Marx on Absolute and Relative Wages and the Modern Theory of Distribution

Abstract

This paper aims at clarifying some aspects of Marx's analysis of the determinants of wages and of the peculiarity of labour as a commodity, focusing on three related issues. The first is that of Marx's notion of the subsistence (or natural) wage rate: the subsistence wage will be shown to stem, according to Marx, from socially determined conditions of reproduction of an efficient labouring class. The second issue refers to the distinction between the natural and the market wage rate that can be found in Marx, and his critique of Ricardo's analysis of the determinants of the price of labour. Here the ‘law of population peculiar to the capitalist mode of production’ (that is, Marx's industrial reserve army mechanism) will be considered both with respect to cyclical fluctuations of wages and to their trend over time. Moreover, a classification of the social and institutional factors affecting the average wage rate will be advanced. Finally, the last issue concerns Marx's analysis of the effects of technical progress on both absolute and relative wages (that is, the wage share). It will also be discussed by relating it back to the longstanding debate on the Marxian law of the falling rate of profit, and addressing some possible scenarios of the trend of wages and distribution.

Acknowledgment

I wish to thank Roberto Ciccone, Duncan Foley, Fabio Petri and Franklin Serrano for their comments and suggestions.

Notes

¹In modern terms, it is given by the price of the bundle of wage goods that comprise the subsistence of the worker.

²Following Marx (e.g. [1867–94] 1961–63, I, pp. 525–527), an operational definition of subsistence should therefore consider what part of an increase in the bundle of commodities given to the labourers truly corresponds to a surplus wage and what part instead merely compensates a greater or faster consumption of the worker as living machinery.

³Marx is referring here to the price of unskilled labour. If we consider labour-power of specialised skilled workers we must add the cost of the commodities necessary ‘to modify the human organism, so that it may acquire skill and handiness in a given branch of industry’ (Marx, [1867–94] 1961–63, I, p. 172). Notice that in order to calculate the wage rates of skilled workers Marx, unlike Smith ([1776] 1976, Book I, chapter X, pt. 1, pp. 113–114) and modern human capital theory, does not suggest that their wages should also strictly recompense the interests on the sum spent to obtain those ‘expensive machines’ (which would otherwise be the sum invested in other activities).

⁴Ramirez (2007, p. 30) correctly observes that in Marx ([1865] 1975, p. 72), the physical minimum is the ‘ultimate limit’ of wages. But he maintained that in the classical economists we find only a bare physiological minimum, while the necessary wage of Smith and Ricardo contains instead a historical and moral element like that of Marx (see, for example, Smith, [1776] 1976, Book V, chapter II, pp. 393, 399–401). Indeed, their notion of subsistence wage can be said to be similar to that of Sen (1987), who posits a relation between commodities and ‘capabilities’ and observes that in a richer country the collection of commodities needed to satisfy the same ‘capabilities’, so that people can avoid the shame of failing to conform to conventions, will increase or change, just as needs can change.

⁵Therefore, no possibility emerges of ascribing to Marx, as for Ricardo (see for example, Samuelson, 1978; Hicks & Hollander, 1977), J.S. Mill's idea that the wage rate will reach subsistence level only in the far off stationary state in which the rates of change of the capital stock and population are zero.

⁶Some recent interpretations of Smith's theory of wages do not concur with Marx's view that Smith's natural price of labour is the subsistence wage (see for example Stirati, 1994). An argument in favour of Marx's view might be traced (see O'Donnel, 1980) in the concomitant facts that (a) Smith argued that ‘the subsistence of the labourer’ will be ‘very different upon different occasions’—that it will be ‘more liberal’ in a society ‘advancing to opulence than in one standing still; and in one standing still than in one going backwards’ (Smith, [1776] 1976, Book I, chapter V, p. 40); and (b), that at the end of chapter VII, Book I of The Wealth of Nations, he stated his intention to show that in every society the natural rate of wages ‘varies according to their circumstances’, that is ‘according to their riches or poverty, their advancing, stationary, or declining condition’ (Smith, [1776] 1976, Book I, chapter VII, p. 71).

