Caste proponents say the census must include questions on caste to establish true caste ratios. Opponents say questions on caste are socially divisive. They also raise a behavioural objection: the very announcement of a caste census could encourage people to claim, fraudulently, that they belong to a caste entitled to reservations.
This behavioural objection applies as forcefully to surveys for determining poverty. The National Sample Survey Organization conducts periodic surveys on consumption, and data from these surveys are used to determine the official poverty head count. The government assumes that people tell NSS surveyors the truth about themselves. Really?
NGO workers in rural areas tell me that if a villager talks to fellow villagers, he will probably exaggerate his living standards to get local prestige. But if a villager is talking to outsiders who may influence the giving of subsidies or other benefits, then the villager will vehemently claim to be poor and deserving of any benefits available. He has no reason to admit any improvement in living standards.
Any psychologist will tell you that such duplicity is both inevitable and universal. People respond to incentives and disincentives. So, a necessary condition for an honest survey is that no benefits or penalties should flow from people’s replies. That is true for surveys on caste, poverty or anything else.
But governments have long been promising ever more subsidies and entitlements targeted at the poor. This is a strong, growing incentive to under-report consumption. The evidence suggests that this has long been happening.
In the 1970s, consumption measured by NSS surveys was 87% of consumption measured by national income statistics. This divergence was bearable. But the divergence has steadily increased over decades. In 2004-05, consumption reported by the NSS was just 48.7 % of that reported by national income statistics.
Why so? Left economists claim that rich people are progressively under-reporting their consumption to the NSS, but poor ones are not. This is implausible. In the 1970s, income tax of 97.75% and wealth tax of 3.5% caused a flight to black money, which nobody reported to the NSS. Consumer imports were illegal and provided by smugglers.
Obviously richer people under-reported such consumption. Yet in the 1970s, the two measures of consumption were not far apart.
But over the next three decades, income tax has fallen to 30%, wealth tax is virtually abolished, imports of consumer goods are legal and lightly taxed, and the share of white money transactions has clearly risen, even in real estate. So, the rich have much less reason to hide their consumption than in the bad old days.
However, poorer folk have been given increasing incentives to under-report living standards. In the 1970s, most freebies were universal, not targeted at the poorest. Subsidized food, power, canal water, fertilizer, bank credit and employment schemes were open to all rural folk, not just the poor.
Targeting came later. The IRDP scheme provided subsidized microcredit to the poor. The Indira Awas Yojana provided housing subsidies to the poorest. BPL (below poverty line) ration cards provided a higher food subsidy than APL (above poverty line) cards. Rural employment in some cases was targeted at the poor, and so was the farm loan write-off in 2008. The DMK in Tamil Nadu provided colour TV sets to those below the poverty line.
Many such schemes were poorly implemented, and often little money reached the poor. Yet villagers realized that some benefits could flow from fraudulently claiming poverty, while no profit at all flowed from honestly reporting improved living standards.
This must have affected the honesty of reporting in NSS surveys. This seems at least a partial explanation of the growing gap between NSS estimates and national income estimates.
Officials continue to treat NSS data as sacred. But companies selling cellphones, shampoo, TVs and other consumer goods view NSS estimates as rubbish. NSS data suggests high poverty and little improvement in rural demand. In fact companies find rural demand is booming, producing record sales. Thanks to rural penetration, cellphone connections are increasing by 19 million per month!
Economist Surjit Bhalla has long denounced official poverty estimates as fiction. By making adjustments consistent with national income data, he reckons actual poverty is less than one-third the official figure. Other adjustments can be made by looking at independent data on rural consumer goods sales.
The government has recently raised the poverty line. Around 37 % of people are below this higher line. Such upward adjustments are appropriate as living standards rise. But this should not distract from the terrible flaws of NSS data. Government policies have made it profitable for rural folk to exaggerate their poverty.