Inequality, myth and reality
Indian inequality is back in the news, thanks to James Crabtree’s excellent description, and analysis, of crony capitalism in India (Billionaire Raj). This book is a must-read for anyone wanting to understand the entrails of crony capitalism, in India or elsewhere. However, Crabtree also wants to paint over a larger canvas — crony capitalism is associated with an increase in inequality. The argument, I believe, is as follows: The rich are getting richer, and maybe the poor are becoming less poor, but the gap between the two is widening.
So far, the argument is unexceptionable. Whether empirically right or wrong, the argument that an increase in inequality accompanies fast growth has an intuitive appeal. But Billionaire Raj wants to go much further — that there was exceptionable increase in crony billionaires in India and that this was associated by an exceptionally large increase in income inequality in India. My comments, and analysis, are for the claim that there was an exceptionally large increase in income inequality in India. There wasn’t, as I point out below — what is ironic is that Crabtree does not need to show any widening inequality in order to make his justifiable case of crony capitalism running amok.
Only three income distribution surveys have been conducted in India. All three have been undertaken by the NCAER — in 1975, 1995 and 2004/5. The last survey was conducted along with the University of Maryland. This exercise (India Human Development Survey, IHDS) was repeated in 2011/12 on the same households as those surveyed in 2004/5. On the basis of all the available data, and the piecing together of various surveys, there is no evidence that income inequality in India increased between 1950 and 2003 — the Gini has essentially been constant at around 42. This popular measure of inequality has a value of 100 if one person has all the income and 0 if all have equal incomes.
The reputable (gold standard) consumption inequality surveys have been conducted by NSSO since 1950. These show declining inequality, and a modestly rising Gini since 1983. In 2011/12, the consumption Gini was 36, just four points higher than that observed in 1983. However, unlike advanced economies, India has a large difference in prices of goods across cities and villages, and across rich and poor states. If one takes the official Tendulkar poverty line (a proxy price index), then there is a 2 to 1 ratio between the price index of the poorest and the richest state. Adjusting for prices, the real Gini index increases from 30 in 1983 to 32 in 2011/12. In the annals of academia, a two-point change over 30 years is trivial, insignificant.
Evidence on increasing inequality certainly does not come from trends in consumption inequality. Where does it come from? In a much cited paper by Praveen Chakravarty and Vivek Dehejia (‘India’s Curious Case of Economic Divergence’) — Crabtree cites it approvingly as well — the authors contend that unlike other countries, India displays a divergence over time. That is, inter-state inequality has increased between 1960 and 2015. The statistical evidence? Very weak. For those so oriented, they achieve a cross-section R2 on only 0.08. They don’t report the statistical significance of their inequality (beta-convergence) variable. We are able to reproduce their results (R2 of 0.08) and obtain significance at only the 57 per cent level; if data for Jharkhand gets omitted, significance increases (22 per cent level) but no cigar. Net result: No evidence of increasing inequality.
Before we go further, let me emphasise that the only source of evidence for high inequality in India (and not increasing inequality) is the IHDS survey for 2004/5 and 2011/12. Income Gini for the two years are estimated as 49 and 51. Not adjusted by prices, this is an increase of nine Gini points between pre-2004 data and 2011/12. For consumption, we had observed an increase of four Gini points.
An IMF study for Asia shows India with a six-point Gini increase between 1990-2010, and for Australia, New Zealand, Japan, Taiwan, Hong Kong, and Indonesia, the average increase observed is around 4-5 Gini points. Not much different — then how is this increase in India an associate cause of crony capitalism, and not so in Australia or Taiwan?
When all else fails, those who want to believe that high growth is associated with increasing inequality cite the Kuznets curve. Simon Kuznets won a Nobel prize for hypothesising that as a country develops, inequality increases; when it reaches maturity, inequality declines — the celebrated inverted U-curve hypothesis. As I have tried to document in two books (Imagine there is no Country and The New Wealth of Nations), the Kuznets curve, like the invisible hand, is nowhere to be seen. Just to make clear, this negation of the Kuznets curve has been observed by many, many scholars.
There are two other “in search of large inequality increase” attempts to buttress the finding that crony capitalism causes (?) inequality increase. The first is a comparison with the robber baron episode in the US. While robber barons existed, and are (maybe) a fair comparison with crony capitalists, the US evidence on inequality increase is underwhelming. Indeed, available estimates suggest that income inequality barely changed in the US between 1870 (Gini of 47) and 1929 (Gini of 49). In the mid-1960s, US income Gini declined towards 40. To be sure, inequality has increased in the US over the last decade, and the Gini is now estimated to be around 55. There is only one study that does point to very high, and increasing, inequality in India. We have commented on this study before (Piketty Inequality in India: Fails Smell Tests, IE, January 20). In that article, I had pointed out that the Chancel-Piketty finding of both the level of inequality in India, and the change, defies credibility. Piketty finds (data from the WID website) that income inequality in India had an average Gini of 43.9 between 1951 and 1990; then averaged 45.8 1991-2000; and then catapulted to 60.6 in 2014 (last available WID data). The average 14-year increase in Gini reported by the Piketty WID database for non-oil countries is two Gini points; there are 389 such observations spanning the period 1870 to 2014. The largest 14-year change was observed for Poland in 2002, 16.3 Gini points. The second largest 14-year increase in Gini: According to Piketty, India had an increase of 13.2 Gini points in 2014. The Gini level in 2000, 2004, 2011, and 2014 were: 47.4, 51.3, 59.2 and 60.6.
Note the close correspondence between the IHDS survey Gini for India for 2004 and the WID estimate for the same year — 49 versus 51. The similarity is not co-incidental; the WID (Chancel-Piketty) estimates are “based” on the IHDS survey of 2004/5. After that, Chancel-Piketty are on their own, and don’t match the survey results that their method is ostensibly based on.
This article is a plea for objectivity in the analysis of the emotive and ideological subject of inequality. The readers, and the academic community, have to ask the choice Cassius asked about Caesar: “Upon what data (meat) doth this our WID (Caesar) feed that” it has become so acceptable?”