⁷Indeed Ricardo used the Malthusian population principle mainly to refute Malthus's denial that capital accumulation will be accompanied by a fall in the rate of profit due to decreasing returns in agriculture. The population principle allowed Ricardo to distinguish between changes in the price of labour due to decreasing returns in agriculture and those arising from a rate of population growth different from that of capital. But in discussing both the secular trend of wages and the effects of taxation on them, Ricardo recognises that the population adjusts slowly to changes in the capital stock, since labour supply ‘cannot be altered at pleasure’. Hence he introduces a different mechanism of adjustment of the actual price of labour to the normal one, which was based upon the relative strength of the parties involved in wage bargaining and the fear of an increasing social conflict if the real wage rate did not reflect that strength (see for instance Ricardo, 1951–73, VIII, p. 196). It is the mechanism that Malthus ([1836] 1986, p. 181) denied when he argued that it was not Smith's notion of ‘common humanity’ but the principle of supply and demand that would prevent wages from falling below the subsistence wage; for, according to Malthus, if the wage did fall below subsistence, the resulting diminution of the population would quickly restore the wage back to subsistence. Marx's concurrence with Say's opinion that Ricardo ‘determines’ the value of labour ‘by demand and supply’ (Marx, [1862–63] 1978, II, p. 400) therefore seems ill-advised to a large extent (see also Steedman, 1982).

⁸See Smith ([1776] 1976, Book I, Chapter VIII, p. 89) and Ricardo (1951–1973, I, pp. 218–219), who sometimes called this average wage the ‘necessary rate’. Note that in Smith and Ricardo, the subsistence wage can itself change in the course of capital accumulation, thus changing the average rate itself; and that wages can be at the subsistence level also in a ‘progressive state’ of the economy depending upon how quickly the supply of labour adjusts to demand. Smith and Ricardo also recognised that the adjustment of labour supply may take place through immigration, changes in working hours and so on, as well as through changes in the population's birth and mortality rates.

⁹This does not exclude the possibility that temporary variations may be transformed into permanent changes. Thus Marx ([1862–63] 1978, III, p. 313) seems to agree with the Ricardian view expressed by Malthus that if a rise in wages can lead, for a period of time, to an increase in the price of certain necessaries, especially food, then ‘for the less proportional returns to increased tillage’, it is possible that ‘part of that increase of price becomes permanent, and thus it prevented a complete reaction from taking place through the principle of population’.

¹⁰‘In times of prosperity, intense expansion, acceleration and vigour of the reproduction process, labourers are fully employed. Generally, there is also a rise in wages which makes up in some measure for their fall below the average during other periods of the business cycle. At the same time, the revenues of the capitalists grow considerably. Consumption increases generally. Commodity prices also rise regularly…’ (Marx, [1867–94] 1961–63, III, p. 437).

¹¹Thus, according to Blaug (1978, p. 251) and Samuelson (1951), who interpreted Marx through the lens of the neoclassical concept of competition, Marx ought to admit that if there are unemployed people, there might not be a tendency to preserve wages at the value of labour-power.

¹²Hence, to explain why wages do not fall below the necessary level, it is not necessary to appeal to non-competing groups or to the hypothesis that the unemployed are kept out of competition with employed workers by means of their transformation into a stratum of ‘broken and degraded workers’ (see Cottrell & Darity, 1988, p. 179), although this phenomenon no doubt occurs. On the other hand, as Green (1991, p. 206) has noted, labour supply is viewed by Marx as particularly elastic ‘in the region of the value of labour-power’. The factors enforcing a degree of wage rigidity are indeed what Smith called the common humanity, tacit combinations among workers and capitalists, the intervention of the State and so on.

¹³According to Marx in this way only 1/2; 1/3, 1/4, … 1/8, … of total capital will be transformed into labour-power, and the remainder into means of production; that is, over time the proportion will be not 1:1, but 2:1; 3:1; 4:1 and so on. In other words, the increase in total capital ought to be progressively greater in order to absorb a given increase in the working-class population. Note that Samuelson's (1957, 886n) critique of Marx not having a quantitative equation to explain why employment is as high as it is appears misleading. The idea that there exists a natural rate of growth determined by the rate of increase of population to which the growth rate of employment must adapt rests in fact on the neoclassical substitution mechanism, which cannot be found in Marx, or in the classical economists.

