6.8 Labor supply falls as the wage rises.

Since an increase in wages will make workers better off, individual workers are just as likely to work fewer hours as more when the wage rate increases. Individual labor supply curves are just as likely then to slope backwards showing falling supply as wages rise as they are to slope forwards.

6.9 An individual labor supply curve derived from extreme and midrange wage levels At the aggregate level, a labor supply curve derived by summing many such individual supply curves could have any shape at all. There could be multiple intersections of the supply curve with the demand curve (accepting, for the moment, that a downward-sloping demand curve is valid). There may be more than one equilibrium wage rate, and who is to say which one is valid? There is therefore no basis on which the aggregate amount of labor that workers wish to supply can be unambiguously related to the wage offered. Economic theory thus fails to prove that employment is determined by supply and demand, and reinforces the real-world observation that involuntary unemployment can exist: that the employment offered by firms can be less than the labor offered by workers, and that reducing the wage won't necessarily reduce the gap.

This imperfection in the theory the possibility of backward-bending labor supply curves is sometimes pointed out to students of economics, but then glossed over with the a.s.sumption that, in general, labor supply curves will be upward sloping. But there is no theoretical or empirical justification for this a.s.sumption.

This strong a.s.sumption would be of little consequence if economists didn't derive such strong conclusions from their model of the labor market. Declarations that minimum wage legislation is ineffective and causes unemployment, or that demand management policies can't alter the rate of unemployment, are hardly insignificant p.r.o.nouncements. Their truth is dependent in part on the supply curve for labor being upward sloping.

6.10 An unstable labor market stabilized by minimum wage legislation For example, if the aggregate demand and supply curves for labor both slope downwards, then the 'equilibrium' of the two could be unstable: falling supply could be met by falling demand, resulting in runaway unemployment. Putting a floor to this process via a minimum wage could actually make the labor market stable and decrease unemployment.

Didactic policy positions should be based upon robust intellectual or empirical foundations, rather than the flimsy substrate of mere fancy. Neocla.s.sical economists are quite p.r.o.ne to dismissing alternative perspectives on labor market policy on this very basis that they lack any theoretical or empirical foundations. Yet their own policy positions on the labor market are based as much on wishful thinking as on wisdom.

Monopoly and monopsony.

The conclusion that workers receive the value of their marginal contribution to output depends upon the a.s.sumption that both the product market and the labor market are perfectly compet.i.tive. The notion of perfect compet.i.tion has already been debunked, but even if it were intellectually sound, it is clearly a dubious thing to a.s.sume for an overall economy.

If we instead accept that in practice both product and labor markets will not be perfectly compet.i.tive, then economic theory predicts that workers will not, in general, receive the value of their marginal contribution to production. In this more general case, economic theory concedes that workers' incomes are determined not only by their contribution to production, but also by the relative bargaining power of workers and employers.

Let's first consider the case in which the product market is not perfectly compet.i.tive: workers are being hired by firms that have to reduce their average selling price to increase output. In this case, the price received per unit falls as output increases. Marginal revenue is thus less than price, and the worker's marginal revenue product is the product of marginal revenue and marginal productivity.

One ironic consequence of this a.n.a.lysis given how vehemently anti-union most neocla.s.sical economists are is that neocla.s.sical theory can be shown to favor the existence of trade unions. Without trade unions, the labor supply will be compet.i.tive and will therefore be 'exploited,' because the wage will be less than the price for which the marginal worker's output can be sold. With a trade union acting as a single seller of labor, however, the price charged for each additional worker will rise as more workers are hired. This situation known as a monopsony or single seller means that the marginal cost of supply lies above the supply curve.

With a monopoly seller of labor confronting non-compet.i.tive purchasers of labor, the wage is indeterminate. It will lie between the minimum set by the marginal revenue product of labor (which means that firms are exploiting workers), and the maximum set by the rising marginal cost of workers (which means that workers are exploiting firms). The final position will be determined by the relative bargaining power of the two groups, which cannot be determined by the market.

Thus while economists normally portray unions as bad because they restrict compet.i.tion in the labor market, this may be a preferable situation to leaving compet.i.tive workers to be exploited by less than perfectly compet.i.tive hirers of labor.

Sraffa's observations on aggregation.

