People's Democracy

(Weekly Organ of the Communist Party of India (Marxist)


No. 04

January 27, 2013


Underconsumption under Capitalism


Prabhat Patnaik


A DEBATE has arisen among two well-known American economists, Joseph Stiglitz and Paul Krugman, on a theme that has been of interest to Marxists for a long time. Stiglitz has argued in a recent book that an increase in inequality in income distribution in society, lowers, other things remaining the same, the level of aggregate demand; and hence gives rise to a tendency towards lower output and employment. It follows from this that greater equality is not only socially desirable but also economically beneficial.


This has been a line of argument, traceable originally to the nineteenth century French economist Sismondi, which has been advanced for long within Marxian economics. There is however one important difference between Stiglitz and the Marxists. The Marxist argument has been that capitalism gives rise to this tendency towards greater inequality, in the form of a decline in the share of wages over time, because, while labour productivity rises over time owing to technological progress, the level of real wages is kept down by the existence of a substantial reserve army of labour: such a decline in other words is endemic to the system. Stiglitz’s argument on the other hand is that such a decline has happened no doubt, with the consequences he has noted, but it is reversible through State action within the system.


On the analytical consequences of such a decline, however, the Stiglitz position does not differ from that of the Marxist tradition, stretching from Rosa Luxemburg to Baran and Sweezy, which has advanced this line of argument and which has been rather misleadingly called the “under-consumptionist tradition”.


I say “misleadingly” because “underconsumption” in the sense of the bulk of the working people “not getting enough by way of consumption” (relative to some “norm”), which is the sense in which Friedrich Engels had used the term, must be distinguished from “under-consumption” in the sense of a secular tendency, other things remaining the same, toward a decline in output, owing to a fall in aggregate demand arising from a decline in the share of consumption in output; this decline in the share of consumption in output, in turn, is caused by a shift in income distribution from the workers to the capitalists. What is called “underconsumption” in this latter sense should really be called a “tendency towards over-production arising from a decline in the share of wages in output (or a rise in the share of economic surplus)”. But the term “tendency towards under-consumption” has been in vogue for so long that we may as well continue using it in this second sense.


This view that capitalism is actually afflicted by a tendency towards “under-consumption” has not been generally accepted by Marxists, and those who have rejected it have also done so for different reasons. It is worth discussing briefly the different arguments within the Marxist tradition against the “under-consumption” view. What is common to all these arguments is the proposition that consumption is not the aim of capitalism, and that even as consumption may decline in relative importance, other sources of demand may make up for it, without necessarily causing any difficulties for the system.




Lenin, for instance, had argued that the very technological progress which raised labour productivity relative to wages and therefore brought down the share of consumption in output, also raised the organic composition of capital and hence the share of output value that had to be used as addition to constant capital. A simple arithmetical example will make Lenin’s point clear. Imagine an economy growing at five per cent per annum. In the initial situation, to maintain this five per cent growth, 15 per cent of output has to be invested. Out of an output of 100, wages, say, are 50, and profits are 50, of which 15 is invested and 35 consumed by capitalists. Now, suppose technological progress is accompanied by a rise in labour productivity unmatched by a rise in wages, so that the share of wages falls to 45 per cent of output. This productivity increase however may entail that a five per cent growth rate will now require a 20 per cent investment ratio, in which case out of 100 output, wages will now be 45, all of which will be consumed, and profits will be 55 of which investment will be 20 and capitalists’ consumption 35 as before. The fall in the share of wages in this case has caused no tendency towards recession or stagnation.


But Lenin’s argument was confined only to saying that underconsumption does not necessarily come in the way of accumulation and growth. The extreme position that it can never come in the way of accumulation and growth, that any decline in the share of consumption in output is necessarily made up by larger investment by capitalists was taken by the Russian economist Mikhail Tugan-Baranovsky. But Tugan’s position amounts to saying that capitalism can never face any problem of aggregate demand, that over-production crises are impossible under capitalism, and hence accepting Say’s Law, which Ricardo had done but which Marx had strongly rejected (thus anticipating the so-called “Keynesian Revolution” by almost three quarters of a century).


As a matter of fact, even the proponents of the “underconsumption” view do not necessarily say that the system will break down or get bogged down in permanent crisis etc. They see “underconsumption” as what economists call an “ex ante tendency”, something that would happen if other countervailing factors did not intervene. The countervailing factor emphasised by Rosa Luxemburg was encroachments into pre-capitalist markets, for which colonialism and imperialism were the instruments; and the countervailing factor emphasised by Baran and Sweezy was military expenditure by the State, in particular, in the post-war period, of the State of the leading capitalist country, the US.


The difference between what the underconsumptionist position says and what Lenin, not Tugan-Baranovsky, had argued consists in this: Lenin had talked about these countervailing factors arising from within capitalism itself, while the proponents of underconsumption argue that countervailing factors from within the system are not enough, and that these have to be necessarily supplemented by the State, by pre-capitalist markets etc. But in the entire debate within the Marxist tradition, nobody has ever argued that a rise in inequality through a shift of income distribution from wages to profits will not lower the share of consumption in output. And this is what Paul Krugman is arguing against Joseph Stiglitz.   