¹⁴Of course when considering heterogeneous labour, it is possible that ‘one section of the workers starves, another section may be better fed and clothed, as may also the unproductive workers and the middle strata between worker and capitalist’ (Marx, [1862–63] 1978, II, p. 561). It is also possible that an increase in the wage rate of a group of workers does not lead to a decrease in the rate of profit, but of the wages of the other workers. Thus, in this case preserving a given scale of differential wage rates becomes a way of obtaining a general increase in wages. For a Marxian analysis of the relationships among workers and of differential wages see Bowles & Gintis (1975), Braverman (1974), Edwards et al. (1982), Elbaum et al. (1979). On differential wages, see also Kurz & Salvadori (1995, pp. 325–338).

¹⁵This does not preclude that circumstances may arise that can reduce subsistence itself, as when there is a large-scale immigration of workers accustomed to a lower standard of living, or the economy falls into a retrograde state.

¹⁶This should convince Gottheil (1962) and other interpreters of Marx that he did not forecast an absolute, but a relative, impoverishment of the working class. See also Lapides (1998), Lebowitz (2003), Meek (1967) and Sowell (1960), who discuss the evolution of Marx's views on this issue.

¹⁷According to Gillman (1957) and Dobb (1945) Marx had picked up the law of the tendency of the profit rate to fall from Smith and Ricardo, and had doubts about its empirical validity. But the fact that Marx ([1867–94] 1961–63, III, p. 233) spoke of the difficulty of explaining why in actual fact the rate of profit does not fall even more seems to indicate that he believed that that law would manifest itself empirically.

¹⁸Marx, who had analysed capital-saving technical progress in depth, observed that ‘the concentration of means of production yields a saving on buildings of various kinds, not only for the actual workshops, but also for storage etc. The same applies to expenditures for fuel, lighting etc’ (Marx, [1867–94] 1961–63, III, p. 79; see also, pp. 82, 96 and 103).

¹⁹Note that Marx assumes that technical progress will lead to a fall in the maximum rate of profit R, that is, of the ratio of living to dead labour, or (S + V)/C, where S is the surplus-value, V the variable capital and C the constant capital. Then, since C/V = (1 + s)/R, where s is the rate of surplus value, an increase in s and a decrease in R must necessarily be associated in Marx with an increase in C/V.

²⁰The inversion point would be increasingly distant the lower the initial rate of surplus-value S/V and the organic composition of capital C/V. Assuming on the other hand a falling maximum rate of profit, the rate of profit would necessarily fall at a certain point in time (see Marx, [1867–94] 1961–63, III, pp. 234, 255).

²¹The usual example given is Smith's analysis of the relation between the extension of the market and the division of labour. On the other hand, both the resources for financing inventions and the introduction and success of innovations seem to be linked to the actual and expected demand (cf. Schmookler, 1966).The stimulus to innovate, however, also arises from the struggle to survive, which explains why, during crises, there is a strong tendency to introduce process innovation.

²²We are not referring here to the case envisaged by Marx regarding a fall in the average rate of profit for a given wage rate, since, as noted by Okishio (1961), a new method of production, if adopted, increases the wage rate for a given rate of profit even if the organic composition of capital increases. Otherwise it is not introduced, as there would not be a reduction in the costs of production. Indeed, technical change may cause both labour productivity and the maximum rate of profit to increase, or the technique with a lower labour productivity may be that associated with lower costs of production at a given rate of profit. Only if there is the same organic composition of capital in all industries, is the technique that minimises the cost of production necessarily that associated with a greater labour productivity (see Steedman, 1977).

²³Here we assume that wages are paid post factum, but none of our conclusions would change if wages were paid ex ante.

²⁴This case resembles the mechanisation of production as defined by Schefold (1976), which in turn is similar to technical progress as considered by Marx when he argues that the maximum rate of profit tends to fall.

