries that insist on behaving like Banana Republics, such as Honduras) military regimes — the traditional hedge against a progressive distributional challenge by the majority — have become obsolete!

As it happened, in many developing countries (especially in Latin America and South Africa) the new process of legitimisation of capital has become so remarkably successful, and the new technologies of power so surprisingly effective, that neo-liberalism has been able to turn the tables on progressive forces and has become (‘low-intensity’) liberal-democracy’s best friend.

The main issue is that there is a big difference between the great majority in many middle- and high-income countries entering into such an unfavourable Nash equilibrium because they are faced with overwhelming odds against the likelihood of succeeding in challenging the ‘pure’ distributional strategy of the élite, and what is happening now when the majority seems to have entered into this Nash equilibrium mostly out of ideological conviction. Mrs. Thatcher was certainly right when she once branded ‘New Labour’ as "my finest creation". If this is the case, the distributional game has probably ceased to be one of ‘chicken’ (or ‘hawk-dove’) — at least for the time being. In fact, all evidence suggests that the great majority in both type of countries are now ideologically prepared to put up with such an unequal distributive outcome as if it was simply their lot in life.

Indeed, it could be argued that this component of the current ideology — its ingrained belief that ‘there is no alternative’ — synthesises the fundamental success of the ‘Anglo-Iberian’ neo-liberal discourse. In terms of its distributive angle, the post-1980 new attitude towards inequality was also best summarised by Margaret Thatcher: “It is our job to glory in inequality and to see that talents and abilities are given vent and expression for the benefit of us all” (quoted in Wade, 2014). Or, as Robert Lucas stated it clearly, “of the tendencies that are harmful to sound economics, the most seductive and […] poisonous is to focus on questions of distribution”. The same for Martin Feldstein: “[those who oppose increases in income at the top are] spiteful egalitarians”. (Ibid.)

37

This discourse resembles an argument put forward long ago by Callicles, a character in Plato’s Gorgias (dialogue): ‘it is natural and just for the strong to dominate the weak, and [… and] it is unfair for the weak to resist such oppression by establishing laws to limit the power of the strong’. In Callicles’ opinion (as in the neo-liberal critique of what was going on pre-1980), ‘the stronger, more aggressive and domineering by nature, has been defanged and domesticated by the legal institutions of the weak demos’.51

One issue here is that perhaps the ‘modern’ neo-liberal ideology may just be to a great extent recycled ideas from the past... Another, of course, is that (despite delusional fantasies of the ‘top’ 1%) the ‘strong’ are not so by ‘nature’ but by ‘environment’. This is the core insight of the Darwinian idea that a subset of members of a population may come to flourish relative to other members simply because they possess a feature, which others do not, that renders them relatively suited to some local environment. The question of the intrinsic worth of those who flourish most is not relevant to this story.52 What is crucial to the understanding of what neo-liberalism is really about is that it deliberately attempts to create a specific economic environment that is most suited to those features that capital has and others do not: in the jungle, capital is king! (And extremely mobile.) For having achieved this most unlikely of Nash equilibria in democracies by a spontaneous consensus type of hegemony, the Latin American élite — and that in the US, and in others of its re-emerging lost relatives — surely deserves an entry in the Guinness Book of Political Records.

And as soon as the 1% succeeded in convincing the majority of its ‘free-markets-supremacy-cum-trickledown’ distributional discourse, there could be only one outcome — as in ‘free’ markets (those with only self-regulation, little market discipline due to lack of competition, multiple artificially-created rents, lavish finance and a ‘sterilised’ state) there can only be one distributional winner. In turn, gains from multiple asset bubbles and easy access to an almost unlimited amount of credit may have helped confirm the ‘trickledown potentials’ part of the story — and help explain the relative stability of the ‘50-50 rule’ — and why it has facilitated enough popular support from the middle and upper-middle for the ‘free’-market-supremacy discourse.53 However, although this stable share may help to explain the ‘popularity’ of neo-liberalism among these groups, one should not underestimate others elements of the neo-liberal discourse that are also particularly appealing to them — such as the promotion of ‘order’ based on freedom, individual initiative, strong enforcement of property rights, sound macroeconomics, fighting paternalism vis-à-vis the poor, and so on.

These insights that emerge from an approach centred (among other factors, of course) on these three issues: i) the artificial creation of an economic environment that favours capital, ii) the financial-gains-and-easy-access-to-credit ‘trickle-down’ part of the story for the middle and upper-middle, and iii) the so far convincing ‘there is no alternative’ part of the distributional game are probably more productive workable hypothesis to explain the increasing levels of inequality since 1980 than the above-mentioned one used by Piketty (2014).

51 See http://www.classicallibrary.org/plato/dialogues/15_gorgias.htm. 52 See especially Lawson (2003). 53 Regarding the famous “disappearing middle “, one should never confuses the remarkable stability of the share of income of the middle and upper-middle in high-income countries with the level of welfare that that (stable) share can offer — given the massive increase in the cost of education, health, debts, and so on.

