Thursday, June 08, 2006

Maximization is impossible

Well, OK. It is possible to solve mathematical optimization problems, and it is possible to find quite a few real-world problems that can be modelled, and solved, using linear programming- see this wonderful article by Virginia Postrel- but John Keay describes the contributions of Herbert Simon, and argues that academic economists are wrong to assume that businesses or individuals maximize anything.

There are some people who are searching for the highest point in England and it is probably good for society that there are. But it is probably not good for them. Some will struggle to the top of Scafell Pike, but more of them are tired and lost in the valleys, taking Prozac.

I think it should be obvious that Firms don’t maximize value- I guess economists merely mean to say that employees should try to act as an owner, also of bounded rationality, would act if in his place- the principal-agent problem.

He discusses the principal-agent problem here, and how the enthusiasm for markets led to a focus on customer satisfaction that ultimately eroded trust in the entire profession of auditing.

Since the audit process is one in which the property owners – the shareholders – pay for the police and the potential burglars – the corporate executives – appoint them, we should expect scrutiny to be costly but not rigorous and frequently this is what we find.

I am not too happy with the the proposed solution, however- seems to be an endless regress.

No comments: