by Utsa Patnaik
Some assumption of this sort is essential to make for a paper money world, once we recognize the wealth demand for money and abandon the view that the value of money is determined by its demand and supply. Since money in such a world is not a produced commodity and is not even convertible at a fixed exchange rate into a produced commodity like gold or silver, there is no cost of production to provide any anchorage to the value of money. And if demand and supply play no role in determining its value, then the only way for there to be a determinate value of money is if this value is tied to one particular commodity. And a commodity that has to be used for producing all other commodities is an ideal candidate for this purpose. Labor-power is such a commodity. Hence money wages being given in any period makes eminent sense in this context.
A problem, however, arises from the fact that the stability of the value of money, which is essential for it to play the role of a wealth-form, then becomes dependent upon the stability of the money-wage rate or the “wage unit” as Keynes called it. And in the isolated capitalist sector that is the conceptual universe for analysis there is no reason to expect the wage unit to be constant irrespective of the level of employment, whether within a given period or across periods.
This is precisely the assumption that Keynes made in the context of his single period analysis and it is crucial for his theory. The whole point of his theory was that a capitalist economy can settle down at any level of employment, depending upon the level of aggregate demand; and since unemployment was socially unacceptable, the state should nudge the economy toward a higher level of employment. But if a higher level of employment meant a higher money wage, then the stability of the wage unit, and with it the entire role of money as a form of wealth-holding, would get jeopardized.
On the other hand, since Keynes accepted the equality of real wages and the marginal product of labor at the “equilibrium” level of employment where the economy settled (given the level of aggregate demand), higher levels of employment would necessarily mean lower real wages (as the marginal product of labor curve was assumed to be downward-sloping). This meant assuming that the workers would be willing to accept the same money-wage rate no matter what the level of employment, even though the real wage rate would be different at different levels of unemployment.
There was thus a basic contradiction between Keynes’s assumption that multiple employment equilibria were possible in the single period he was considering, and that the wage unit nonetheless remained stable in this period. This is the contradiction that was later exploited by his critics to bring about a revival of monetarism.
Keynes glossed over this contradiction, through postulating a “money illusion” that afflicted the workers. He put forward the proposition that any lowering of real wages as employment increased would go unnoticed by the workers as long as the money wages remained unchanged, because workers were focused only upon the money-wage rate—wherein lay their “money illusion.” If their real wage rate were to be cut through a cut in the money-wage rate, then they would resist it, but they would not notice a cut in real wage rate through a rise in prices at a given money-wage rate.
Keynes thus sought to reconcile, through his theory of money illusion, the requirement that the wage unit should be stable in a money-using economy with his theory that such an economy could settle at any one of a range of levels of employment, and therefore should be pushed toward full employment by state intervention in aggregate demand. The Phillips Curve supported Keynes’s idea of a “money illusion” afflicting the workers.
This last point can be seen as follows: If the rate of growth of money wages is a function of the unemployment rate, that is, [(dw/dt)/w] = f(u), and if the rate of growth of prices is the same as the rate of growth of unit labor cost, that is, [(dp/dt)]/p = [(dw/dt)/w] – β, where β is the rate of growth of labor productivity, then it follows that at every level of unemployment the workers’ effort to raise the wage share of output is frustrated. Real wage growth can never rise above productivity growth, even though in demanding and obtaining money wage increases, this is what workers are aiming to achieve. They believe that prices will not change while obtaining money wage increases, but this is precisely what happens, preventing them from getting any real wage increases in excess of productivity growth.
Taking β to be zero, what this means is that the workers never actually get any real wage increase, even though the whole point of their demanding and getting higher money wage increases was to raise their real wage. They are frustrated by price increase, and yet the Phillips Curve postulates that they never notice the price increase, for if they did then they would anticipate inflation and incorporate it into their money-wage demand, which would make inflation accelerate. The basic underlying assumption behind the Phillips Curve, as is well known, is static price expectations on the part of the workers, which means that they do not notice inflation and therefore do not expect it. This is a form of “money illusion.”
But it is obvious that though “money illusion” among workers may prevail temporarily, it cannot be assumed to be a permanent feature of a capitalist economy, in which case Keynesianism gets into difficulties, a point perceived by Joan Robinson when she postulated an “inflationary barrier.”10 She saw that the rate of unemployment could not fall below a certain level, the level at which unbounded inflation set in. That is, unbounded inflation set a floor for the level of unemployment to which an increase in aggregate demand (including through state intervention) could push the economy.
The “inflationary barrier,” however, does not do away with the idea of “money illusion.” All it suggests is that at some level of unemployment, when the real wages have fallen sufficiently low, the fall will be noted, the “illusion” will break, and workers will wake up to demand higher money wages, which, if offset by price increases, will go on and on rising, resulting in a veritable explosion. Before the point at which the “illusion” breaks, it must be there.