You will remember from Chapter 5 that Sraffa had two criticisms of economic demand and supply a.n.a.lysis: one for a broad definition of an industry, the other for a narrow definition. The labor market is clearly a broadly defined industry, and Sraffa's first critique is therefore relevant to it.

The critique was that, with a broad definition of an industry, it is not feasible to draw independent demand and supply curves, since any change in supply will have income distributional effects which will in turn alter demand.

This is clearly the case when the supply curve refers to the entire labor force. Remember that the aggregate demand curve, in this market, is supposed to represent the aggregate marginal revenue product for labor. This in turn is a product of physical labor productivity on the one hand, and the price for which output produced by that labor is sold.

If an increase in supply requires an increase in the price of labor if, in other words, the supply curve for labor is upward sloping then this is clearly going to alter income distribution, the demand for commodities, and hence their prices. This means that a different 'demand curve' for labor will apply at every different point along a labor supply curve.

This means that multiple equilibria will exist, none of which can be said to be more fundamental than any other. It is also quite feasible that 'perverse' outcomes will apply: that, for example, a higher wage could be a.s.sociated with a higher level of employment rather than a lower one (this dilemma is explored in detail in Chapter 7, in the context of the demand for capital).

6.11 Interdependence of labor supply and demand via the income distributional effects of wage changes The economist's ubiquitous tool of supply and demand a.n.a.lysis is therefore particularly unsuited to a.n.a.lyzing this crucial market.

Freedom and labor.

The vision of a worker deciding how many hours to work on the basis of his preferences between income and leisure, and offering more labor as the wage rises, is, like so much else of economic theory, superficially appealing. But, again like so much else in economics, it implicitly raises a question which undermines the superficial appeal. In this case, the question is 'how can one enjoy leisure time without income?'

If there is a positive relationship between the wage rate and hours worked, then as the wage rate falls, so too will the number of hours worked. As a result, income the product of the hourly wage times the number of hours worked falls even faster. So according to economists, a fall in the wage rate should mean that workers will substantially reduce their incomes, and simultaneously devote more time to 'leisure activities.'

In reality, the only 'leisure activity' which one can devote more time to with less income is sleeping (just ask a homeless person). Most leisure activities are just that active and cost money. The only way that workers could behave as economics fantasizes is if they have alternative sources of income.

This in effect is the economic vision of a worker: someone who has alternative means to generate income at his disposal, and has to be enticed by the wage to undertake wage labor for an employer over the alternative of working for himself.

For that choice to be a reality, workers need something else: capital, his own means of production.

Some workers are so endowed. Some farmers can be enticed into working as farm laborers if the wage is high enough, and if it's not, then they can work their own land. Some office workers have the alternative of working for a wage, or operating as independent consultants out of their home offices. Some 'wage slaves' can make the transition from employee to employer by an innovative idea, hard work, good luck, skill or good timing or fraud.

But the majority do not have that choice or rather don't have it to the degree that they could avoid bankruptcy or starvation by turning to self-employment. For this majority, work is not an option but in the absence of a very generous social security system a necessity. Rather than smoothly choosing between work and leisure, in a completely free market system they face the choice of either working or starving. In a market economy attenuated by the welfare state, this choice is less stark, but still present.

A three-horse race.

This point will become clearer in later chapters, when I outline the monetary approach to economics that I take, in which bankers are treated as a separate social cla.s.s to capitalists. The precis for now is that bankers' incomes depend on the level of debt, and if a Ponzi scheme develops, then the level of debt can escalate dramatically. This then transfers income from both workers and capitalists to bankers, and to the detriment of society in general since it also normally results in a lower level of real investment.

This issue might seem arcane now, but it has serious implications during a financial crisis, such as the one we are currently in. Neocla.s.sical efforts to get out of such a crisis once they've gotten over the shock of one actually happening, and revert to form after behaving like 'born-again Keynesians' when the crisis begins invariably argue that wages have to fall to end the crisis, because high employment clearly indicates that wages are too high.

In fact, policies based on this notion actually make a debt deflation worse, because they drive down the general price level and actually increase the debt burden on society. What is really needed is not lower wages, but lower debt levels and paradoxically that can be achieved by increasing wages. A boost to money wages during a depression can cause inflation far more effectively than 'printing money,' and this inflation can reduce the real debt burden.

If such a policy is ever proposed, you can bet your bottom dollar that the main opposition to it will come from neocla.s.sical economists and their advice, as always, will be wrong.

