OPINION I The Financial Express

Any debate must be bolstered with data, not ancedotes.

This is the season for quoting, and misquoting, Keynes. “When the facts change, I change my mind. What do you do, sir?” This is attributed to Keynes, though there is no actual evidence Keynes ever said, or wrote, this. However, both Paul Samuelson and Joan Robinson have attributed this to Keynes. Statements have also been incorrectly attributed to Sherlock Holmes. “Elementary, my dear Watson,” is the obvious example. But there is one statement clearly made by Sherlock Holmes. I am quoting this from A Scandal in Bohemia. There is a slightly different, and earlier, version in A Study in Scarlet. “It is a capital mistake to theorise before one has data, insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” Whether Keynes or Holmes, there is emphasis on facts and data. Many problems with policy-making in India stem from paucity of data, understandable for an economy that is largely self-employed, informal/unorganised and even rural. These constraints are mentioned in the 2001 Report of National Statistical (Rangarajan) Commission and continue despite amendments to Collection of Statistics Act (1953). Indeed, the present National Statistical Commission is examining ways to revamp India’s statistical system. Sometimes, there are holes in data. Sometimes, there are time-lags.

You can’t devise policy in 2016 on the basis of Census (2001) or National Sample Survey (2011-12). Sometimes, policy requires a census, surveys won’t suffice. SECC (Socio Economic Caste Census) is an instance of a census, where, to identify beneficiaries of government programmes, the challenge is to update after 2011-12. Most assertions about MSMEs are based on a census conducted in 2006-07. In a way, calling this a census is unwarranted. It is a census only for registered MSMEs. For unregistered MSMEs (which constitute the overwhelming bulk), it is really projections based on a survey. There was a time when respectable surveys, particularly on consumption expenditure, income, access to public goods and services, were only conducted by government, NCAER being a bit of an exception. By “respectable”, I mean a survey with a proper sampling design and large samples, spread across all states, not one where one thrusts a microphone before 10 people in Delhi and extrapolates to all of India. To the best of my knowledge, there are two surveys that satisfy this respectability test. There are two products CMIE brings out in collaboration with BSE. And there is PRICE (People Research of India’s Consumer Economy). PRICE did a survey in August 2016, with more than 300,000 households, straddling 216 districts, 1,217 villages, 487 towns and 25 states.

There was an earlier survey in 2014. Therefore, you can look at the 2016 snapshot. You can compare improvements between 2014 and 2016. Since a panel of 12,000 households is constant between 2014 and 2016, that improvement scrutiny is even more robust. There is a wealth of data. Let me focus on a few. A remarkable 98.7% of households have bank accounts (98.9% urban, 98.6% rural), a fact worth remembering when citing dated data on financial inclusion. (This doesn’t mean financial products have penetrated.) Between 2014 and 2016, income share of the bottom 20% of the population has increased, while that of the top 20% of the population has declined. Surplus income (defined as what is left of income after routine expenditure is met) of bottom 20% has increased, while that of top 20% has declined. Between 2014 and 2016, the largest increase in income (three times the increase in national average) has been for urban labour, while the sharpest decline in income has been for rural labour employed in agriculture. Ninety-one percent of households have electricity. (A word of caution for those who quote Census figures.) Ninety-two percent of households own mobiles. (Also a caution for those who quote Census.) Fifty-seven percent have LPG connections, while fifty-five percent have tap water. (For both these indicators, Census figures were less than half these numbers.)

For CMIE, the two products (available on the internet) are on consumer sentiments and unemployment. There is again a wealth of data generated from 158,624 households. Take the unemployment rate, last updated on November 30, 2016. It is given separately for rural and urban India, but let me cite all-India figures. That rate was 9.84% on August 31, 8.97% on September 30, 6.34% on October 31 and 5.69% on November 30. There is a tendency to ascribe everything to what happened on November 8 and also highlight rural distress based on anecdotal accounts. The decline in rural unemployment rate mirrors the all-India trend. For urban, there is a decline to 6.89% on October 31, followed by an increase to 7.05% on November 30. Till end of November, this indicator doesn’t reveal a great loss in jobs, either rural or urban. In case one thinks unemployment rate is an imperfect indicator, let’s look at the consumer sentiments index. For rural India, the dip occurs from end of August to the end of September and there is a steady upwards climb thereafter. There is no spike because of November 8. For urban India, the dip occurs from end-September to end-October, with a steady climb thereafter, with no spike because of November 8. My intention is not to debate impact of November 8, but to point out that any debate must be bolstered with data, not anecdotes.

CMIE’s unemployment product is a treasure trove of data, state-wise too. For instance, there seems to be a sharp increase in voluntary unemployment.