38

Conclusions

We have analysed the contrasting nature of the centripetal and centrifugal forces at work in terms of income distribution, as well as how only few oligarchies seem to be able to get away with the levels of inequality found in some middle-income countries — and more and more in high-income ones, in a changing world where increasingly the former seems to be showing the latter ‘the image of their own future’.

Taking a lead from Sartre’s ideas, one should always reject mechanical determinisms and ‘external’ factors — characteristic of so many explanations of the increasing inequality found currently across many regions in the world coming from different quarters — and insist on individuals and societies’ ultimate freedom and responsibility: “I am my choices” was his worldview (see also epigraph). “I am my freedom” was the key passage in one of his plays. It would be difficult to state more emphatically than this the dimensions of human freedom and individual responsibility. For him, every act is a self-defining one, and no act can truly be blamed on ‘external’ or ‘exogenous’ factors (see, for example, Sartre 1981).54

If we have the income distribution we do, it is because it’s basically our choice; any other ‘structuralist’, ‘exogenous’ or ‘path-dependent’ explanation may help contribute to the understanding of a really complex (and probably ‘over-determined’) whole but, as mentioned above, ultimately, and as Stiglitz has stressed, “inequality is a choice” (2013b; see also epigraph).

In other words, from the perspective of the recurrent debate between ‘structure vs. agency’, as far as income distribution is concerned I am clearly on the side of agency — although emphasising that this agency may well fail if it fails to understand structure! The huge diversity of distributional outcomes across the world seems to support this view.

This does not mean, as explained above, that one should underestimate the many obstacles facing a ‘progressive distributive agency’ in middle-income countries today, or in many high-income countries. The same is true — given the remarkable stability of the sub-optimal and highly unequal distributional Nash-equilibrium found in Latin America today — regarding the obstacles ahead in this region for it to be able to go much further in its current attempts to reduce inequality.

Of course these are issues as big as they come, because as someone famously said in his analysis of events in France in 1848, “people make their own history, but they do not make it as they please; they do not make it under circumstances they themselves have chosen, but under given and inherited circumstances with which they are directly confronted”. So, under these constraints, what makes us take particular choices? What helps form collective beliefs? How do spontaneous consensus types of hegemonies emerge?

And the fact that democracies get to be more and more something resembling the government of the 1% for the 1% and by the 1% (Stiglitz, 2011 and 2012) — and as the type of 1% found in the Nordic countries is a species in serious danger of extinction (if not already extinct) — the choices of an ever more weird minority, able to impose their will in the distributional struggle both by ideology and force, seem increasingly to be what really counts. And part of their success in being able to impose their own brand of choices on society rest on their continuous success in convincing the majority that under

54 A similar proposition is found in a speech in King Lear: “This is the excellent foppery of the world, that, when we are sick in fortune — often the surfeit of our own behaviour — we make guilty of our disasters the sun, the moon, and the stars: as if we were villains by necessity, fools by heavenly compulsion, knaves, thieves, and treachers, by spherical predominance, drunkards, liars, and adulterers, by an enforced obedience of planetary influence; and all that we are evil in, by a divine thrusting on: an admirable evasion of whoremaster man, to lay his goatish disposition to the charge of a star!” (1.2.132)

39

the current domestic and international constraints, the assemblage of the necessary social constituencies for progressive agendas is off the political map.

As is so often the case, Warren Buffett explains all this beautifully and succinctly: “There’s class warfare, all right, but it’s my class, the rich class, that’s making war, and we’re winning.”55

And in terms of ‘choice’, in the same way that more often than not “every nation gets the government it deserves” the same probably relates to income distribution. In turn — as Solt’s database shows (Solt, 2014) — the same could also be said for fiscal policy, one of the key instruments for dealing with inequality. From this point of view, Schumpeter’s assertion (1918) could not be more accurate, as the contrast between Chile and Finland above indicates, “The fiscal history of a people is above all an essential part of its general history”.

55 http://www.nytimes.com/2006/11/26/business/yourmoney/26every.html?_r=0.

40

Appendix 1: Sample

In this paper, in order to construct my sample (131 countries) I use the following sources:

i).- OECD (2014) for high-income OECD countries, and other non-Latin American

countries for which this dataset provides information (Czech Republic, Estonia, Hong-Kong,

Hungary, Israel, Poland, Russia, Singapore, Slovakia, Slovenia, South Africa, and South

Korea);

ii).- SEDLAC (2014) for all Latin American countries;

iii).- Taiwan (2014) for Taiwan; and

iii).- World Bank (2014; the WDI dataset) for the rest. In this source, I only included

countries with data after 2002 (as a result, Botswana, Trinidad and Tobago, Turkmenistan

and Zimbabwe were excluded). I also excluded countries with a population of less than 1

million (Belize, Bhutan, Comoros, Djibouti, Fiji, Iceland, Luxembourg, Maldives,

Montenegro, Saint Lucia, Sao Tome and Principe and Suriname).

iv).- For the historical analysis of the shares of D5-D9 and D7-D9 I also use the World

Bank-WYD database (Milanovic, 2002, and updates — the last available one refers only to

2005).

41

Appendix 2: Has the ‘50-50 rule’ been stable over time? Has there been a trend towards ‘convergence’ where change has taken place?

This is a rather difficult question to answer properly as reliable time-series data are scarce, and in the majority of countries simply non-existent. So, regrettably, for the time being there is not much we can say about this with certainty. Of the data we have, let’s start by looking at two of the few countries for which we have historical evidence — and two that are exact opposites in this respect — the US and my own country, Chile.

i).- The US and Chile

In the case of the US, at least as reported in the annual data of the US Census Bureau, there is a remarkable stability in time of the share of deciles 5 to 9, and especially 7 to 9. (Figure 18).

FIGURE 18

● 3-year moving averages (to highlight trends). P=percentile.

● Source: US Census Bureau (2014).

First, in the left-hand panel, there is an unmistakable trend of ‘stability in the middle vs. instability at the tails’ when the population of the US is divided in the way we have done so far. Furthermore, as the source reports data for the top 5%, it was also a great chance to divide the population in a slightly different way to that discussed so far: top 5%, an enlarged middle (D5–D9 plus P91–P95) and bottom 40%; left-hand panel). There are two remarkable features here. One is the changing fortunes of the top 5% and bottom 40%: starting in 1947 from a situation in which both received the same income-share — about 17% of the total —, by the mid-1970s the bottom 40% was getting three percentage points more. However, after 1980 the top 5% began their remarkable comeback (sometimes called the ‘revenge of the rentier’; see Palma, 2009), and ended up appropriating more than eight percentage points than the bottom 40% (21.3% vs. 13%). The other noteworthy feature is that the 55% of the population who make up this ‘enlarged middle and upper middle’ appropriates a remarkably stable share of income throughout (about two-thirds). In fact, the whole range in which the share of this 55% of the population fluctuates in this 64-year period is less than four percentage points of

42

income. Basically, this share first increased by 2.9 percentage points between 1947 to 1980, and then fell by 2.7 ones between then and 2012 — i.e., returning to where it started.56 As in so many other areas, in terms of income distribution the post-1980 economic reforms seem to have been initially just a ‘backlash phenomenon’, led by an ideological compass whose 'magnetic north' was simply the reversal of as many aspects of the previous development strategy as possible — and although this data indicates that for this group it stopped there, for the top and bottom of the distribution, once this process had started, it took a life of its own (and went much further).

So, as soon as we have data for a sufficiently large number of countries, I suggest we also try a Palma Ratio “v2”: the ratio of the top 5% vs. that of the bottom 40% (see Figure 19) — but for this new statistic to add significantly in terms of information from what we already have from the Palma Ratio “v1”, we will have to wait until household surveys become able to capture properly the income of the very rich.

FIGURE 19

● Source: as in Figure 18. 3-year moving averages.

There are no prizes for guessing the correct answer to the question: “What happened in 1980?”

In turn, the limited historical evidence we have on some particularly unequal developing countries, like Chile and South Africa, indicates a very different picture in terms of the stability of the share of the middle and upper-middle — not least because the only way that the seemingly unstoppable ‘centrifugal forces’ at the top of these middle-income countries can continue to operate once the bottom 40% has been squeezed almost out of existence (remember, in today’s South Africa it only manages to appropriate 6.4% of the overall national income), is by squeezing the middle and upper

56 In turn, the coefficient of variation of the share of this ‘enlarged middle and upper-middle’ for this period is just 0.02 — compared with 0.13 for the top 5%, and 0.10 for the bottom 40%.

43

middle. Also, and not surprisingly, political shocks in developing countries do take their toll on the share of this group — leaving short- as well as long-term effects in their wake. The case of Chile, a country that suffered a vicious political shock in 1973 is revealing. As noted before, Chile is one of the few countries in the developing world that has a relatively robust set of historical data (at least for the ‘Greater Santiago’, where almost 40% of Chile’s population live). See Figure 18 for the short- and medium-term effects of Pinochet’s rule after its coup d’état in 1973 (from 2 to 3 in the graph) — with an indication of its long-term ‘ratchet effect’ (already evident in Figure 10 above).

FIGURE 20

● Black lines are harmonic means between 1957 and the election of Allende in 1970 (49.4%); and between the return to democracy and 2010 (45.8%; 2010 was the last year for which I was able to get these data).

● From 1 to 8 and source, as in Figure 9. 3-year moving averages.

The key issue one has to understand is that among other factors, the outcome of the vicious cocktail made by brutal political repression, mixed with Milton Freedman and Friedrich Hayek’s proposition that “economic freedom” is what really matters (while political ones are just optional extras), adding Arnold Harberger and his ‘Chicago-Boys’ — and their many (self-)privatisations57 — and topped with a fundamentalist understanding of the Washington Consensus policies and plenty of artificially created asset bubbles, was a 51% increase in the income-share of the top 10%. This share jumped from 34.2% of national income (1973) to no less than 51.7% (1987). As James Bond could have suggested, an appropriate ‘family motto’ for the Chilean oligarchs (especially the new ones) might be “The World Is Not Enough” — as they tried to move into a distributional outer space propelled into dark matter by their special brand of neo-liberal ideology and insatiable greed. In order to do so, even though the bottom 40% had made substantial improvements under Allende, there was clearly not enough there to squeeze from. Not

57 See, for example, Monckeberg (2001), and Palma (2013b).

44

surprisingly, the middle and upper middle — ironically, Pinochet’s bedrock support — had to be their next target. And this group lost no less than 12 percentage points of their share in national income from the heights where Allende had left them in 1973 (even though they had been some of his fiercest opponents) — see from 2 to 3 in Figure 20. With the return to democracy (which was only possible because Pinochet lost his plebiscite to remain in power), this group recovered a significant part of their losses (even though a large segment of this group had supported Pinochet in his plebiscite to remain in power).58 However, as Figure 20 indicates, for D5-D9 the new distributional struggle seems to have settled at a steady state with a lower average than pre-1973.

So, for a Chilean of my generation it is hardly surprising to find that the middle and upper-middle classes in middle-income countries are as much part of the distributional struggle as the rich and the poor — ivory towers are reserved only for tenured academics! In fact, in my 2011 paper I devote part of an appendix precisely to discussing why in Chile the middle and upper middle have been politically so weak (vis-à-vis their counterparts in other parts of the world) both to defend themselves against Pinochet’s ‘flexibilisations’, as well as from benefiting fully from the return to democracy — after 24 years since the return to democracy, Chile’s D5-D9 now only manages to appropriate 45.5% of national income, the 8th lowest share in this 131 country-sample (only after Namibia, South Africa, Zambia, Rwanda, Central African Republic, Guatemala and Honduras, respectively).

However, what still needs to be understood properly both in developed and developing countries — fortunately since my 2011 paper some people have been doing serious work trying to understand this — is why D5-D9 often tend to succeed in the struggle to get their half, but seem quite unable to improve upon it. It is like a football team that has to fight really hard not to go down to a lower division (in a league totally rigged by the top 10%), but no matter how hard they try, they cannot improve their relatively precarious position. But what is patently clear is that if they don’t (or can’t) fight tooth and nail for their share, they will be relegated to a lower division — as happened in Chile. Give the oligarchs half a chance, and they will take it with pleasure — see Figure 21 for the ‘winner-takes-all’ type distributional game during Pinochet-cum-Chicago-Boys’ ‘modernisations’.

58 I would be great to hear an exponent of the ‘Rational Choice’ school for an explanation of this conundrum — one in which (as Gramsci had suggested) for D5-D9 their ideology was clearly more important than their pocket!

45

FIGURE 21

● Source: as in Figure 9.

It seems that not only there was an anti-Robin-Hood-type redistribution in operation —one in which the poor were robed by the rich — but also one where a more ‘post-modernist’ Robin Hood was also at play: one that robed the rich to give to the very rich!

ii).- Any further evidence on whether the ‘50-50 rule’ has been stable over time?

As far as the more ‘aged’ OECD countries are concerned, the OECD dataset indicates that, as in the case of the US above, the share of the middle and upper middle seems to have changed very little in the last quarter of a century (see Table 2).

TABLE 2 OECD (19) average c of v average c of v difference difference as %

1985 2010 1985-2010 of share in 1985D10 22.0 0.125 23.5 0.091 1.5 6.9

D1-D4 22.6 0.118 21.3 0.103 -1.3 -5.8

D5-D9 55.4 0.016 55.2 0.021 -0.2 -0.4

D7-D9 37.4 0.025 37.6 0.030 0.1 0.3 ● There are data for only 19 of the 23 high-income OECD countries for this period.

● Source: constructed using data form OECD (2014).

Even though the average change in D5-D9 is negligible, there was some movement around it; while in some countries there was a positive change, in others this was negative. However, in half of them their change in this 25-year period was equivalent to less than 1 percentage point of national income; and, with the exception of (rent-seeking-prone middle and upper-middle) Spain, in all the rest the change was below two percentage points. In all, the change was negative in 11 countries, and positive in the

46

other eight. This is in marked contrast with the other two groups, in which there was a clear trend: 18 of the 19 countries had a decline in the share of D1-D4 (only in Ireland remained stable), and in 16 there was an increase in that of D10 (only in Belgium, Greece and Spain there was either stability or a tiny decline). I think this contrast is in itself relevant as well; especially because in 1985 the average of those countries that had a fall in this share was higher than that of those with an increase — in particular among those where the change was greater than one percentage point. Therefore, without overstressing this point at this stage, it seems that these data begin to indicate the presence of some sort of ‘centripetal force’ at work leading towards ‘convergence’ (more on this below).

Regarding D7-D9, not only is the average change in this quarter of a century just 0.1 percentage points, but in 15 of the 19 countries it was less than 1; and in the other 3, just 2 (Japan) or lower. Moreover, these tiny changes followed the same pattern mentioned above (‘convergence’). So, we find a remarkably stable middle and upper middle in this group of countries during this quarter of a century; and it seems that there was logic to what little change there was. But, needless to say, what has happened in the more geriatric OECD cannot be automatically generalised to the rest of the (more restless) world.

Regarding Latin America, the SEDLAC dataset indicates a similar phenomenon of “convergence” — though more intense (Table 3).

TABLE 3 LA (18) average c of v average c of v difference difference as %

1990 2010 1985-2010 of share in 1990D10 40.7 0.133 37.5 0.127 -3.2 -7.9

D1-D4 11.1 0.226 12.7 0.142 1.6 14.4

D5-D9 48.1 0.068 49.7 0.063 1.6 3.3

D7-D9 35.8 0.052 36.5 0.051 0.7 2.0 ● Source: : constructed using data form SEDLAC (2014).

Of those countries that had a change in their D5-D9 share in this 20-year period, those with an increase had an average share in 1990 that was 2.7 percentage points lower than those with a decline. Furthermore, those that had an increase of more than three percentage points during this period had in 1990 an average share for this group that was no less than 6 percentage points below that of those with a decline of this magnitude. So, this dataset seems to indicate a much stronger trend towards ‘convergence’ for the shares of D5-D9 in this region.

And in D7-D9, although there was practically no change in the overall regional average (just a 0.7 percentage-point change), the initial difference between the two group-averages (those with an increase vis-à-vis those with a decline) was one percentage point in favour of the latter; and for those countries with a change greater than 2 percentage points (either way), it was 3.1 in the same direction. So, centripetal forces seem to have clearly worked among the Latin American upper-middle as well.

Is this the case for the rest of the developing world, too? As unfortunately the World Bank-WDI dataset is not entirely reliable for historical trends, it makes more sense to turn to another World Bank dataset that attempts to address this issue more systematically — the World Bank-WYD one (Table 4).59

59 The World Bank (2014 — WDI) has little or no data for the OECD countries and dynamic Asia (e.g. Korea or Taiwan); and for Latin America, its connected group (SEDLAC) provides more systematic data. Also, this database may be a great source for current income distribution in Sub-Saharan Africa, Eastern Europe and countries in the former Soviet Union, but has problems for

47

TABLE 4 57 countries average c of v average c of v difference difference as %

1988 2005 1988-2005 of share in 1988D10 29.4 0.309 31.0 0.246 1.6 5.4

D1-D4 17.7 0.314 16.9 0.287 -0.8 -4.5

D5-D9 52.8 0.082 52.0 0.059 -0.8 -1.5

D7-D9 37.3 0.059 36.9 0.035 -0.4 -1.1 ● Source: : constructed using data form Milanovic (2002, and updates).60 There are data for only 57 countries for this period.

Again, although there is little overall change in the average for D5-D9 (just a 0.8 percentage point decline in this 17-year period), the initial gap between the averages of the two groups in 1988 (those whose shares increased during this period, and those with a decline) was 4.3 percentage points again in favour of the latter. And for those countries with a change greater than 1.5 percentage points, it was 8.2 in the same direction. So, again, further evidence of a distributional ‘centripetal’ force at work for the middle and upper-middle.

In terms of regions, this dataset indicates that among high-income OECD countries (19), the overall average changes less than 1 percentage point, and apart from Switzerland, in no country the share of D5-D9 increased by more than 1 percentage point. In turn in Asia (10 countries), although there was no difference at all in the overall averages during this period, in those countries that this share changed more than 1.5 percentage points (either way), we find again an initial gap in their averages of 2 percentage points in favour of those that had a decline. In Latin America (12 countries), while the average for D5-D9 in the two end years was also almost identical (just a -0.1 percentage-point change), the difference in the 1988-averages between those two groups was 3 percentage points (again, in favour of those that had a decline). In Sub-Saharan Africa (only 8 countries), the overall average changed by just 1 percentage point; however, the initial difference in averages for the two groups was no less than 8 percentage points (in the same direction).

And among ex-communist countries of Eastern Europe and the former Soviet Union (just 6 of them), this time the overall average fell by no less than 5 percentage points during this period, and (not surprisingly) in all 6 countries there was a fall of their initial high share — initially, the average share of the D5-D9 group was as high as 59% of the total, but by 2005 (and a lot of ‘transition’) it had fallen to a more ‘respectable’ 54%.

