Looking for something else altogether, I found this:
What Denrell sets up is a fairly simple model, in which a capitalist hires a worker to achieve a particular project. There is one important variable, which he calls the amount of discretion the worker has over how to do his part of the job; one could just as well think of it as the level of skill. He assumes that the income of the firm is an increasing, concave function of the level of discretion. He also assumes that if the worker quits once a project has begun, the value of the project takes a very big hit, and may indeed become negative; this cost is larger the more discretion the worker had. There is some value-maximizing level of discretion for the worker. But now suppose that they agree on a wage and the work begins. Under certain conditions, which Denrell specifies, the worker could credibly threaten to walk off unless he gets paid more ("hold-up", in the jargon). Because that would be so costly, he'd get it. Now, being economists, we must assume that if the worker can do this, he will. The capitalist knows this, so he would prefer a lower level of discretion, one which reduces the total revenue, but does not leave him vulnerable to hold-up. The capitalist gets more at the inefficient level than he does, after the hold-up, at the efficient level, and so we see a level of discretion (skill, etc.) which maximizes his profit, not production. (Cue "The mode of production becomes a fetter on the further development of the forces of production", etc., to the expropriation of the expropriators.)
Incomplete contracting is important here. If they could find some way of working at the higher level of discretion, both the capitalist and the worker would be better off, but there are many circumstances where it would be very hard to devise, let alone enforce, a contract that effectively kept either side from pulling a hold-up. Denrell, being a Scandinavian, notes that unions play an important function here, by making it less costly --- for both sides! --- to enforce the labor contract. Reputations in repeated games can play the same role.
Note that if there's some statistical risk of the worker leaving the job, even if there's no hold-up, a similar logic applies --- capitalists will prefer lower levels of discretion, because that reduces their exposure to that risk, even if it is technically inefficient. Denrell suggests some ways in which one might hope to discriminate between these two possibilities empirically. He also notes that the model implies no progressive tendency for capitalists to prefer to lower the level of discretion over time, which accords with the empirical evidence that there is no so such tendency overall. It's also important that the capitalist can lower the level of discretion with only a fairly small reduction in output; if it fell of precipitously, then he'd just have to take his lumps.
Overall, then, we have a very nice model which cleanly captures many of the intuitions going into one of the more plausible ideas of radical political economy, and shows that they do indeed deliver the promised result, under certain specifiable conditions. (It is also notable that it works within methodological individualism; classes as collective actors, etc., aren't needed.) This is a good starting point for understanding when and how that prediction applies in the real world, refining our understanding of the situation, and maybe even doing something about it.
I will mention two limitations, both relating to the capitalist. First, Denrell does not remark at all on his assumption that the capitalist gets to set the level of discretion, to make it whatever he wants. This is where power enters the picture, as Bertrand Russell ("power is the production of intended effects by some men on other men"), Spinoza, etc., could've told him. Consider the difference between Denrell's situation and a model of the relationship between firms in a supply chain. Firm A contracts firm B to make some specialized component for a project A is undertaking, and this can be set up with differing levels of discretion on B's part, etc., just as before. This leaves A vulnerable to hold-up, so it might very well want to force B to operate on a tighter leash. But it can't just dictate terms to another firm in that way --- unless it possesses enough economic power (or buys B outright; this is a plausible argument for how we might get inefficient vertical integration). Second, his capitalist still wants to maximize his income. He doesn't actually mind changes that let the worker get more, provided he doesn't have to pay for them, and in principle wouldn't care if the worker made more than he did. Denrell does not, therefore, consider the case of someone who prefers to get a smaller absolute income, if that is the price they must pay to make sure their relative income increases. But then, neo-classical economics has always been hampered by its sunny view of human nature.
Posted by crshalizi at October 11, 2004 23:44 | permanent link