Lies, damn lies, and Indian economic statistics. Can someone comb out the lies from these statistics and tell us the truth?
Just to clarify, we are not referring to data on the current state of the Indian economy, which is plummeting fast with no end in sight. We are referring to those golden years, from 2011-12 to 2017-18, when India was among the world’s fastest-growing economies. The last three years of the UPA government with Manmohan Singh as prime minister and the first four years of the NDA government with Narendra Modi at the helm.
The Economic Survey data show that the Indian economy grew at an average of about 7% a year, or a cumulative 48% over the entire sevenyear period. An impressive feat by not just Indian but even global standards.
But a draft of the consumer expenditure survey 2017-18 results from the National Sample Survey Organisation (NSSO), leaked to the press, shows that the average per-capita monthly spending, adjusted for inflation, fell by 3.7% to Rs 1,446 in 2017-18 from RS 1,501 in 2011-12; per-capita spending in the rural areas fell by 8% and in the urban areas increased by 2%.
Is that even possible? The answer is no. It is not possible to have a 7% annual growth in GDP for seven years, but a stagnation, or slight fall, in consumer spending. So, whom should we trust? NSSO or the Economic Survey? The rosy GDP growth statistics from the Survey have been endorsed by the likes of the International Monetary Fund (IMF) and the World Bank. One might say that their endorsements are largely based on government data.
So, these endorsements might be under the GIGO — Garbage In, Garbage Out — effect. Yet, it is not possible to keep up the pretence of being one of the fastest-growing economies for seven years when, in fact, per-capita spending was stagnant or even falling.
NSSO surveys suffer from credibility issues. Many have questioned its recent report on employment. Many have been questioning its consumer expenditure surveys for years.
One can also take the view, which may seem partisan (read: anti-BJP), that the economy was growing fast during UPA’s last three years, and it began to tank in the NDA years. But this view certainly does not gel with Narendra Modi’s two electoral victories.
It is hard to believe that rural spending was plummeting in the year before Modi won a thumping majority in the general elections. Political choices are not too far removed from economics.
So, what’s going on? One possible theory is that with the proliferation of welfare schemes that provide benefits based on income, survey respondents have an incentive to under-report theirincomes and consumption expenditures. They will probably not underreport on weekly consumption of rice and wheat, but may decide not to mention purchases of high-end items like a television set, a motorcycle, home renovations to the surveyors.
NSSO can test this theory by matching sales data on specific consumerdurables (motorcycles, TVs and otherlarge items) of multiple years with survey reports of purchases and ownership. Importantly, NSSO can devise surveys to check whether respondents under-report, which items and to what degree.
Note that data inconsistencies are not the exclusive domain of Indian economic statistical organisations. Astudy (bit.do/fhSRh) by economists Bruce D Meyer, Wallace K C Mok and James X Sullivan found that the US Census Bureau annual surveys (Annual Social and Economic Supplementof the Current Population Survey) report only half the amount received through food stamps, welfare cash transfers, workers’ compensation.
In the same survey, 45% of the respondents do not report their income, which is then imputed. Further, the non-response rate is higher at the tail ends of the distributions and more for families that are likely to fall at the bottom end of the distribution (based on demographic characteristics such as age, education, race, ethnicity). According to a presentation by Trudi Renwick of the US Census Bureau, non-response of personal income has risen from 28% in 1997 to 45% in 2015.
In short, NSSO surveyors are in good company. They can learn from the experience of other countries in detecting if survey respondents are under-reporting, and whether there are ways in which certain data on consumer spending can be supplemented from other sources to adjust for under-reporting.
No Date With Data
When faced with data that shows the Indian economy in poor light, successive Indian governments have historically resorted to not releasing the data, or discontinuing future surveys. When the 2004-05 Demographic and Health Survey (DHS) showed high incidence of infant mortality, stunting and underweight, the Manmohan Singh government simply decided not to conduct the next DHS survey, which was expected in 2010-11.
Now, the Modi government has announced not to release the 2017-18 NSSO survey. These are bad practices.
Despite massive under-reporting on welfare receipt and incomes in the US surveys, these surveys remain in the public domain. The US Census Bureau is not hiding the data, but is working on findings ways of improving it. The NSSO and the government should follow the same mantra.
The writer is professor, social policy, Columbia University, US