Finally, some preliminary evidence for India (King and Rajan, 2013) seems to indicate that the homogeneity in the middle and upper-middle also hold at sub-national levels. They conclude that the proposition made in my 2011 paper also holds for Indian state- and district-level in 1986, 1995, and 2004:

... the fifth to ninth income deciles consistently account for roughly half of total

comparison in time. For example, it is difficult to believe that the share of the top 10% in Kenya fell by 15 percentage points in two years (from 47.9% to 32.8% of income). Or that in Côte d'Ivoire it first increased by 11 percentage points, only to fall again by 8 in a short period of time. Or that in Mauritania the share of D10 also fell by more than 14 percentage points in just three years. Or that in 2000s South Africa it first went up by 13 percentage points, only to fall again immediately by 6. The list is endless. So, it makes more sense to use this dataset for the latest available information than for historical trends, as recent data tend to be more reliable (due to household surveys becoming more accurate), and more comparable (as methodologies to homogenise them become more robust). 60 Note how in this table and those before, the coefficients of variation of D5-D9 and D7-D9 are, again, just a small fraction of those of D10 and D1-D4.

48

consumption expenditure; whereas the shares accounted for by the top 10% and bottom 40% each vary much more widely.

So, to summarise, as sufficiently reliable historical data are still very scarce, we simply cannot say with certainty how stable in time was the share of the middle and upper-middle across the world, and how strong was the logic of change when this occurred. However, the evidence we have so far shows that with the exception of ex-communist countries (where there was a downwards ‘correction’ towards the ’50-50 rule’), there was little or no change in terms of regional averages. Furthermore, in terms of changes in individual countries, there seems to be clear evidence in these data for some sort of centripetal force working towards a ‘50-50’ world for the middle and upper-middle in all regions — including ex-communist countries, when in ‘transition’. The little data we have show a quite systematic indication in this direction.61 Therefore, the fact that the middle and upper middle is so homogenous in current data is perhaps, at least in part, the outcome of some sort of a ‘time-convergence’ trend towards this ‘50-50 rule’.

How much would the bottom 40% living in middle-income countries wish for a similar phenomenon of ‘convergence’ for them! And how much the top 10% from my part of the world would think of this as a fate almost worse than death if this also were to happen to them.

Finally, developing countries indicate (not surprisingly) that political shocks — as happened in Chile, Eastern Europe, the former Soviet Union, and in many Sub-Saharan countries (and taking place now in the Arab world) — are most likely to have short and long-term distributional effects — and the latter, as it is clear in the Chilean and Brazilian cases, could easily take the form of an inequalising ‘ratchet effect’ — as those who benefit from reduced inequality tend to be less able to retain their gains as those that benefit from increased inequality.

61 What is needed now is to corroborate this findings with other datasets, such as the “UTIP Global Inequality Data Sets” (2014); the “World Top Incomes Database” (2014).

49

Bibliography

Acemoglu, D. and J. A. Robinson (2006), Economic Origins of Dictatorship and Democracy, Cambridge University Press.

Alvaredo, F., A. Atkinson, T. Piketty and E. Saez, (2014), ‘The World Top Incomes Database’, at http://www.parisschoolofeconomics.eu/en/news/the-top-incomes-database-new-website/.

Arantes, P. (2007). Extinção, Boitempo Editorial. Atkinson, A. B. (1997), ‘Bringing income distribution in from the cold’, The Economic

Journal 107(4). Cepal, (2010), La hora de la igualdad: brechas por cerrar, caminos por abrir, Cepal. Chang, H-J., (2014), Cline, W. (1997), Trade and Income Distribution, Institute for International Economics. Cobham, A. and A. Sumner (2013a), ‘On inequality, let’s do the Palma (because the Gini is

so last century)’, at http://oxfamblogs.org/fp2p/on-inequality-lets-do-the-palma-because-the-gini-is-so-last-century/.

Cobham, A. and A. Sumner (2013b), ‘Is It All About the Tails? The Palma Measure of Income Inequality’, Center for Global Development, Working Paper 343, September, at http://www.cgdev.org/sites/default/files/it-all-about-tails-palma-measure-income-inequality.pdf.

Cobham, A. and A. Sumner (2013c), ‘Putting the Gini back in the bottle? ‘The Palma’ as a policy-relevant measure of inequality’, at https://www.kcl.ac.uk/aboutkings/worldwide/initiatives/global/intdev/people/Sumner/Cobham-Sumner-15March2013.pdf.

Cornia, G. A. (2012), ‘Inequality Trends and their Determinants: Latin America over 1990-2010’, WIDER Working Paper No. 2012/09.

Cornia, G. A. and F. Steward, (2014), Towards Human Development: New Approaches to Macroeconomics and Inequality, Oxford University Press.

FACEA (2012), Encuesta Trimestral de Ocupación y Desocupación en el Gran Santiago’, available at http://www.empleo.microdatos.cl/encuesta_ocupacion/.