Put differently, the following three propositions, each of which characterizes the Keynesian system, cannot logically hold together without workers being afflicted by “money illusion” at least over a certain range: (1) Real wages = marginal productivity of labor (with a downward-sloping MPL curve); (2) money wages are given in the period in question; and (3) the level of employment can be altered by state intervention through altering aggregate demand. And if “money illusion” is seen to be an unreal phenomenon, then this makes the Keynesian system, at least in its original form as developed by Keynes,11 logically untenable.
One can, of course, go along with the Kaleckian formulation, rather than the Keynesian one, and do away with number 1 above. One can postulate instead that the price is a markup over unit prime cost, which in the case of an isolated capitalist economy producing ex hypothesi all its own raw materials internally, is simply the unit labor cost that is constant as long as all production coefficients are given and unchanging. The fixity of the money-wage rate even in this case, however, cannot hold since the workers’ real wage demand (for given productivity) will be negatively related to the rate of unemployment, with there being a certain floor real wage rate which it cannot fall below. At the level of unemployment where the real wage demand begins to exceed this floor, money wages will rise, and since they will be “passed on” in the form of higher prices, accelerating inflation will ensue (because there is no “money illusion”). But this level of unemployment will then become the floor below which unemployment cannot fall (analogous to the “inflationary barrier”). And since it is at this level of unemployment that money wages start increasing through workers’ bargaining for higher real wages, this floor unemployment rate can be quite high, which means that the capacity of the government to push down unemployment remains correspondingly constricted.
But for the same reason that pushing down unemployment remains constricted, any buoyancy in aggregate demand that pushes the unemployment level below this threshold also pushes the economy into accelerating inflation. In an is
olated capitalist economy with no regulation of aggregate demand, there is nothing to prevent such accelerating inflation.
We can distinguish here between two cases. One, where at the prevailing level of aggregate demand and hence output, there is both unutilized capacity and unemployment above this threshold rate the wage unit will be stable in such a case. Two, where at the prevailing level of aggregate demand, there is either an ex ante excess demand at full capacity or an ex ante tendency for unemployment to fall below the threshold level. In this case the wage unit will get destabilized through accelerating inflation.
The stability of the wage unit therefore requires both things: the maintenance of unutilized capacity (so that there is no tendency toward a rise in prices through demand-pull pressures) and the maintenance of sizable unemployment, above the threshold where money wages start rising. This dual requirement is impossible to meet in a capitalist economy in which aggregate demand is not planned by some central authority.
Besides, if both these conditions held then we would never witness any increase in money wages above the rise in labor productivity, for that would simply engender accelerating inflation. Since we do see increases in money wages above the rate of growth of labor productivity, and yet no accelerating inflation, as Phillips has shown through his empirical analysis, it follows that unless we believe in “money illusion,” the conceptualization of capitalism as an isolated system with only workers and capitalists is wrong.
Even the Keynesian analysis of a money-using economy, which recognizes its characteristics so much better than the other theoretical systems we have analyzed, the Walrasian, Marshallian, and Ricardian, is fraught with logical contradictions. True, these would not arise if we believe with Keynes that workers suffer from “money illusion,” but once we abandon this concept as unreal (thereby rendering Keynes’s original system logically untenable) and introduce an alternative pricing-distribution system in its place that is compatible with “involuntary unemployment,” it turns out that the stability of the wage unit, without which a money-using economy cannot survive, becomes difficult to explain. Clearly the problem lies with the conceptual universe in terms of which we are looking at capitalism.
Capitalism cannot be conceived as an isolated system consisting only of workers and capitalists, as economic theory in all its different traditions has done. But before going further into how it should be conceptualized, we will examine the Marxian theoretical system in the next chapter.
CHAPTER 3
The Marxian System and Money
The Marxian system, like the Keynesian, recognizes that people hold a part of their wealth in the form of money, and consequently Say’s Law does not hold and involuntary unemployment can and does occur in such an economy. It also recognizes that because there is a wealth demand for money, the value of money vis-à-vis the world of commodities cannot be determined by the demand for and the supply of money. Rather, it is fixed independently of its demand and supply.1 Since Marx, unlike Keynes, was talking not about a paper money or credit money world, but about a commodity money world (with paper money being statutorily convertible into the money commodity at some fixed rate), he saw the relative amounts of labor directly or indirectly embodied in a unit of the money commodity, as compared with a unit of the basket of non-money commodities being produced, as the determinant of the value of money. In short, he saw the value of money vis-à-vis commodities as being determined by the labor theory of value.
