Sunday, April 15, 2007

Prescott on the Cambridge Capital Controversy

"I will argue that much of the failure of monetary economics to progress over the last 25 years is the failure to construct models that provide guidance on how the national accounts should be extended to include production of 'financial services'. In the 1960s there was the famous Cambridge capital controversy. This controversy bears on the issue 'What is money?' The Cambridge capital controversy was a silly one, as pointed out so clearly by Arrow (1989). Arrow, being a general equilibrium theorist, pointed out that there are multiple types of capital goods and with multiple capital goods only under very special conditions is there an aggregate capital stock. I emphasize that this does not mean that a model with a single capital good, which is matched to the value of some capital good statistic, is not useful in drawing scientific inference. I use such models along with other national accounts statistics to draw quantitative inference concerning a variety of phenomena, including business cycle fluctuations, secular movements in output and hours worked in the market sector, depressions and prosperities, and even in the behavior of stock prices. Rather, it means that for some purposes this single capital stock abstraction is not a good one for drawing inference." -- Edward Prescott (2005). "Comments on 'Inflation, Output, and Welfare' by Ricardo Lagos and Guillaume Rocheteau", International Economic Review, V. 46, N. 2 (May): 523-531

5 comments:

Gabriel Mihalache said...

Can we now all be friends, or what? :-)

Robert Vienneau said...

I don't see how a dismissal of price theory as "silly" would lead me to be friends with Prescott - not that intellectual agreement or disagreement is a basis of friendship. Furthermore, I agree with Paul Samuelson, Amartya Sen, Geoff Harcourt that the CCC wasn't concerned solely with how to aggregate capital. (Harcourt makes this point elsewhere in the referenced interview than where I quoted.)

I like slightly better Gabriel's sometimes statements that he may someday examine the CCC, than Prescott's nonsubstantive dismissal. My problem is that I don't think Gabriel states issues correctly. As I understand it, the textbook model he mentions is the Solow growth model augmented with a representative agent, where the representative agent engages in intertemporal utility-maximization. Contrary to Gabriel's claim, this model does not have 1 consumer good, 1 capital good, and 1 type of labor. It does have one type of labor. But it has one other commodity, which is both a capital good and a consumer good. Having one commodity (other than labor-services) doesn't prevent there being a price ratio among consumables at specific dates. I think one needs to be clear on this matter to understand why, due to price Wicksell effects, the equilibrium interest rate is, in general, not equal to the marginal product of capital. This inequality arises in models with a single capital good. (Doesn't Prescott make the same mathematical mistake in the quotation in my post?)

In the one-good model, the one commodity can be composed of multiple commodities in fixed proportions. I don't think of that as a genuine multiple-commodity generalization.

I haven't read the Prescott-Summers interchange, as opposed to comments on it. I suspect the Keynesian objection isn't to Pareto optimization, as such. Rather it is to discarding the possibility of macroeconomic discoordination, by assumption. Sometimes industrial economies have recessions, in which some are out-of-work and all markets do not clear. It isn't the case that the unemployed have all just chosen to take long vacations. (I realize the exceptions in the Lucas-tradition where ad-hoc constraints on prices or wages are introduced.) I think Solow has expressed his surprise on how far Lucas and his followers have gone with their crazy idea that recessions do not occur.

Gabriel Mihalache said...

What we say is that you can turn 1 unit of the consumption good into capital. It simplifies the investment dynamics.

Fix proportions make sense in the steady state. (Sidenote: hence the illusory empirical success of multipliers or rules-of-thumb). So I wouldn't rule out Hicks-Leontief goods aggregation.

In the end, I still think that the difference between us is regarding how we judge models and what we expect out of them.

I find insights, meaning and use in models where there are no Wicksell effects or capital-related stuff. You don't.

Everyone should be aware of the CCC and their ramifications. That being said, this is one of MANY open issues. Science is a set of open issues.

Utility functions misrepresent welfare at times and simple representations of capital obscure some of the complexity there. And we could go on for every issue.

So you try and move forward anyway you can. For me, RBC is going forward since it answers some questions. It's not a framework for studying distributive issues. I'll do a post on this later today, hopefully.

Robert Vienneau said...

If the capital good were a different good than the consumption good, the possibility of price Wicksell effects would arise.

abhishek narangwal said...

capital controversy was silly?? I do not agree with you. Capital is a capitalistically produced means of production, its not a factor of production. Labor is only the factor of production. the demand and supply approach of Neo-classical in every analysis seems to be silly. Their various conceptions on the meaning of capital itself shows how contradicting is their approach.