'A benevolent central authority'

I've saved the unkindest cut of all for last: even though neocla.s.sical economists are normally vehement opponents of the redistribution of income by the state everything, they normally argue, should be decided by the market their own theory of demand and supply only works if, and only if, a 'benevolent central authority' (Mas-Colell et al. 1995: 117) redistributes income in order to 'keep the ethical worth of each person's marginal dollar equal' (Samuelson 1956: 21).

This nonsensical condition is yet another 'proof by contradiction' that neocla.s.sical economics is unsound. Starting from the a.s.sumption that the market economy maximizes social welfare, it concludes that this is possible only if, prior to the market operating, a dictatorship redistributes wealth so that everyone in society is happy with the resulting distribution.

This is, of course, absurd. Rather than using neocla.s.sical economics to justify dictatorships, that neocla.s.sical theory literally needs a dictatorship to make its model work is a reason to abandon neocla.s.sical theory. The fact that neocla.s.sical economists not only cling to their theory but argue against income redistribution in policy debates also shows how little they understand their own theory.

Normally this happens because the a.n.a.lysis that establishes bizarre results like this is only in the journal literature that most neocla.s.sical economists don't read in this case, Samuelson's 1956 paper 'Social indifference curves.' However, here I have to thank Andreu Mas-Colell and colleagues for putting this nonsense in their market-dominating PhD textbook Microeconomic Theory, which makes it impossible for neocla.s.sical economists to hide behind their ignorance of their own literature. This section is worth reiterating here, even though I previously cited some of it in Chapter 3: For it to be correct to treat aggregate demand as we did individual demand [...] there must be a positive representative consumer. However, although this is a necessary condition for the property of the aggregate demand that we seek, it is not sufficient. We also need to be able to a.s.sign welfare significance to this fictional individual's demand function. This will lead to the definition of a normative representative consumer. To do so, however, we first have to be more specific about what we mean by the term social welfare. We accomplish this by introducing the concept of a social welfare function [...]

The idea behind a social welfare function is that it accurately expresses society's judgments on how individual utilities have to be compared to produce an ordering of possible social outcomes [...] Let us now hypothesize that there is a process, a benevolent central authority perhaps, that [...] redistributes wealth in order to maximize social welfare [...] this indirect utility function provides a positive representative consumer for the aggregate demand function [...]

If there is a normative representative consumer, the preferences of this consumer have welfare significance and the aggregate demand function can be used to make welfare judgments [...] In doing so however, it should never be forgotten that a given wealth distribution rule is being adhered to and that the 'level of wealth' should always be understood as the 'optimally distributed level of wealth.' (Mas-Colell et al. 1995: 11618; emphases added) Ahem; please, stop snoring that was important! In the turgid and boring prose of a neocla.s.sical textbook and one which has been used in the training of virtually every American PhD student since the late 1990s you've just been told that neocla.s.sical economics has to a.s.sume the existence of a dictator (benevolent of course!).

Most neocla.s.sical economists don't realize this if they did, they would, I hope, abandon the neocla.s.sical approach as a waste of time. But instead it's likely they don't even read this section of their 1,000-page instruction manual, let alone realize the import of what it says at this point.

I hope you do, however. Certainly, this conundrum makes anything neocla.s.sical economists have to say about the distribution of income irrelevant.

So what?

Few issues provide better examples of the negative impact of economic theory on society than the distribution of income. Economists are forever opposing 'market interventions' which might raise the wages of the poor, while defending astronomical salary levels for top executives on the basis that if the market is willing to pay them that much, they must be worth it. In fact, the inequality which is so much a characteristic of modern society reflects power rather than justice. This is one of the many instances where unsound economic theory makes economists the champions of policies which, if anything, undermine the economic foundations of modern society.

Economics should accept that labor is unlike any other commodity, and develop an a.n.a.lysis suited to its peculiarities, rather than attempt to warp this most personal of markets to fit the conventional cloth of supply and demand.

Keynes did just that in the General Theory. But mainstream economics after Keynes pulled away from this innovation on the basis that Keynes's argument 'did not have good microeconomic foundations.' As this and the preceding three chapters have shown, conventional microeconomic theory itself has unsound foundations. And things get even worse when we turn our attention to problems with the other 'factor of production,' capital.

