Why the Core of ex-CEA’s Argument on India’s GDP ‘Overestimation’ Stands

The “gross domestic product” (GDP) is a concept rooted in an epistemic position which is intrinsically incapable of recognising the existence of a “surplus” in society. A simple example will make this clear. Suppose we have an agrarian economy in which 100 peasants produce 100 units of food; and suppose 50 of these are taken by an overlord through taxes, for consumption by his family and hangers-on. These 50 units will be readily recognised as constituting a “surplus” out of the total output of 100. But the concept of GDP would not recognise this. Instead it would claim that the GDP of the country is 150, consisting of 100 units of food and 50 units of “services” rendered by the overlord and his retainers. In fact if there is an increase in the degree of exploitation of the peasants by the overlord, through a rise in the magnitude of taxes to 60 from the original 50, then this increased exploitation will appear as an increase in GDP to 160 from the original 150. Increased exploitation of producers will appear as a 6 2/3 per cent rate of economic growth, from 150 to 160!

This is why Marxists have never taken GDP estimates, and the growth rate measure based upon them, seriously. But the recent claim by the former chief economic adviser (CEA) to the government of India, Arvind Subramanian, that there is a gross over-estimation of India’s recent growth rate, deserves serious attention, because his claim relates not just to the “services” sector, which is where the trouble usually lies, but to a large extent to the material commodity producing sector of the economy, notably, manufacturing.

Subramanian’s argument is that the new method of calculating the GDP, which came into effect recently, has tended to overestimate the growth rate quite substantially. In fact, he shows that the growth rate for the period 2011-2016 is way out of line with a whole lot of other indicators which normally should be increasing with the GDP. Based on this, he estimates the degree of overestimation to have been around 2.5%. Instead of 7% growth rate as officially claimed, the correct growth rate is about 4.5%, in fact anywhere between 3.5% and 5.5%, with 4.5% as the mean figure.

From Subramanian’s statistical exercise about the degree of overestimation of the growth rate, one cannot of course deduce why such an overestimation has occurred. But he gives a number of reasons why it could have arisen.The new method of estimating GDP moves away, in the sphere of industry, from reliance on volume calculations, such as the index of industrial production, to value calculations for which it relies on company statistics made available to the Ministry of Company Affairs. There should be no difference between the two in a situation where input coefficients per unit of output remain unchanged. But when output per unit of input increases, through better management, then the two diverge. There would, in the normal course, be some improvement in “productive efficiency” in this sense, but not as much as the Indian GDP data of late have been showing.

The shift to company-finance-based estimates in a situation of where world oil prices have been falling, is also likely to overestimate the GDP growth rate.This is because of the fact that input and output are not being deflated by their respective separate price indices but both are being deflated by a common index, namely, the output price index; in such a case, a fall in the input price will lead to an overestimation of GDP growth rate, as the following example will show.

Suppose Rs 100 worth of output is produced with Rs 50 worth of an input, so the value added which enters the GDP estimate is Rs 50. Now, suppose the input price is halved, so that the new input cost is Rs 25. If input and output were being separately deflated by their respective price indices, then there should be no change in the “real” value added (that is, at base prices). But if both input and output are being deflated by the same price index, namely the output price index (which has not increased at all), then the real value added at base prices will now appear to be Rs 75. The sector will appear to have grown, from Rs 50 to Rs 75, while there has been no change whatsoever in the level of activity in this sector. Now, oil being an imported good, a fall in its world price, when there is only output price deflation, will show itself as an increase in value added, even though there has been no change in the level of domestic activity.

The fact that India’s GDP growth was being overestimated has been noted by many economists. What is striking about Subramanian’s paper is that a person who till yesterday was the chief economic adviser to the government, is now making this point, which gives the argument greater weight. Not surprisingly, his paper has been much criticised by official spokespersons; and even his motives in writing the paper after having left office, are being questioned. But the core of his argument stands.

If the overall growth rate was less than is claimed, then the sectoral growth rates, too, must have been lower. If we take the sectoral growth rates between 2011-12 and 2016-17 and scale them down in equal proportion, in keeping with the scaling down of the overall growth rate from 7% to 4.5%, then it turns out that the growth of the non-services sector must have been around 2.3%, while that of the services sector must have been around 6%. This growth rate of the material commodity producing sectors is well below what was experienced on average during the pre-liberalisation, or dirigiste, period.

It is well-known that the first decade after economic liberalisation in 1991 had seen no increase in the material production sectors’ overall growth rate compared with earlier; such an increase in growth rate that had occurred then, was in the services sector. Whatever increase may have occurred in the material production sectors’ growth rate, would have been only after the first decade of “reforms”. But now it turns out that in the period after 2011-12, the overall growth-rate, even including the services sector, has been only around 4.5%, which is about the same as the GDP growth rate before liberalisation, and which means that the growth of the material production sector was actually lower than before liberalisation.

It follows, therefore, that barring at best a brief period of a decade or so early this century, the material production sectors’ growth rate has been no higher than, and often lower than, what it had been in the dirigiste period. And since this is the sector that really counts, and not the services sector, whose inclusion in GDP is fraught with serious conceptual problems, it follows that barring a few years at best, the Indian economy has not done all that much better after liberalisation compared with before.

At the same time, the elasticity of employment with respect to output, or the percentage change in employment divided by the percentage change in output, in the material commodity producing sectors has declined compared with the dirigiste period. It is not surprising, therefore, that unemployment has emerged as perhaps the foremost problem for the Indian economy.

All this time, when a hullabaloo was being made about economic liberalisation, the argument had been advanced that the growth rate of the economy had accelerated greatly in this period. The immense increase in income and wealth inequality that had emerged in this period had been justified on the grounds that this was a necessary price to pay for such greatly accelerated growth. But, if the growth rate itself, barring at best a brief interregnum, is not much higher than in the dirigiste period, then this entire line of argument collapses.

The period of economic liberalisation appears to have been one during which the growth rate scarcely moved up compared with earlier; but a thin upper stratum of the population helped itself to a rapidly growing share of the national income, unlike earlier.