Marx set out the notion that a ‘reserve army’ of unemployed labour is more or less continuously maintained in the course of capitalistic development. In the initial phases, this reserve army may be created through the destruction of the pre-capitalistic modes of production while, in later phases, a systematic bias in favour of labour-displacing innovations could serve the same purpose. This entails a broad vision of capitalistic development under extremely elastic supply conditions for labour where the actual level of wage employment is usually demand-determined. This means that the supply of labour tends to adjust to its demand through various routes such as, higher participation rate (e.g. as more married women join the labour force or the average schooling period is shortened), interregional and international migration of labour etc., all this taking place against the background of continuous induced innovations. Under these circumstances, it is not very useful to think of a ‘natural’ rate of growth, set by the growth of labour force and of labour productivity, as the maximum feasible growth rate of a capitalist economy (Marglin 1984 pp. 103–8).

The elastic nature of the labour supply and its adjustability to the level of demand entail the existence of open or disguised unemployment as an untapped reservoir of labour in the normal course of capitalist development. However, such disguised unemployment although real is a somewhat amorphous phenomenon in an advanced capitalist economy for two distinct reasons. First, under normal circumstances, many potential entrants (e.g. married women, late school-leavers, young people on the farm) may not actually even try to enter the labour market unless demand is seen to be high with all sorts of job vacancies exceeding their corresponding numbers in registered unemployment. Second, the economic nationalism in the richer countries often takes the form of strictly regulating the migration of ‘guest-workers’ so that open unemployment in the (potentially) labour-exporting countries, rather than disguised unemployment in the advanced capitalist countries, becomes the normal pattern. And yet, prolonged stagnation in economic conditions in an advanced capitalist country may make this phenomenon of disguised unemployment more visible, as the redundant workers either seek various forms of self-employment with virtually no invested capital or try to sell their labour services directly as porters, odd-jobmen, domestic servants, farm-hands etc. (Robinson 1956, pp. 157–8). Their earnings in these peripheral jobs would then become the ‘reservation price’ of this marginalized labour force. When unemployment dole and social security set a higher reservation price, some of this unemployment may come out in the open instead of being disguised. In this sense, it is probable that the growth of the welfare state may openly register as unemployed some who would have been otherwise unemployed in a disguised fashion earlier. And, the reverse could happen if the social security measures are cut by the government.

The existence of such disguised unemployment on a significant scale is usually accommodated by a secondary or informal labour market mostly in the service sector. This is much more easily visible in the phenomenon of massive migration to urban centres from rural areas in many developing countries. While all such migrants from rural areas aspire to limited job opportunities in organized industries located in urban areas, only a small fraction among them are actually able to find proper jobs at any given point of time. The rest spend their time, waiting in search of appropriate jobs. In the meantime, they somehow manage to disguise their unemployment either by self-employing themselves with tiny amounts of invested capital (e.g. polishing shoes, cleaning cars etc.) or by selling their labour services directly in odd jobs or even, simply taking recourse to the support of the elaborate kinship system in more traditional societies e.g. by living off better-placed relatives and migrant workers from their home areas. Thus, the phenomenon of disguised unemployment in the urban areas of many developing countries becomes closely linked with the massive migration from rural areas during the course of industrialization.

A distinguishing feature of disguised unemployment in such an informal sector is the irregular and often long hours of work per day. This is evident enough in the case of most self-employed persons in the informal sector; but even those who are employed on a wage-labour basis usually have highly flexible wage contracts in many respects (e.g. domestic servants, odd-jobmen etc.). Partly the explanation lies in the lower unionization of this sector. However, a deeper explanation lies in the fact that most self-employed persons as well as workers paid at the piece-rate have to work extended hours per day simply to make a livelihood. But this also could have a limited advantage for some of them insofar as the entire family can participate in the work (e.g. traditional carpet making, weaving and other types of artisan work are often done by many members of the family working together). In this context, we have to make a sharp distinction between labour-service and the labourer providing such service: the same amount of labour service (say, 18 hours per day) may be spread out over several family members working as labourers (say, three). In some cases, each family member (labourer) may on an average have a lighter work load (of only six hours) per day compared to an average worker in the organized industry. This brings us to a somewhat different analytical dimension of disguised unemployment: some persons may be unemployed in a disguised manner not only in the sense of having a very low earning rate i.e. income-wise unemployment but also in the sense of relatively light work-intensity per day, i.e. time-disposition-wise unemployment. And, unless one believes in the neoclassical proposition that income necessarily reflects the marginal product, one would have to devise, a third (and separate) criterion of disguised unemployment in terms of abnormally low productivity of labour. However, given the structure of reward in a capitalist economy, one needs to be careful in applying these concepts. Thus, an ‘important person’ belonging to the board of directors of several large corporations, may be making a well-above-average income by attending only a couple of board meetings per month. Such a person may very well be considered to be disguised unemployed by the time-disposition criterion and even perhaps by the labour-productivity criterion although, he cannot, by any means, be considered unemployed, disguised or not, by the income criterion! Also recall in this context that ‘unproductive labour’ was a common category used in the classical tradition of political economy and, all those engaged in unproductive labour (e.g. ‘priests, prostitutes and professors’ according to a picturesque phrase employed by Rosa Luxemburg) could be considered to be disguised unemployed by the productivity criterion.