PART 2 | COMPLEXITIES.

ISSUES OMITTED FROM STANDARD COURSES THAT SHOULD BE PART OF AN EDUCATION IN ECONOMICS.

7 | THE HOLY WAR OVER CAPITAL.

Why the productivity of capital doesn't determine profits.

The economist Dharma k.u.mar is said to have once remarked that 'Time is a device to stop everything from happening at once, and s.p.a.ce is a device to stop everything from happening in Cambridge.'

Nevertheless, a lot did happen at Cambridge during the 1960s and 1970s, where 'Cambridge' refers to both Cambridge, Ma.s.sachusetts, USA, and Cambridge, England. The former is home to the Ma.s.sachusetts Inst.i.tute of Technology (better known by its initials MIT); the latter is the home of the famous University of Cambridge. MIT was the bastion for the leading true believers in economics, while the University of Cambridge housed an important group of heretics.

For twenty years, these two Cambridges waged a theoretical 'Holy War' over the foundations of neocla.s.sical economics. The first shot was fired by the heretics, and after initial surprise the true believers responded strongly and confidently. Yet after several exchanges, the leading bishop of the true believers had conceded that the heretics were substantially correct. Summing up the conflict in 1966, Paul Samuelson observed that the heretics 'merit our grat.i.tude' for pointing out that the simple homilies of economic theory are not in general true. He concluded that 'If all this causes headaches for those nostalgic for the old time parables of neocla.s.sical writing, we must remind ourselves that scholars are not born to live an easy existence. We must respect, and appraise, the facts of life' (Samuelson 1966: 583).

One might hope that such a definitive capitulation by as significant an economist as Paul Samuelson would have signaled a major change in the evolution of economics. Unfortunately, this was not to be. While many of the bishops have conceded that economics needs drastic revision, its priests preach on in a new millennium, largely unaware that they lost the holy war thirty years earlier.

The kernel.

The term 'capital' has two quite different meanings in economics: a sum of money, and a collection of machinery. Economists a.s.sume that they can use the two terms interchangeably, and use the money value of machines as a proxy for the amount of machinery used in production. They prefer to abstract from the complexity that there are many different types of machines, many of which (such as, for example, blast furnaces) are solely suited to producing one particular commodity, and instead work with the generic term 'capital' as if there is some ubiquitous productive substance which is just as suited to turning out sheep as it is to producing steel. For the economic theories of production and distribution to work, the behavior of this hypothetical generic substance must be little different from the behavior of the actual real world of many different machines.

However, a careful a.n.a.lysis of production as a system by which commodities are produced by combining other commodities and labor shows that the money value of machinery cannot be used as a proxy for the amount of machinery used in production. As a result, the economic theory of how commodities are produced is wrong, and the theory's argument that profit is a reward for capital's contribution to production is also wrong. This reinforces the observations made in Chapter 6, that the distribution of income is not the result of impersonal market forces, but instead reflects the relative power of different social cla.s.ses.

The roadmap.

This quite difficult chapter begins with an outline of the economic theory of the production of commodities by 'factors of production,' with its a.s.sumption that all machinery can be lumped into the aggregate called 'capital' and measured by the money value placed upon those machines. Then Sraffa's 'abstraction-free' a.n.a.lysis of production is outlined. It is shown that, rather than the rate of profit depending upon the amount of capital, as neocla.s.sical economists argue, the measured amount of capital in fact depends upon the rate of profit.

Measuring capital.

Though the war began in earnest only in 1960, the possibility of conflict was first flagged by Piero Sraffa in his 1926 paper 'The law of returns under compet.i.tive conditions' (discussed in Chapter 5). In pa.s.sing, Sraffa observed that an essential aspect of the economic theory of production was the a.s.sumption that the interdependence of industries could be ignored. The problem was that this a.s.sumption was invalid when changes in one industry's output affected the costs of many other industries, which in turn determined the costs facing the first industry. As Sraffa put it, the a.s.sumption becomes illegitimate, when a variation in the quant.i.ty produced by the industry under consideration sets up a force which acts directly, not merely upon its own costs, but also upon the costs of other industries; in such a case the conditions of the 'particular equilibrium' which it was intended to isolate are upset, and it is no longer possible, without contradiction, to neglect collateral effects. (Sraffa 1926) Sraffa spent the better part of the next thirty-five years turning this observation into a rigorous theoretical argument. The product was a book with the bland but descriptive t.i.tle of The Production of Commodities by Means of Commodities (Sraffa 1960), and the rather more revealing but still oracular subt.i.tle of 'Prelude to a critique of economic theory.' Essentially, Sraffa provided the techniques needed to highlight fundamental internal inconsistencies in the economic theory of production.

