THE WEALTH OF DATA ON WHAT INDIANS CONSUME

THERE is a celebrated sherlock Holmes quote: “it is a capital mistake to theorise before one has data.

- BIBEK DEBROY

Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.”

This is a more familiar version of the adage from ‘A Scandal in Bohemia’. a slightly different and less colourful version exists in ‘A Study in Scarlet’, conveying the same intent. Whether a capital mistake or not, it is a current mistake to theorise, and passionately assert, in the complete absence of data.

Witness the discourse and debate on inequality, which makes blood pressures shoot up. inequality is defined with respect to a metric, a variable whose distribution is being mapped. inequality in the distribution of wealth, which is a stock, is one measure. inequality in the distribution of income, which is a flow variable, is another. one should rightly be sceptical of estimations of wealth, especially if imputation of the value of real estate or shares is involved.

Strangely, a lot of people who pontificate on policy do not seem to know that india, like many other countries, does not officially collect data on incomes. That’s because data on income is believed to be unreliable. instead, we collect data on consumption expenditure, an exercise the national sample survey office (nsso) started in 1950,.

Inequality in the distribution of consumption expenditure will be lower than in income. nevertheless, from the former, we can deduce something about the latter, at least as a trend. There is a difference between a census and a survey. a census is a complete enumeration of the population. a survey is based on a sample. if the sample is truly representative, the survey will tell us something about the characteristics of the population. For consumption expenditure, nsso has surveys known as the household consumption expenditure survey (HCEs).