Monday, October 09, 2006

I'd Like To Hear Second Thoughts On The CCC From Joseph Stiglitz

Stiglitz's first thoughts can be seen in his 1974 review of Harcourt's 1972 survey book. About that time, Stiglitz also reviewed Pasinetti's presentation of the Kahn-Kaldor-Pasinetti-Robinson theory of growth and distribution. Neither Harcourt nor Pasinetti were happy with Stiglitz's respective review. Morishima (1977) is an outgrowth the JEP's editor becoming aware of Pasinetti's unhappiness.

I don't want to hear second thoughts from Stiglitz so much on the technical content below. I rather hear second thoughts on a point from the sociology of "knowledge". What determines how some views become worth considering in mainstream economics, while others are written out? I think Stiglitz might have something interesting to say on that based on his recent experiences, and I wonder if that could inform how he might look back at the Cambridge Capital Controversy.

Stiglitz writes:
"It is also true that with profit-maximizing competitive firms, in long-run equilibrium where all relative prices are constant, the rate of interest is equal to the own rate of return of every capital good (the marginal productivity of every capital good in terms of itself..." -- J. E. Stiglitz (1974)
I find Stiglitz less than forthright in failing to address the conception of capital as finance. Is Stiglitz clear that, in neoclassical theory, no theorem asserts the equality in equilibrium of the interest rate and the marginal product of (finance) capital? I think Stiglitz's formulation about own rates, as if that was a point in dispute, is likely to mislead readers.
"...At any moment, there is a given vector of capital goods and of labor. Under the extremely simplified models conventionally used, these endowments determine the marginal productivity of the different capital goods and the rate of interest. Given the savings behavior, this determines the change in the stocks of capital goods; eventually the economy converges to a state where the rate of interest is equal to the rate of growth divided by the savings propensity; still, at each moment, it is the 'capital goods-labor ratios' which determine the rates of return on the different capital goods." -- J. E. Stiglitz (1974)
(Notice Stiglitz does not say marginal productivities determine factor prices.) I think this emphasis on very short run equilibrium paths has not held up well; I see no reason to think a capitalist economy will follow such a path. Anyways, a consensus still does not exist on whether this approach is resistant to Cambridge capital critiques.

Consider:
"All that [reswitching] implies is that the weak qualitative assumptions we conventionally make in economics - that is, convexity of the technology, with its implications of diminishing returns - do not have any strong implications for comparisons of economies in steady state: that is, reswitching establishes that the derivation of simple comparative dynamics propositions requires stronger assumptions than those required for the existence of competitive equilibrium and the derivation of qualitative properties concerning economies with given initial endowments...

...Moreover, it is easy to develop, using steady-state analysis, all manner of paradoxes. For instance, the opening of free trade may actually lower steady-state consumption (this does not contradict the classical propositions concerning gains to trade). One can show that of all the feasible steady states in a life-cycle model, the one which maximizes steady-state utility is not sustainable by a competitive equilibrium (without appropriate lump-sum transfers), and conversely. (This does not contradict classical welfare propositions.)" -- J. E. Stiglitz (1974)
(I think the classical theory of free trade is more about the comparison of steady states than very short run paths with given initial endowments of capital goods.) I don't think this comment has aged well, either. Economists have not been able to state general assumptions that rule out capital-reversing - which is not the same as reswitching. Furthermore, if reswitching is so non-threatening, why hasn't it entered the mainstream introductory textbooks? Why are economics students at all levels not taught to be clear on the structures of their models and their implications? Surely, one should not still be able to surprise economists with paradoxes whose possibility was established decades ago?

References
  • Harcourt, G. C. (1972). Some Cambridge Controversies in the Theory of Capital, Cambridge University Press
  • Pasinetti, Luigi L. (1974). Growth and Income Distribution: Essays in Economic Theory, Cambridge University Press
  • Stiglitz, Joseph E. (1974). "The Cambridge-Cambridge Controversy in the Theory of Capital: A View from New Haven: A Review Article", Journal of Political Economy, (Cowles Foundation Paper 410), V. 4
  • Stiglitz, Joseph E. (1975). "Growth and Income Distribution: Essays in Economic Theory", Journal of Economic Perspectives, V. 13, N. 4 (Dec.): 1327-1328
  • Morishima, Michio (1977). "Pasinetti's Growth and Income Distribution Revisited", Journal of Economic Perspectives, V. 15, N 1 (Mar.): 56-61

3 comments:

radek said...

