For The ‘Leaked’ NSS Data To Be Right, Almost All Other Consumption Indicators Need To Be Wrong

The 2017-18 National Sample Survey’s (NSS’s) Consumption Expenditure Survey (CES) is yet to be released. However, a leaked report has emerged in a newspaper today (15 November). The report shows a finding that is consistent with the Periodic Labour Force Survey’s (PLFS’s) consumption module. That is, it shows a fall in consumption for the first time in nearly four decades.

However, the findings are far from being true for many reasons. For starters, the implication of a decline in consumption should correspond to an increase in poverty rates in 2017 compared to 2011. Nearly every other indicator shows India’s poverty to be in single digits and very close to 5 per cent in 2018.

Therefore, something is amiss as both these findings are not consistent as it is highly unlikely that poverty would have shot up till 2017, and then drastically reduced in just one year to a single digit. That would imply one of the biggest reduction in poverty levels in a single year anywhere in the world.

The problems with PLFS’s consumption module were discussed by Dr Surjit S Bhalla in an Indian Express article (you can read it by clicking here).

In a joint article with Dr Surjit S Bhalla for the Indian Express, we showed that as per the Tendulkar Committee poverty line, only 4.5 per cent of Indians were classified as poor. Of course, there is a need to raise the poverty line but that is a discussion for another day. For now, let us take the line to be given and evaluate the findings of the NSS CES 2017-18.

There are significant problems with the assertion that consumption expenditure has reduced between 2011 and 2017 as this finding is inconsistent with what we already know about the Indian economy.

For starters, we know that offtake of rice increased by 7.2 per cent, of wheat increased by 2.3 per cent. That is, total offtake increased by 5.1 per cent.

If we move beyond food, we find that the number of air passengers increased by 95 per cent while oil consumption increased by 37.2 per cent.

During the period, consumption of electricity and the number of mobile phone users increased by 43.9 per cent and 44.3 per cent respectively.

Therefore, clearly at the aggregate level consumption has increased.

Moreover, as Dr Bhalla highlights in his Indian Express article, total national accounts consumption per capita has increased by 34 per cent, per capita agriculture at 11 per cent, per capita oil consumption increased by 25 per cent and per capita electricity consumption increased by 29 per cent.

Therefore, clearly, there are problems associated with the CES and its findings as they don’t match with what we know from administrative data regarding the economy.

Now the question is whether CES is correct or the administrative data.

For starters, it is worth noting that administrative data is of actual consumption while sample surveys tend to provide granular details with respect to the consumption pattern at the household or district level.

However, in general they both should be complementary in nature, that is, the finding of the administrative data and the survey data should ideally be close. The closer they both are, the better is the quality of data.

Unfortunately, both these data points have been diverging for a long time and several committees have looked into exploring these issues in detail. The gap between the National Accounts Statistics (NAS) and Household Consumption Surveys used to be in the range of 20 per cent in 1980s but it had gone up to 45 per cent in 2009-10. Therefore, the problem of CES underestimating consumption expenditure is not a new problem.

It is worth mentioning that the comparison of NSS estimates for consumption expenditure on healthcare and education with focussed social expenditure surveys of NSS also reveal a discrepancy.

That is, for the focussed surveys we find the per capita consumption to be higher as per the modified mixed reference period.

Therefore, clearly CES is underestimating the expenditure on healthcare and education. There have been concerns with respect to the ability of the Consumption Expenditure Survey and its ability to capture consumption of services.

Though the 2017-18 round was supposed to have a detailed section on service consumption, however, results indicate that it has failed to capture this aspect. A major reason behind this is a lengthy questionnaire that often results in respondent fatigue.

The lack of adequate capture of service expenditure is important for a growing economy as a bulk of the increase in consumption is bound to happen on services rather than in consumption of food and other articles.

Therefore, a growing divergence in the NSS and NAS reflects the problems with our consumption expenditure survey – a fact further corroborated by the divergence between NSS’s own focused surveys with the CES.

There is an urgent need to address this problem in order to ensure good quality data is available for economic policymaking.

Clearly, the national statistical commission has a tough task ahead for itself.

As far as concerns of falling consumption levels are concerned, they may give one sensational headline however there is nothing more to it than sensationalism.

India’s poverty estimates as per the Tendulkar poverty line is indeed 4.5 per cent and a recent joint paper with Dr Surjit Bhalla reveal that based on the World Bank middle income poverty line (less than $3.2ppp per day) poverty in India is at 31 per cent from 57 per cent in 2011-12.

This is the fastest pace of poverty decline India has ever experienced so clearly, India’s record of poverty alleviation is by far one of the best.