Unemployment has many facets. Let us just look at technological unemployment. This is basically due to careless exposition in some elementary textbooks the casual student of economics is frequently left with the impression that technological innovation cannot lead to unemployment.
THE UNFORTUNATE Luddites, imagining themselves displaced by machinery, being the victims of vulgar error rather than historical circumstances. In extenuation of this belief in the impossibility of technological unemployment it must be conceded that until not so long ago the dominant view was, and had been for over a century, that neither could unemployment arise from a deficiency of total demand. For with the great equilibrium systems of the Classical economists in mind, it was not easy to envisage underemployment equilibrium.
The existence of involuntary unemployment, it was believed, exerted a downward pressure on money wages. And if wages were in the process of decline the position, by definition, could not be one of equilibrium. Moreover, the gradual decline in money wages, it was argued, tended to restore full employment, since any reduction in money wages implied a corresponding reduction in the price level and, through the operation (1) of the cash-balance effect in reducing interest rates that were imperfectly stabilized by speculative activity; and (2) of the asset-expenditure effect in reducing the real volume of saving, the level of employment would rise until all underemployment-pressure on prices ceased.
The conclusion that involuntary unemployment could exist only in the short run did not, however, satisfy Keynes, simply because the time required for market forces effectively to reduce such unemployment might be many years. The long-run solution was therefore irrelevant: 'In the long run we are all dead!' The practical thing was for the government to intervene immediately in stimulating effective demand by the several means open to it.
There is something of this older complacency in the popular fallacy that technological unemployment cannot exist. Thus (even if we ignore production functions with regions of negative marginal returns to the abundant factor) a labour-saving innovation may well be such that the existing supply of labour in the economy could be employed only at a real wage materially below that in existence. In a competitive economy in which wages were completely flexible the new technique and the new wage would come to prevail and anyone who attempted to continue with the old technique would (paying the new wage and rentals) cost himself out of the market? Again, therefore, in some 'long run', full employment would be restored at this lower absolute wage. But if, for institutional reasons, there were determined attempts to maintain the real wage at its pre-automation level, technological unemployment could continue for many years.
Even this argument, however, concedes too much, for two reasons: (1) because with the new methods of production, factor substitution may be negligible. True, there may be several alternative ways of producing a set of goods, but if they are all highly labour-saving then unemployment with zero wages, as the market solution, is consistent with each method - at least, such a solution would prevail in a market with completely flexible factor prices. However, if institutional forces could determine wages and rentals, unemployed labour would continue to exist until capital accumulation reduced the labour-capital ratio to the point at which the existing supply of labour could be employed again at positive wages.
A possibly 'academic' solution for highly labour-saving innovations might be to reduce drastically the hours put in by each labourer, though such a solution might well act to reduce technical efficiency.The other reason (2) is that the assumption of homogeneous labour is too abstract a concept for the relevant economic models. In automation, complex capital equipment is combined with skilled 'brain labour' both in the producing and the operating of the machinery: the whole idea is to dispense not only with unskilled labour but also with highly skilled manual labour and clerical labour.
Thus, the proportion of such labour required in any automated process must be expected to decline with the evolution of technology.Again, it is open to people to be complacent and talk of a long-run 'transformation' of old-type manual labour into new- type brain labour - though even here, since substitutability is likely to be small, it may be a long time before there were sufficient capital to employ the transformed supply of labour. However, it is far from being impossible that in the not-too-distant future a large proportion of the adult population would be unemployable simply because they would not be endowed with the innate capacities necessary to acquire the highly developed mental skills which may be called for by a more complex technology. Of course, they may all be comfortably maintained by some transfer of the increased output produced by the reduced working population. There need be no economic crisis, only a social one.