I rather hear second thoughts on a point from the sociology of "knowledge". What determines how some views become worth considering in mainstream economics, while others are written out?

Yeah that is an interesting question and one which Stiglitz probably has an interesting thing or two to say about.

But Stiglitz is also a good illustration of the fact that "mainstream economics" isn't as homogenous as some folks think. Other "mainstream" economists might disparage his popular books (with some good reasons me thinks) but I haven't met one who didn't have anything but profound respect for Stiglitz actual research. Politics or no politics. John Roemer is another example, though less known outside his area.

As far as why the CCC has never made it into "mainstream" thought... I got my own ideas on that and maybe I'll start a blog where this topic is the launching point. Succinctly I think that IT DID make it into mainstream thought. Aggregate production functions disappeared for 15, maybe 20 years. You couldn't find an economist in the 70's who came near a APF unless it was to slap it around some more (like Franklin Fisher) (maybe I'm wrong on this, it's not like I did any rigorous research). Then in early 80's you got New Growth Theory and RBC which basically brought the APF back. But at that point, no one cared about theory-theory, much less arcane debates in methadology of theory-theory, as this paper shows:

http://www.nber.org/papers/w12526

Instead economists were happily playing with their more powerful computers and new fangled estimation techniques. Most PhD programs these days churn out empirical economists who cross themselves when they hear the words "Existance Proof" and reach for garlic when History of Economic Thought comes up.
Not a really good development, in my opinion, but I doubt ideology or conspiracy has much to do with it.
Ok, that wasn't succint. I could talk about how Deirdre McCloskey's ideas on economics as metaphor fit into it but then I'll have nothing to write about when I do start my own blog. In the mean time real life teaching and reasarch duties interfere.

h economicus said...

I like Rubinstein's characterization of economics as being a set of conventions, and I would agree that there is considerable levity in terms of what ideas can be explored, *provided* that they can be expressed in terms of these conventions. If I was to try to characterize these conventions, I would include (i) methodological individualism, (ii) axiomatic reasoning, and (iii) equilibrium concepts. Presumably there is an economic reality out there, and I feel a certain amount of scepticism about whether the above elements are sufficient tools to understand these phenomena, whatever they may be.

You (Radek) mentioned Rubinstein's work on bounded rationality. I find the exchange between him and Simon Herbert interesting. Simon argues that what Rubinstein is attempting is pointless, since psychologists have already developed models decision making under bounded rationality, and that what is required is not the development of further models, but rather the acquisition of further knowledge on actual regularities in decision making (i.e. additional observation of the actual economic world). I think one of the reasons that "bastard" theory endures is that data acquisition in economics is so difficult, and that existing data is so poor. As a result, the only inexpensive avenue for research is in the realm of theorizing.

With regards to the CCC, I think part of the problem is that the economics discipline has embraced instrumentalism without being willing to fully acknowledge it. The CCC vanished because the neoclassical theory of distribution is both simple and appealing, provided that one doesn't think about the arguments of the production function too much. A full acceptance of instrumentalism would be problematic, as I see it, because there would be no place for normative economics. The standard conceit of understanding the intuition of a model would also be largely pointless, since these imply an understanding of underlying process. I suppose this "intuition" would retain a heuristic role.

Robert Vienneau said...

Thanks for the comments.

I like that while Radek is more willing than I to defend the content of mainstream economics, he seems to be critical of the culture of economists too. I find particularly stupid an attitude that if literature more than a decade old had anything to teach us, its lessons would already be incorporated in recent literature. But I have found that some mainstream economists believe that at some level.

I agree that the CCC did temporarily affect the practice of mainstream economists. But they never wrote the techniques and arguments into the textbooks. So the most recent generation doesn't seem to be aware of the arguments.

Much of the long arguments I present seem to me to be about the logical implications of optimization, applied in a manner at least as consistent with methodological individualism as mainstream practice. I think using theory to root out logical errors is a sound use of theory, whatever one may think about shifts in the balance among economists in the role of theory and econometrics.

I agree with H. that some resist these arguments because of beliefs in instrumentalism. This has never made much sense to me. How does this methodology defend illogic?