7.1 The standard economic 'circular flow' diagram.

This theory argues that commodities everything from cornflakes to steel mills are produced by 'factors of production.' These are normally reduced to just labor on the one hand, and capital on the other. This concept is normally embodied in a 'circular flow diagram' like that of Figure 7.1, which shows factors of production 'flowing' from households to the factory sector, and goods flowing from the factory sector to households.

For this flow to be truly circular, households must transform goods into factors of production, while factories must transform factors of production into goods. The factories-to-households half of the circle is reasonable: factories can transform capital and labor inputs into goods. To complete the circle, households must transform the goods they receive from factories into factors of production labor and capital.

The proposition that households convert goods into labor is unproblematic. However, the questionable proposition is that households also convert goods into capital. This raises a vital question: what is capital, in the context of this diagram? Is it machinery, etc., or is it financial instruments? If it is the former, then this raises the question of where these machines are produced. The model implies that households take goods produced by firms and internally convert them into machines, which are then sold to firms by households. Clearly this is nonsense, since in this case 'households' must also be factories. Therefore, the flow of capital from households to firms must be a financial flow.

However, economic theory treats this financial flow as directly contributing to production: the 'capital' from households to firms generates a profit flow back from firms to households, where that profit reflects the marginal productivity of capital.

One way this would be possible is if financial instruments directly produced output (in combination with labor) which clearly they don't.

There is only one other solution, which is to acknowledge that the model is not complete. Factories actually produce capital machines, and this is left out of the diagram. The flow of capital from households to firms is therefore a financial flow, but hopefully there is a direct and unequivocal relationship between the measurement of capital in financial terms and its physical productivity.

7.2 The rate of profit equals the marginal product of capital.

A standard 'education' in economics simply ignores these complexities, and explains profit just as it explains wages: the payment to capital represents its marginal productivity. The argument goes that a profit-maximizing firm will hire capital up to the point at which its marginal contribution to output just equals the cost of hiring it. The cost of hiring it is the rate of interest, while its marginal contribution is the rate of profit. The two are equal in equilibrium, so the demand curve for capital slopes downwards just like all other demand curves reflecting rising demand for capital as the cost of capital falls.

7.3 Supply and demand determine the rate of profit.

The sum of all the individual demand for capital curves gives the market demand curve for capital, while the supply curve the willingness of households to supply capital rises as the rate of interest increases. The point of intersection of this downward-sloping demand curve with the upward-sloping supply curve yields the equilibrium rate of profit.

This argument should already be looking somewhat suspect to you, after the previous chapters. For instance, production is supposed to occur in the short run, when at least one factor of production can't be varied. That notion appears at least arguably OK when capital is the fixed factor though we've shown it to be invalid even there. But it makes no apparent sense to imagine that machinery is now variable while labor is fixed. Surely machinery should be the least flexible factor of production so that if it can be varied, then everything else can be varied too?

The arguments put by Sraffa against the concept of diminishing marginal productivity can also be applied here in a simple and devastating critique, which was first put formally by Bhaduri in 1969. As with the labor market, the 'capital market' is a broadly defined 'industry': there would be thousands of products being lumped together into the general rubric of 'capital,' and there is no industry which does not use some 'capital' as an input. This raises Sraffa's argument in Chapter 5, that a change in the price of such an input would affect numerous industries, and therefore alter the distribution of income. This is a similar point to that made earlier for the labor market, but it can now be put in a more explicit form.1 If we notionally divide all people into either workers or capitalists, then total income will be the sum of wages and profits. Profits in turn are the product of the rate of profit, times the amount of capital hired. Applying this at the level of the single firm, this gives us the relationship that: Income equals.

(a) the wage rate multiplied by the number of employees plus.

(b) the rate of profit multiplied by the stock of capital.

