Wednesday, February 12, 2014

Steve Keen, Debunking Economics, Chapter 6: Wages

I review Chapter 6 of Steve Keen’s Debunking Economics below, which is a discussion of wages and labour markets.

Neoclassical economics analyses labour as a commodity, like any other, governed by the law of supply and demand (Keen 2011: 129).

Two crucial requirements of standard neoclassical analysis of labour markets are that (1) labour demand curves are necessarily downward-sloping and supply curves upwards-sloping, and (2) each worker tends to be paid the marginal product of labour (Keen 2011: 130–131).

Yet labour is fundamentally different from other commodities: whereas demand for some commodity like bread is determined by consumers and supply decisions by producers, the supply of labour is offered by consumers, and demand decisions are made by producers (Keen 2011: 129).

Steve Keen sees a number of problems with the neoclassical analysis:
(1) the labour supply curve can “slope backwards”: e.g., a fall in the wage rate can induce an increase in the supply of labour;

(2) the market power of some employers can result in unfair wages even in neoclassical theory, so that worker trade unions or collective bargaining can make wages fairer;

(3) standard supply and demand analysis can be inappropriate when applied to labour markets in light of Piero Sraffa’s aggregation problem;

(4) the fundamental explanation of labour supply as workers choosing between leisure and work is flawed;

(5) that market demand curves, including labour demand curves, necessarily obey the law of demand is unrealistic.
In regard to (1), Keen notes how a higher wage rate can result in the same income level for a worker if he or she works fewer hours: therefore less labour might be supplied as the wage rises (Keen 2011: 133–134).

Attempts to overcome this problem with the substitution effect are not convincing:
“… it makes no sense to separate the impact of an increase in the wage rate into its substitution effect and income effect: the fact that the substitution effect will always result in an increase in hours worked is irrelevant, since everyone will always have twenty-four hours to allocate between work and leisure.

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.

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 assumption that, in general, labor supply curves will be upward sloping. But there is no theoretical – or empirical – justification for this assumption” (Keen 2011: 134).
The problem Keen identifies here is that labour supply curves need not be well behaved.

This is just as easy to see in reductions in wages. A strong general characteristic of most households is that they wish to maintain their standard of living, as they face fixed contractual obligations like debt, and hence the need to maintain income levels (Lavoie 1992: 222).

Therefore labour supply often depends on a perceived target wage rate and past standards of living (Lavoie 1992: 222–223), not necessarily on actual movements of the wage rate. If wages fall, this may well increase labour supply as a breadwinner or other members of the household decide to look for more work to maintain household income.

To turn to point (2) above, the real world is far from the perfect or near competition models of neoclassical theory.

Even if one wants to assume that workers should be paid their marginal product, firms with market power will pay wages below this value, so that a trade union acting as a single seller of labour will drive wages higher, so that wages will be fairer (the so-called monopsony argument).

In regard to point (4), neoclassical theory holds work and leisure to be two “goods,” between which workers freely choose as the wage rate changes. In a truly laissez faire society with no welfare or social security, this idea is of course nothing more than a sick joke: either you work for whatever wages so can obtain or starve.

Even in modern welfare states, the idea is still dubious: for most forms of leisure require money and income, and in most full-time work hours worked are strictly set by employers and not often negotiable.

Keen also notes how in recessions or depressions where there is a very high level of (normally) fixed private nominal debt, cutting wages and prices (and hence profits, which are the income of businesses) to increase demand for labour will induce debt deflationary pressures, a self-defeating exercise (Keen 2011: 138).

Keen, Steve. 2011. Debunking Economics: The Naked Emperor Dethroned? (rev. and expanded edn.). Zed Books, London and New York.

Lavoie, Marc. 1992. Foundations of Post-Keynesian Economic Analysis. Edward Elgar Publishing, Aldershot, UK.


  1. "If wages fall, this may well increase labour supply as a breadwinner or other members of the household decide to look for more work to maintain household income."

    I can buy this as a possible correlation, but might the causality be in the opposite direction?

    If so, it might strongly resemble the entry of women into the workplace and diminishing family incomes. Today, with both spouses working, family income is lower than it was in the past.

    1. Well, the increase in part time work and the entry of partners/wives of breadwinners into the labour force (as well as the surge in private debt) to increase household income over the past 30 years is generally seen as the consequence of real wages falling since the mid-1970s.

      This in turn was caused by the weakening of trade unions and the anti-labour policies of governments since the 1970s.

  2. I read two pieces by Auld. Here's one.

    Devastating. Consider just the Q stuff at the beginning. Keen introduces a contraint on Q. Constraining Q changes the problem. This has been pointed out to Keen repeatedly. This is MF level buffoonery from Keen.
    Rowe's piece is even better.

    1. I am not seeing any devastating demolition of Keen's work there or of Post Keynesian theory.

      Auld is arguing over some technical points about mathematical models:

      "I also think both Nick’s blog post and my previous piece make an error in possibly leaving the reader with the impression that Keen’s argument is correct if the competitive model is unrealistic or fails empirically, but that’s not the issue. "

      Yes, it is. All economic theory boils down to the questions: is it unrealistic? and does it fail empirically?

    2. Maybe all economic theory does, but not all arguments do. Getting the math wrong for instance. Keen's argument is destroyed by Auld because it is a failed mathematical derivation. For the rest of the book, I won't read it, as I won't read MF's magnum opus on economics.

    3. I disagree on point of fact here Ken.

      I have been trying to work out a key point of fact to make it clear, but I think Auld goes astray around here,
      "Let’s consider that claim. Yes, firm 1′s profits in equation (1) depend on firm 1′s own output and on the output of all other firms, R. No, that does not imply that we solve firm 1′s profit maximization problem by taking the derivative of equation (1) with respect to total output." much of what follows for the next two paragraphs is incomprehensible to me, and doesn't appear to match the Keen papers thinking.

      Now you need to understand the total derivative (Auld mentions several things its not, but doesn't discuss what a total derivative is). We are talking about the first definition from here, In this model there are strategic interactions between the firms, so in order to find out what the effects of one firm changing its output level one must also consider the effects of that change of output level on other firms output levels. The fact there is strategic interaction is established earlier on in the Keen paper. That is why its a total derivative, the output of firm q1 effects the profit maximization output level of firm q2 ... qn in this model. e.g because the firm q1 changes its output level the other firms optimal output levels change in response to this change..

      Auld discusses an industry where all firms attempt to maximize profits by equating mc=mr but doesn't establish that that its impossible to do better by following any different strategy (as far as I can see).