Krugman’s argument is that if a dollar is transferred from the workers to the capitalists, the proportion of it that latter spend on consumption is no less than that of the former. From the fact that in any period the rich appear to consume less than the poor we cannot infer anything about their spending behavior because there will be a disproportionately large number of people with temporary windfalls among the rich in any period and there will be a disproportionately large number of people who are temporarily “passing through bad times” among the poor. Since such temporary gains and losses do not immediately affect consumption, from the observed fact of lower consumption per unit income for the rich than for the poor, we can infer very little about the intrinsic consumption behavior of the rich compared to the poor.


Krugman therefore tries to settle the issue empirically, by turning to US data. And here he finds that private savings as a proportion of GDP were down, not up, before the crisis which contradicts the view that an increase in inequality lowers the consumption ratio (and hence raises the savings ratio). He also finds that the savings ratio went up after the crisis, even though there had been no sudden or significant increase in inequality in the interim. He contends therefore that the “underconsumption” view is fundamentally flawed.


Krugman’s argument however is both theoretically untenable and empirically unconvincing. Let us look at the theoretical issue first. And for this it is useful to talk not in terms of “rich” and “poor” (and hence personal income distribution), but in terms of “capitalists” and “workers” (and hence class income distribution). Suppose Krugman is right and whether a rupee accrues to workers or capitalists makes no difference to aggregate consumption; then any excess demand in a situation where output happens to be supply-constrained should be pushing up the price-level to infinity. This is because demand (let us assume for simplicity a closed economy) consists of two parts: investment demand and consumption demand. If this demand is too much relative to the output that can be produced, then it is adjusted to the latter not, typically, through any lowering of investment, but through a rise in prices relative to money wages, and hence a rise in profit margin, ie, a shift of income distribution from wages to profits, which curtails the overall consumption, and hence demand in the economy. Such a curtailment in demand, and hence a ceiling on the extent to which price needs to rise relative to money wages (which are given in the short run) occurs precisely because capitalists’ consumption out of a unit of extra income is lower than that of workers, ie, precisely because of the very proposition that Krugman is denying.


It may be thought that any such rise in profit margin, and hence the accrual of larger profits for a given output, was like a temporary windfall; and hence from the fact of its immediate non-consumption, which is what ensures price stability in the system, nothing can be inferred about consumption behavior, as Krugman has argued. But unless there is a systematic difference between capitalists’ and workers’ consumption behavior, prices will still hit infinity. If capitalists do not increase their consumption because of temporary gain and workers also do not reduce their consumption because of temporary loss, then again the problem of excess demand will never get eliminated. Hence there must be a difference between capitalists’ and workers’ consumption ratios out of income, if the system is to be able to resolve the potential problem of its price instability at the expense of the workers. (This point incidentally was made not by any socialist or Marxist economist, but by John Maynard Keynes through his theory of “profit inflation”, which argued that booms under capitalism were invariably financed by squeezing the workers). And once we accept that there is such a difference between consumption ratios, then any shift in income distribution, no matter at what level of capacity utilisation, from workers to capitalists, will have the effect of pushing the economy to a lower output and employment.


The empirical point about savings ratio in the US falling when income shifts from workers to capitalists and rising sharply in a crisis, even when no income shift takes place, is easily explicable and does not negate the difference in consumption ratios of the two classes. It arose empirically from the fact that workers’ consumption was sustained through credit even when the share of wages in output was falling. A simple numerical example will make the point clear.


Suppose the output of an economy is 100, divided between workers and capitalists in the ratio of 50:50. Workers consume their entire wages and capitalists habitually consume only half of their profits. Then workers’ consumption is 50, capitalists’ consumption is 25 and capitalists’ savings are 25 which exactly equal investment (otherwise 100 of output will not be realised).


Now suppose income distribution between workers and capitalists changes to 40:60; and let us assume (just for simplicity) that already employed workers (and not any newly employed ones) can get credit from banks to maintain their consumption. Since total profits must equal in this case the sum of investment, capitalists’ consumption and loans to workers, it follows that the new output will be 116 2/3, so that profits out of it, at 60 per cent, will be exactly 70 (to be just right for financing investment, capitalists’ consumption and loans to workers); and wages will be 46 2/3. Out of the 70 profits, 35 (or exactly half ex hypothesi) will be consumed by the capitalists, and 35 saved; but 10 of these savings, deposited let us say with banks, will go to finance workers’ consumption in excess of the wage bill, and 25 will finance investment, which remains at 25 as before. Workers’ consumption will be 56 2/3, of which 10 will be financed by credit from banks and 46 2/3 from the wage bill. For the economy as a whole, the private savings ratio (which in this case is the same as the overall savings ratio) has fallen from 25 per cent to 21 3/7  per cent (which is 25 divided by 116 2/3). A fall in the private savings ratio occurs in this case despite the shift from wages to profits.


Now let us see what will happen if this credit is suddenly withdrawn, as is the case in a financial crisis. Total profits will now equal the sum of investment and capitalists’ consumption alone; output therefore will fall suddenly to 83 1/3, of which profits, at 60 per cent, will be 50, and wages will be 33 1/3. Out of the profits of 50, capitalists’ consumption will be 25 and their savings will be 25. The private savings ratio in the economy will shoot up to 30 per cent, even though no change has occurred in income inequality during the crisis.


In short, everything that Krugman adduces as evidence against the theory of underconsumption is perfectly explicable within the context of that theory. It constitutes no argument either against that theory, or against Stiglitz’s view that a rise in income inequality tends to depress aggregate demand in a capitalist economy.