To GE or not GE

That is the question. While looking on the web for material for a paper I'm working on, "The Past and Present of the Theory of the Firm", I came across a number of discussions on the evils or the non-evils of general equilibrium theory. Some people argue McKenzie, Arrow and Debreu et al are the great satans who are out to destroy economics - if not the world. Others believe GE theory is the greatest achievement in the history of economics. This interested me because I end my paper with the following observation (a couple of footnotes have been removed)

A final point about the models of the firm discussed in this essay is that they highlight a general issue to do with post-1970 microeconomics, that is, the movement away from the use of general equilibrium (GE) models. All the models considered above are partial equilibrium models, but in this regard the theory of the firm is no different from most of the microeconomic theory developed since the 1970s. Microeconomics such as incentive theory, incomplete contract theory, game theory, industrial organisation etc, has largely turned its back, presumably temporarily, on GE theory and has worked almost exclusively within a partial equilibrium framework. This illustrates the point made at the beginning of the paper that there is a close relationship between the economic mainstream and the theory of the firm; when the mainstream forgoes general equilibrium, so does the theory of the firm.

For the case of contract theory, a field closely related to that of the theory of the firm, Salanie (2005: 2) has argued,

``[t]he theory of contracts has evolved from the failures of general equilibrium theory. In the 1970s several economists settled on a new way to study economic relationships. The idea was to turn away temporarily from general equilibrium models, whose description of the economy is consistent but not realistic enough, and to focus on necessarily partial models that take into account the full complexity of strategic interactions between privately informed agents in well-defined institutional settings".

As was noted in Section 4.1 the post-1970s literature on the firm can be classified into two general groups based on which of two of the standard assumptions of GE theory, namely symmetric information and complete contracts, is violated when modelling the firm. This requirement to violate basic assumptions of GE theory so that we can model the firm, suggests that as it stands GE can not deal easily with firms, or other important economic institutions. As Bernard Salanie notes,

``[ ... ] the organization of the many institutions that govern economic relationships is entirely absent from these [GE] models. This is particularly striking in the case of firms, which are modeled as a production set. This makes the very existence of firms difficult to justify in the context of general equilibrium models, since all interactions are expected to take place through the price system in these models". (Salanie 2005: 1).

This would suggest that creating GE models that can account for information asymmetries, contractual incompleteness, strategic interaction and the existence of institutions is needed if GE theory is to be able to deal with the issues - including those related to the firm - that microeconomists wish to tackle.

Now what I'm wondering is, Is the problem with GE a problem with the idea of general equilibrium as such or a problem with the Arrow-Debreu approach to GE? I can't help thinking that the world is GE and thus to handle the "real world" we need general equilibrium models. But the way we model GE leaves much to be desired. The current generation of GE models don't deal well with things like information asymmetries, contractual incompleteness, strategic interaction and economic institutions. Thus do we need at new GE which doesn't just try to model the extreme decentralisation of the perfectly competitive framework? Not that it is obvious what such a form of GE would look like. How much, if any, of the current approach to GE can be saved? Or is the Arrow-Debreu framework the right way to go? Or should we just stay with partial equilibrium models and not worry about interactions between markets?

Things you ponder on a Thursday afternoon.