Ricardo and Marx on Money and Value
It is often taken for granted that the labor theory of value of Marx is more or less identical with that of Ricardo. This is untrue. Indeed, it simply cannot be true in view of Marx’s total rejection of Say’s Law, which Ricardo accepted. Rejection of Say’s Law entails accepting that money has a wealth demand, quite apart from its transaction demand; the magnitude of this wealth demand can change even when there is no change in the value of the commodities being circulated through the medium of money. Marx saw people hoarding money, quite apart from the money that was circulating, the size of which could fluctuate for independent reasons having nothing to do with the value of the total magnitude of commodities circulating. If there is such a wealth demand for money, then an increase in the supply of money also need not enter immediately into circulation (at least not the whole of it) and create additional demand for commodities, raising their prices in the short run, as Ricardo believed.
The value of money, or its obverse, the price level of commodities, does not respond to an increase in the supply of money, either fully (as monetarism, including that of Ricardo, would suggest), or even at all, since all the addition to the supply of money may simply be added to the hoard that people hold. Correspondingly, if there is an increase in the transaction demand for money owing to an increase in the value of commodities circulating in the economy, this need not cause an increase in the value of money, since this extra transactions demand may be accommodated through a fall in the relative size of the hoard.
What is more, even if the supply of money remains unchanged, as does the supply of commodities that have to be circulated with the help of money, a desire on the part of people to hoard a larger amount of money, that is, not throw it back into circulation, can lead to an ex ante change in the aggregate demand for commodities and hence in their price level (to be followed by quantity adjustment). Since neither an increase in the supply of commodities (and hence in the transaction demand for money required to circulate them) nor an increase in the supply of money need make any difference to the value of money, it follows that in a money-using economy the value of money has to be explained by some factor other than the demand for and supply of money. Then, whatever effect the changes in the demand and supply of money have on the economy has to be examined independently. Supply and demand do not provide any anchorage to the value of money.
This other factor determining the value of money cannot be the obverse of the “prices of production” of commodities, as in Ricardo. This is because monetarism, as we have seen, is a necessary ingredient of Ricardo’s theory, indeed of any theory that determines the value of money in terms of these prices of production (or “Sraffa prices,” as some may prefer to call them).
It follows that Ricardo’s acceptance of monetarism even in the short run cannot characterize the Marxian system. This in turn means that money in the Marxian system cannot be part of the charmed circle of commodities across which the rate of profit and the wage rate are equalized in a free competition economy. The value of money has to be independently determined, and the determination of all the prices of production of non-money commodities has to be based on this independent specification. The relative exchange rate between money and non-money commodities in Marx is not determined as the relative exchange rates within the world of nonmoney commodities, but is determined independently, and underlies the latter. This determination is by the relative quantities of labor embodied directly and indirectly in a unit of the money commodity relative to a unit of the basket of non-money commodities.
It follows from this that when the prices of production deviate from labor values in the case of the non-money commodities, this does not mean any change in the rate of exchange between money and the nonmoney commodities, which, in the Marxian system, continues to equal the relative labor values. This exchange rate remains unchanged, unaffected by the transformation of values into prices; rather it constitutes the basis on which the transformation of values into prices within the non-money commodities world occurs.
Putting it differently, even if labor values do not actually determine the long run (or “center-of-gravity”) equilibrium prices among the non-money commodities, the relative quantities of labor embodied still constitute the determinant of the exchange ratio between money and the world of nonmoney commodities in the Marxian system. The prices of commodities in terms of money, in other words, are determined on the prior specification of an independent meta-rule that determines the value of money. The labor theory of val
ue provides such a meta-rule.
An example will clarify the point. If 100 units of the money commodity require as much labor directly and indirectly as the total output of nonmoney commodities, then the money value of the latter is 100; or the value of money in terms of the basket of commodities produced is .01. If the money wage rate is 0.5, then the prices of production can be determined, given this value of money. If the value of money was not given by this metarule, then, unlike in the Ricardian system where money enters the charmed circle of commodities across which the wage rate and the rate of profit are equalized, prices of production could not be determined.
Of course, the exchange ratio between money and the basket of other commodities need not empirically correspond to the relative quantities of labor directly and indirectly embodied in a unit of each. This empirical fact may suggest that we need an alternative explanation for this exchange ratio. But any different explanation can always be subsumed under a relative labor–embodied explanation by bringing in, say, the monopoly rent of gold mine owners, which may explain why the gold-versus-commodity ratio is not exactly identical with the relative quantities of labor embodied.
In other words, a fixed ratio between money and the world of nonmoney commodities, even if it does not empirically correspond to the relative quantities of labor embodied, can still be explained as being based upon it. It is this fixed ratio that is an essential part of Marx’s theory, as underlying even the determination of the prices of production of the nonmoney commodities on the basis of given money wages. Thus there is a fundamental difference between Ricardo and Marx on the question of money and value theory. This in turn arises from their difference over the wealth demand for money, and hence Say’s Law, and the possibility of generalized ex ante overproduction.