Fisher, M. (2013), ‘How the world’s countries compare on income inequality’, at http://www.washingtonpost.com/blogs/worldviews/wp/2013/09/27/map-how-the-worlds-countries-compare-on-income-inequality-the-u-s-ranks-below-nigeria/.

Fisher, M. (2014), ‘40 more maps that explain the world’, at http://www.washingtonpost.com/blogs/worldviews/wp/2014/01/13/40-more-maps-that-explain-the-world/.

Fishlow, A. (1972), ‘Brazilian Size Distribution of Income’, The American Economic Review, 62(1/2).

Foucault, M. 2004. Naissance de la Biopolitique: Cours au Collège de France, 1978–1979, Paris, Gallimard Seuil

Green, D. (2012), ‘It's the share of rich, stupid’: brilliant inequality stats and politics from Gabriel Palma, Oxfam blog at http://www.oxfamblogs.org/fp2p/?p=9964.

Hazledine, T. (2014), ‘Does the centre hold? Testing Palma’s proposition, Development and Change, forthcoming.

ILO (2014), KILM, at http://www.ilo.org/empelm/what/WCMS_114240/lang--en/index.htm.

Kaldor, N. (1959), ‘Economic problems of Chile’, El Trimestre Económico, abril-junio. Keynes, J. M. (1919), The Economic Consequences of the Peace, Harcourt, Brace, and

Howe, Inc. Keynes, J. M. 1931, ‘The Grand Slump of 1930’, In Essays in Persuasion, Macmillan. Khan, M. H. (2000), 'Rent-seeking as Process.' In Khan, M. and J. K. Sundaram, (eds.),

Rents, Rent-Seeking and Economic Development: Theory and Evidence in Asia, Cambridge University Press.

50

Khan, M. H. (2005), ‘The capitalist transformation’, in Jomo, K. S. and E. S. Reinert (eds.), The Origins of Development Economics: How Schools of Economic Thought Have Addressed Development, Zed Books.

Kindleberger, C. 2000. Manias, Panics, and Crashes: a History of Financial Crises, 4th edition, Wiley.

Krugman, P., and R. Lawrence (1993), ‘Trade, jobs and wages’, NBER WP No. 4,478. King, L and K. Rajan (2013), ‘An alternative measure of income inequality? The ‘Palma

Index’ and public health in India’, mimeo, Cambridge. Lall, D. (1983), The Poverty of ‘Development Economics’, Institute of Economic Affairs. Lawson, T. 2003. Reorienting Economics, New York, Routledge. Leibbrandt, M., I. Woolard, A. Finn and J. Argent (2010), ‘Trends in South African Income

Distribution and Poverty since the Fall of Apartheid’, OECD Working Papers, No. 101.

López R., E. Figueroa and P. Gutiérrez (2013), La ‘parte del león’: nuevas estimaciones de la participación de los súper ricos en el ingreso de Chile’, at http://www.econ.uchile.cl/uploads/publicacion/306018fadb3ac79952bf1395a555a90a86633790.pdf.

Marcel, M. and J. G. Palma( 1989), ‘Kaldor on the discreet charm of the Chilean bourgeoisie’, Cambridge Journal of Economics, Kaldor Memorial Issue, April. Also in T. Lawson, J. G. Palma and J. Sender, Kaldor's Contribution to Political Economy, Academic Press; and El Trimestre Económico.

Milanovic, B. (2002), True World Income Distribution, 1988 and 1993: First Calculation Based on Household Surveys Alone, The Economic Journal 112 (476),

Milanovic, B. (2010), The Haves and the Have-Nots: A Brief and Idiosyncratic History of Global Inequality, Basic Books.

Minsky, H. 1992. The Financial Instability Hypothesis, The Jerome Levy Economics Institute of Bard College, Working Paper No. 74.

Monckeberg, M. O. (2001), El Saqueo de Los Grupos Económicos Al Estado Chileno, Ediciones B.

OECD (2013), Ending Poverty, Development Co-operation Report, OECD. OECD (2014), at http://stats.oecd.org/. Oliveira, F. de (2003). ‘The duckbilled platypus’, New Left Review, 24. Ortega y Gasset, J. (1918), ‘Impresiones de un viajero’, Hebe, 5. Palma, J. G. (2008), ‘De-industrialisation, premature de-industrialisation and the Dutch

Disease’, The New Palgrave Dictionary of Economics, 2nd edition, Palgrave Macmillan.

Palma, J. G. (2009), ‘The Revenge of the market on the rentiers: why neo-liberal reports of the end of history turned out to be premature’, Cambridge Journal of Economics 33(4). An extended version is available at http://www.econ.cam.ac.uk/dae/repec/cam/pdf/cwpe0927.pdf.

Palma, J. G. (2010), ‘Why has productivity growth stagnated in most Latin American countries since the neo-liberal reforms?, in J. A. Ocampo and J. Ros (eds.), The Handbook of Latin American Economics, Oxford University Press. An extended version is available at http://www.econ.cam.ac.uk/dae/repec/cam/pdf/cwpe1030.pdf.

