Measuring Unemployment Trends in India Prabhat Patnaik

Unlike in the advanced capitalist countries, a reduction in employment opportunities in India takes the form not of a larger proportion of the work-force being shut out of employment, but of almost everyone having lesser number of days of work. This is a reflection of the fact that only a tiny segment of the work-force is employed on a full-time basis. Most are either self-employed, like peasants or shop-keepers, where many more or fewer family-members can share a given amount of work; or are causal workers who may get work on a given day but not on another. The same number of persons in other words gets more or less work; unemployment can be said to have fallen in the former and risen in the latter case. There is no close correspondence between the magnitude of work and the number of persons working. The question arises: how can we measure unemployment trends in such a situation?

Because of this problem, official statistics in India have for long used not one but several different measures of unemployment. We shall not go into all these measures here. But the problem with all of them can be illustrated by taking any one of them. Let us take for illustration the concept of “current weekly status” unemployment: if a person has not got even one hour of employment in some gainful activity, despite looking for work, over the preceding week (prior to his being interviewed), then that person’s “current weekly status” is being unemployed; otherwise the person’s “current weekly status” is being employed. But consider a person who had four hours of work over the preceding week during the survey in a particular year; during the next survey a few years later, such a person has only two hours of work during the preceding week. Clearly then in reality the person is more unemployed at the later date compared to the earlier date; and, if that person happens to be typical, then the unemployment situation has worsened in the country between those two dates. But the “current weekly status” unemployment rate as measured by official statistics would have remained unchanged.

This being more or less true of every official measure, none of them is any good at capturing unemployment trends. How then can we capture unemployment trends? When the amount of work available per head of the work-force is decreasing, that is, when unemployment is actually increasing, it is perfectly reasonable to believe that an average worker’s real earnings per hour of work could not be increasing. But since at such a time the number of hours of work per worker would be decreasing, it follows that the product of the two, which is the real earnings per capita of the work-force, must be decreasing. It follows that a symptom of rising unemployment, other things being equal, would be decreasing real earnings per capita of the work-force.

Correspondingly, when unemployment is decreasing, the per capita real earnings of the work-force should be increasing. Hence, a reasonable proxy for the trend in unemployment rate, if other things can be taken to remain unchanged, would be to take the movement in per capita real earnings of the work-force. The problem however is that other things do not remain unchanged. In particular when per capita earnings of the workers decline, and the share of economic surplus increases in total output, the size of the class of “hangers on” who subsist on this surplus, such as lawyers, business executives, advertisement agencies and such like, also increases, and they are also usually classified as belonging to the “work-force”. The “work-force” in other words does not just include workers, peasants, agricultural labourers, artisans, craftsmen, fishermen, and such like, whose unemployment rate is our real concern; it also includes corporate lawyers, executives etc. Hence, while movements in the per capita earnings of the former alone could have given us a reasonable proxy for unemployment trends, the observed per capita earnings of what is called the “work-force” might be misleading as a proxy.

There is however a way out. It is reasonable to believe that when the per capita real earnings of the section of the “work-force” belonging to the upper crust, which really consists of surplus earners, increases or remains unchanged, their per capita absorption of foodgrains does not decrease. When the observed per capita absorption of foodgrains declines, then one can certainly infer from this fact that the per capita real earnings of the “work-force” consisting of the genuine working people must have decreased.

It follows that movements in the per capita absorption, or availability, of foodgrains, in any situation where there are no constraints of foodgrains supplies, can be taken as a reasonable proxy for the trends in unemployment rate.

This can be objected to on the grounds that it makes trends in unemployment almost synonymous with trends in hunger. But unless an increase in hunger is caused by an independently-arising shortage in the supply of foodgrains, it reflects a growing shortage of purchasing power in the hands of the working people, which is symptomatic of an upward movement in unemployment. And almost throughout the neo-liberal period, there have been more foodgrain stocks with the Food Corporation of India than are considered necessary for its purposes, so that there is no question of any independently-arising shortage in the supply of foodgrains. Our proxy measure of unemployment trends therefore becomes quite a legitimate one.

Of course, foodgrain availability fluctuates from one year to the next; what is relevant for us however is the trend in it, and we examine it by comparing the average annual per capita availability for two triennia, one belonging to the initial dates and the other to the terminal dates of the period of our study.

For the triennium 1989-91, that is, just on the eve of economic “liberalisation” in India, the average annual per capita net availability of foodgrains was 180.2 kilogrammes; for the triennium 2016-18, just prior to the pandemic, for which we have somewhat firm data, it had come down to 178.7 kilogrammes. By our argument therefore there was a rise in the unemployment rate over the period of neo-liberal economic policies. Lest it be thought that this result arises because of the particular dates we have chosen, we should clarify that for no subsequent triennium during the entire neo-liberal period was the per capita net availability of foodgrains higher than in 1989-91. The conclusion about its decline over this period as a whole therefore is indubitable.

There is a further fact that strengthens this conclusion. After the introduction of the Mahatma Gandhi National Rural Employment Guarantee Scheme under the first UPA government, some employment was provided under this scheme which added to the earnings of the working people and hence to their demand for foodgrains. If despite this there was a drop in per capita foodgrain availability over the neo-liberal period, then it follows that the drop would have been even sharper in the absence of MGNREGS.

The matter can be put another way. One can distinguish between the employment provided by the operation of the neo-liberal capitalist system and the employment provided by the MGNREGS which is a supplementary source of employment. The fact that unemployment, despite taking into account MGNREGS employment, increased over the neo-liberal period, suggests that the system’s ability to provide employment was even more restricted than appears from our proxy measure, that the increase in the unemployment rate would have been even greater if the neo-liberal capitalist system alone was providing employment.

The reason for this phenomenon of rising systemic unemployment is two-fold. First, in addition to the natural growth of the working population, an influx takes place into the cities of distressed peasantry looking for wage employment. This happens because of the strangulation of peasant agriculture under neo-liberalism, through the withdrawal of input subsidies by the government and the dismantling of the price-support mechanism in the case of cash crops. Secondly, the removal of all restrictions on technological-cum-structural change, together with the opening up of the economy to competition from imported goods, forces the adoption of such change, which is typically labour-displacing and hence causes very little employment growth, no matter how high the rate of GDP growth.

These two factors together ensure that the unemployment rate keeps increasing in an economy under neo-liberal capitalism. This increase however escapes being fully captured by conventional official statistics in India. It does show up however if we look at proxy measures of the sort we have suggested here.

(This article was originally published in People’s Democracy on October 17, 2021)