Inequality measures can be relative, absolute, or intermediate. A relative measure is one whose value remains unchanged when all incomes in an income distribution are raised or lowered by the same proportion. An absolute measure is one whose value remains unchanged when all incomes in a distribution are raised or lowered by the same absolute amount. The French economist Serge-Christophe Kolm observed that in a period of labour unrest in the late 1960s, French workers agreed to an across-the-board increase of 13% in all remuneration. It was only later that they realized that there was a specific sense in which this arrangement could increase inequality among wage earners. Imagine two employees with remuneration of 100 francs and 1,000 francs, respectively, before the wage hike. The difference in their remuneration is 900 francs. After the wage hike, their remuneration becomes 113 francs and 1,130 francs, respectively, and the gap between these increases from 900 francs before the hike to 1,017 francs after the hike!

In the presence of income growth, relative measures tend to behave like “rightist" measures, and absolute measures like “leftist" measures, as Kolm put it. He, therefore, saw the case for more “moderate", “centrist/intermediate" measures, which register an increase in value when all incomes in a distribution are raised equi-proportionately, and a decline in value when all incomes are raised equally. These are what one may call income-centrist measures. An analogous problem is encountered with a reckoning of inequality in the presence of population changes. Most extant inequality measures are population-relative, in the sense that when the numbers of persons at all income levels are raised equi-proportionately, the value of the inequality index remains unchanged. In contrast, a population-absolute index would register a k-fold increase in its value for a k-fold replication of the population at each income level. Population-intermediate measures, which avoid the “extreme values" of both relative and absolute indices, would typically register, in the presence of equi-proportionate increases of population at all income levels, a less-than-proportionate increase in inequality.

There is a strong case, from the perspectives of both logical and ethical acceptability, for employing comprehensively intermediate (that is, income-cum-population-intermediate) measures of inequality in empirical work. The predominant mainstream practice, however, is to employ strictly relative measures. One such measure is the Gini coefficient of inequality. What happens when we replace the relative Gini with the intermediate Gini coefficient? It is instructive to consider this question in the context of economic inequality in India and the world as a whole.

In India, we do not have systematic data on the distribution of personal incomes, though we do have data, from the periodic surveys conducted by the Central Statistical Office’s National Sample Surveys, on the distribution of consumption expenditure. If we look at data from the 1970s to the 2010s, we find that the relative Gini had displayed a moderately rising trend in urban India and no significant trend in rural India so that, given that the rural population predominates, the overall all-India trend is not an alarmingly increasing one.

This has served as a basis for neoconservative commentators to claim that growth in India has not been seriously non-inclusive, and has also been good for poverty reduction. Such an assertion misses the point that a greater emphasis on redistribution would have secured further reductions in poverty than has been the case with the jobless, trickle-down growth that the country has seen in the last three decades. It also misses the point that the trend in the intermediate Gini coefficient is a clearly increasing one for both rural and urban India. This trend is particularly and severely apparent in the over-time distribution of household wealth.

Exactly the same sorts of results hold for the world as a whole, as revealed by the contrast in the over-time behaviour of the relative and intermediate Ginis for global income. From the late 1980s to the late 2000s, the global relative Gini has remained roughly constant, as revealed in the important work of experts like Branko Milanovic of the City University of New York, whereas the global centrist Gini has registered a rising trend.

Both diagnosis and policy prescription are dependent on the findings from measurement. Yet, if there is one thing which measurement teaches us, it is that its underlying conceptual basis is no bedrock of value-free certainties. On the contrary, measurement is shot through with ambiguity, doubt and uncertainty. As the logician Frank Ramsey, who died tragically young at the age of 28, said: “We can make several things clearer, but we cannot make anything clear."

One thing which is reasonably clear though, despite protocols of measurement which tell us otherwise, is that inequality in the world is increasing. Why we must do something to stop this is a subject we won’t even get around to discussing if we keep denying that it is happening. Meanwhile, the coexistence of poverty as widespread destitution and affluence as concentrated obscenity is moral and political grotesquery that bodes ill for the prospects of our country’s continued democratic functioning.

This concludes my 10-part series of essays on economic inequality.

S. Subramanian is an economist.

Comments are welcome at theirview@livemint.com.

Close