Palma, J. G. (2011), Homogeneous middles vs. heterogeneous tails, and the end of the ‘Inverted-U’: the share of the rich is what it’s all about, Development and Change 42(1). An extended version is available at http://www.econ.cam.ac.uk/dae/repec/cam/pdf/cwpe1111.pdf

Palma, J. G. (2013a), ‘Cómo fue que nos graduamos de país de “ingreso alto” sin salir del subdesarrollo’, at http://ciperchile.cl/2013/07/15/como-fue-que-nos-graduamos-de-pais-de-%E2%80%9Cingreso-alto%E2%80%9D-sin-salir-del-subdesarrollo/.

Palma, J. G. (2013b), ¿Cuánto habrá que esperar para que los Chicago Boys & Asociados respondan por el botín que algunos se llevaron?, at

51

http://ciperchile.cl/2013/09/12/%C2%BFcuanto-habra-que-esperar-para-que-los-chicago-boys-asociados-respondan-por-el-botin-que-algunos-se-llevaron/.

Paraje, G. (2004), Inequality, Welfare and Polarisation in the Greater Buenos Aires, 1986-1999, Ph D thesis, Cambridge University.

Pérez, C (2002), Technological Revolutions and Financial Capital, Elgar. Pesaran, H., N.U. Haque and S. Sharma (2000). ‘Neglected heterogeneity and dynamics in

cross-country savings regressions’, in J. Krishnakumar and E. Ronchetti, (eds.) Panel Data Econometrics — Future Direction, Elsevier Science.

Piketty, T. (2014), Capital in the Twenty-First Century, Harvard University Press. Rowthorn, R. (2014), ‘A Note on Thomas Piketty’s "Capital"’, mimeo, Cambridge. Sáez, E. (2013), ‘Striking it Richer: The Evolution of Top Incomes in the United States’, at

http://elsa.berkeley.edu/users/saez/saez-UStopincomes-2012.pdf. Schumpeter, J. (1918), ‘The Crisis of the Tax State’, International Economic Papers, 2(4),

1954. Sen, A. (1973), On Economic Inequality, Norton. Shepherd, A. (2013), ‘How do we get to zero on poverty – and stay there?’, in OECD

(2013). SEDLAC (2014), Socio-Economic Database for Latin America and the Caribbean, available

at http://sedlac.econo.unlp.edu.ar/eng/. Solt, F. (2014), ‘The Standardized World Income Inequality Database’, at

http://thedata.harvard.edu/dvn/dv/fsolt/faces/study/StudyPage.xhtml?studyId=36908&tab=files.

Sokoloff K.L. and S.L. Engerman (2000). ‘Institutions, factor endowments, and paths of development in the New World’, Journal of Economic Perspectives, 14(3).

Stiglitz, J. E. (1976), ‘The Efficiency Wage Hypothesis, Surplus Labour, and the Distribution of Income in LDCs’, Oxford Economic Papers, 28(2).

Stiglitz, J. E. (2011), ‘Of the 1%, by the 1%, for the 1%’, at http://www.vanityfair.com/society/features/2011/05/top-one-percent-201105.

Stiglitz, J. E. (2012), The Price of Inequality, W.W. Norton & Co. Stiglitz, J. E. (2013a), ‘Inequality Is Holding Back the Recovery’, at

http://opinionator.blogs.nytimes.com/2013/01/19/inequality-is-holding-back-the-recovery/.

Stiglitz, J. E. (2013b), ‘Inequality Is a Choice’, at http://opinionator.blogs.nytimes.com/2013/10/13/inequality-is-a-choice/?_php=true&_type=blogs&_r=0.

Sutcliffe, B. (2001), 100 ways of seeing an unequal World, Zed Books. Taiwan, (2014), National Statistics, available at http://eng.stat.gov.tw/mp.asp?mp=5. Tregenna, F. (2012), ‘What are the distributional implications of halving poverty in South Africa when growth alone is not enough?’, Applied Economics, 44(20). US Census Bureau (2014), at

http://www.census.gov/hhes/www/income/data/historical/inequality/. UTIP (2014), ‘UTIP Global Inequality Data Sets’, at http://utip.gov.utexas.edu/data.html. Wade, R. H. (2014), ‘The strange neglect of income inequality in Economics and public

policy’, in Cornia and Stewart (2014). Williamson, J.G. (2009). ‘Five centuries of Latin American inequality’, NBER WP No. 15,305. Wolfson, M. (1997). ‘Divergent inequalities: theory and empirical results’, RIW 43(4). Wood, E. M. (1999), ‘The politics of capitalism’, Monthly Review, 51(4). World Bank (2014), World Development Indicators. World Top Incomes Database, (2014), at

http://www.parisschoolofeconomics.eu/en/news/the-top-incomes-database-new-website/.