²⁵We are assuming here that, as in Marx, both 1/R and ω increase with the introduction of the new technique. The same result could be illustrated by adopting the basket of consumption goods as the numeraire of the price system, and introducing hypotheses regarding the use of the net product and the consumption-steady growth rate relationship (see Kurz & Salvadori, 1995, pp. 114–116 for a graphic treatment of the wage share in this case). Also to be noted is that, unlike what was advanced by Marx, the new dominant technique might be characterised by both a lower labour productivity and a lower maximum rate of profit. Paradoxically, in this case, since with the new technique both the rate of profit and the wage share increase, an unlikely fall in the wage rate (and a further increase in r) would have to occur to keep the wage share unchanged.

²⁶It is worth noting here that Sraffa viewed Marx's arguments on the law of the falling rate of profit as referring to cases in which no proper technical progress (i.e., inventions) is taking place (see Gehrke & Kurz, 2006).

²⁷On the influence of political factors, see Kalecki (1943). With respect to the effect of technical progress, Marx himself argued that the pressure of competition will compel capitalists to invest and innovate, and that technological waves (see, for example, Marx, 1867−[94] 1961–63, III, pp. 70–71) and radical innovations exist which (directly or indirectly) fuel the rate of accumulation.

²⁸In these models, assuming, for the sake of simplicity, s _w  = 0, since g* = s _p r* = s/C* (where g* is the warranted rate of growth, C* is the desired capital-output ratio, r* is the normal rate of profit, s is the overall propensity to save, and s _w and s _p are respectively the propensities to save of the workers and of the capitalists), it is argued that, if r* increases, g* increases. But the rate of growth g* actually only reflects income distribution (cf. Vianello, 1985; Garegnani, 1992), and has no relation to the actual pace of accumulation, which may differ from g*, as the actual capital/output ratio C, although tending to C*, may differ from it.

²⁹This argument does not rule out the possibility that a decrease in the rate of profit in a particular country relative to the profit rates in other countries may have a negative effect on investment in that former country as an aspect of international competition.

³⁰Here we can see why the normal rate of profit is not satisfactorily explained by the rate of accumulation as suggested in some Post-Keynesian models: an increase in the actual pace of accumulation is, in effect, ‘financed’ by an increase in output per unit of capital and in the amount of productive capacity, given the wage rate and the methods of production (see Garegnani, 1992).

³¹Since the rate of profit is equal to r = (Π/Y)(Y/Y _f )(Y _f /K)  =  (1− w/π)(u/k), where π is labour productivity, Y _f the real full capacity output, u the degree of capacity utilisation, and k the ratio of the stock of capital to Y _f , then given w and π and k, r increases if u increases. In particular, assuming s _p  =  1, it is argued that the rate of accumulation h will be positively related to the discrepancies between the actual (u) and desired (u*) degree of capacity utilisation, while the savings per unit of capital will be equal to r. Thus, if du/dw > 0, then dh ^m /dw = dr/dw > 0, with h ^m  = r = (1− w/π)u ^m , where h ^m and u ^m are respectively the average rate of accumulation and the average degree of capital utilisation, which may differ from the normal utilisation rate u* corresponding to the desired or normal capital/output ratio.

³²The possibility of wages sharing in the net product is implicit in a notion of subsistence as historically determined. On the fact that we need not take the rate of profit as the independent variable in the price system if the wage rate is not at the subsistence level, see Levrero (2000).

³³Given the target real interest rate r ^T and the expected rate of inflation p ^a , the money interest rate i* should thus be such that (1 + i*) = (1 + r ^T )(1 + p ^a ). This implies that, if the actual inflation rate p is equal to the expected one, and i = i*, then the real rate of interest r = (i*− p)/(1 + p) = r ^T . Of course, if the workers' target real wages happen to be incompatible with the target rate r ^T , then a change in the inflation rate will be set up if no change occurs in the normal profits of enterprise or the technical conditions of production. Note that a different money interest rate i* should be viewed in the light of the price index chosen by the monetary authorities to calculate the actual and expected inflation rate.

³⁴For an analysis in this direction see Levrero (1999) and Levrero & Stirati (2004).