If we now consider changes in output (which we have to do to derive the marginal product of capital), then a rule of mathematics tells us that the changes in output have to equal the changes in wages and profits. Another rule of mathematics lets us decompose the change in profits into two bits: the rate of profit times the change in capital, and capital times the change in the rate of profit.2 This yields the relationship that: Change in income equals.

a) change in the wages bill (which we leave aggregated), plus.

b) change in profit (which we disaggregate).

Disaggregating changes in profit leads to the statement that: Change in income equals.

a) change in the wages bill, plus.

b) the rate of profit multiplied by the change in capital, plus c) the amount of capital multiplied by the change in the rate of profit At the level of the individual firm, economists a.s.sume that (a) and (c) are zero: a change in the firm's level of output caused solely by hiring more capital has no impact on either the real wage or the rate of profit. Thus the relationship can be reduced to: Change in income equals.

a) change in wages [zero], plus.

b) the rate of profit multiplied by the change in capital [one3], plus c) capital multiplied by the change in the rate of profit [zero]

Canceling out the terms we know are zero or one yields the desired relationship: Change in output due to a change in capital (marginal product) equals the rate of profit However, while this is a reasonable approximation at the level of the individual firm, it is not true at the level of the overall economy. There, any change in capital will definitely have implications for the wage rate, and for the rate of profit. Therefore the aggregate relationship is Change in output due to a change in capital (marginal product) equals.

a) change in wages due to change in capital [non-zero], plus.

b) the rate of profit, plus.

c) the amount of capital multiplied by the change in the rate of profit due to the change in capital [non-zero]

The rate of profit will therefore not equal the marginal product of capital unless (a) and (c) exactly cancel each other out.4 Thus at the aggregate level, the desired relationship the rate of profit equals the marginal product of capital will not hold true. This proves Sraffa's a.s.sertion that, when a broadly defined industry is considered, changes in its conditions of supply and demand will affect the distribution of income.

A change in the capital input will change output, but it also changes the wage, and the rate of profit. These changes alter the distribution of income between workers and capitalists, and will therefore alter the pattern of demand. Exactly the same argument applies to wages, so that in general a person's income will not be equal to their marginal contribution to output. As a result, the distribution of income is neither meritocratic nor determined by the market. The distribution of income is to some significant degree determined independently of marginal productivity and the impartial blades of supply and demand.

This adds what mathematicians call an additional 'degree of freedom' to the model of the economy. To be able to work out prices, it is first necessary to know the distribution of income; and there will be a different pattern of prices for every different division of the economic cake between workers and capitalists. There is therefore nothing sacrosanct about the prices that apply in the economy, and equally nothing sacrosanct about the distribution of income. It reflects the relative power of different groups in society though it is also constrained by limits set by the productive system, as we will soon discuss.

This contradicts economic theory, which says that the distribution of income is uniquely determined by the market (via the mechanisms discussed in these two chapters), and therefore there's nothing that policy-makers can or should do to alter it.5 Instead, rather than prices determining the distribution of income as economists allege, the distribution of income determines prices. Within limits, the distribution of income is something which is determined, not by market mechanisms, but by relative political power.

Bhaduri's critique still accepts the a.s.sumption that it is possible to define a factor of production called capital. However, as I intimated above, the machinery aspect of the term 'capital' covers too great a mult.i.tude of things to be easily reduced to one h.o.m.ogeneous substance. It includes machines and the buildings that house them; trucks, ships and planes; oil wells, steel works and power stations. Each of these items itself consists of numerous other sub-a.s.semblies which are themselves commodities. A truck contains an engine, which contains valves, springs and cables, the manufacture of which requires inputs from other types of capital, and so on.

The only thing that such disparate commodities obviously have in common is a price, and this is how economists would prefer to aggregate capital. But the price of a piece of capital should depend on the rate of profit, and the rate of profit will vary as prices change: there is an impossible circularity in this method of aggregation.

This problem was explicitly considered by Sraffa in his 1960 magnum opus. His purpose was to provide a firm foundation upon which a critique of the economic theory of production and income distribution could be built. He built his argument up stage by stage, with great care taken at each stage to make sure that the a.n.a.lysis was sound.

This meticulous method uncovered a number of paradoxes that invalidated the simplistic beliefs economists held about the relationship between productivity and income. Just as the peculiar conditions of 'production' of labor complicate the argument that the wage equals the marginal product of labor, so do the more conventional conditions of the production of capital disturb the argument that profit represents the marginal productivity of capital.