In the normal organization of factory work under the capitalist system, the threefold distinction between income-wise, time-wise and productivity-wide disguised unemployment may not be particularly relevant. Thus, an unemployed industrial worker is both income– and time-wise unemployed and of course, he does not have much a chance to be productive either. However, such a distinction can be highly relevant in the context of traditional, family-based agriculture, especially for characterizing such phenomena as rural poverty or the existence of surplus labour. Consider for example a typical rural woman in the poorest strata: in addition to all her other work inside and outside the house, she may have to spend long hours collecting wood for fuel and carrying water home from a distance. Although she has exceptionally hard and long working hours every day and must be considered time – and disposition-wise fully employed and certainly productive in every normal sense of the term, in keeping her family going under most difficult circumstances, in all probability she would not be classified as ‘gainfully employed’ by the income criterion. Indeed her case is the opposite of that our ‘important person’ who has a high income by attending a couple of board meetings every month. It is to be noted that the worst kind of rural poverty is often concentrated among people who are fully employed by the time-disposition criterion, but may be described as disguised unemployed by the income criterion, because of their miserably low earning rate per hour of work. After all, this is what the phrase ‘eking out a living’ usually means.

There can hardly be any serious doubt that in the backward agriculture of many populous countries (e.g. in South Asia), a high proportion of the population engaged in cultivation have extremely low income and, in this sense suffers from disguised unemployment by the income-criterion. Nevertheless, it is far more problematic to identify what such disguised unemployment by the income criterion implies in terms of either the time-disposal or the productivity criterion. If one were to believe in the ideologically potent neoclassical slogan that all ‘factors of production’ including labour always tend to get paid according to their marginal product even in pre-capitalist, backward agriculture, then that proportion of population with extremely low income could be said to be rather unproductively engaged in agriculture. Their low income would be the ‘evidence’ of their low productivity which in turn would imply a corresponding level of disguised unemployment in agriculture. But this would involve implicit theorizing based on the dubious assumption that income (earning) is always positively associated with productivity, even in traditional agriculture.

Such implicit theorizing apart (a sophisticated example of which is the so-called ‘efficiency wage’ hypothesis e.g. Bliss and Stern 1978) the important question remains as to whether there is any meaningful sense in which one can argue about the existence of significant surplus labour and disguised unemployment in backward agriculture, judged by the productivity criterion. This would imply that some surplus labour can be withdrawn from agriculture without adversely affecting the level of agricultural output. Or, in more textbookish jargon, ‘at the margin’ labour contributes nothing to output so that, the marginal product of labour is zero in such agriculture. Put in such general terms, the formulation is too fuzzy to be useful. For instance, if by ‘margin’, one means the intensive margin of higher labour input per unit of land, then considerable empirical evidence exists, at least in India, to suggest that the smaller-sized land holdings usually do use family labour more intensively, both in current agricultural operations and in direct investment of labour for improving land quality. As a result, the total output, taking all crops together over the year, tends to be higher per unit of land on smaller holdings (Bharadwaj 1974, chs. 2, 3 and 7 provide an excellent account). This tendency towards an inverse relation between farm size and productivity per acre in traditional agriculture would tend to cast doubt on the simple-minded proposition that the ‘marginal’ product of labour is zero, especially if the notion of intensive margin is used.

Without going into such finer points of intensive and extensive margin, Schultz (1964, ch. 4) proposed the ‘epidemic test’: the 1918–19 influenza epidemic in India killed 6.2% of the 1918 population and 8.3% of the working population in agriculture (the latter according to Schultz’s estimate). Schultz found that, although the weather conditions were roughly similar in 1916–17 and in 1919–20, in the latter year agricultural output was lower by about 3.8%, providing circumstantial evidence that withdrawal of labour from agriculture did affect output level. However, apart from many statistical and conceptual problems (e.g. the relation between acrage change in the sense of extensive margin and output change which is a resultant of both extensive and intensive margin in his macro-level statistical investigation), this ‘epidemic test’ must be deemed to be over-simplistic despite its apparent ingenuity. At best, it showed that a random x% withdrawal of labour from cultivation did affect the acrage and/or output level. But it does in no way establish the impossibility of selectively withdrawing x% labour through suitable reorganization of agricultural production at the family and regional level (e.g. Sen 1967). And yet, most of the important initial proponents of the ‘surplus labour’ doctrine had in mind such selective (but not random) withdrawal of labour that may be induced by industrialization and expansion in urban employment opportunities (Nurkse 1953; Lewis 1954). And, once it is recognized that such withdrawal of labour from agriculture can be accompanied by reorganization of labour in the family farm through adjusting the hours of work of the family members staying back on the farm or through higher availability of land per cultivating family, it seems plausible to argue analytically (e.g. Takagi 1978) as well as empirically that, labour can usually be released from agriculture without adversely affecting the level of agricultural output. Indeed, post-revolutionary experiences of agrarian reorganization in China and Vietnam demonstrated the possibility of using surplus labour to improve the quality of land through better drainage and irrigation without significant drop in short-run agricultural output, despite all the serious problems of lack of adequate incentive to private production in cooperative and collective agriculture.

See Also