I wrote:
>one big problem with the Sraffa tradition (that of reswitching, not >the early stuff) is that it is highly dependent on the assumption of >equilibrium. That is, the equilibrium points are knife-edges, >unstable, which makes them inappropriate for most empirical work. This >is unlike the neoclassical tradition, in which the assumption of >smooth curves allows the attainment of a new equilibrium after an old >one has been disturbed. However, it is appropriate to use Sraffian >equilibrium models to criticize neoclassical (or other) equilibrium >models. It is the latter where Sraffa has been most successful.
Paul:
This is an intriguing post with some original thinking. I am not sure I have fully grasped what Jim meant (I am no Sraffa-mathematical expert). As I am sure Jim knows, just before her death Joan Robinson turned against Sraffian ideas (which she had long promoted). Along the lines of the "pure" Keynesian view, she especially criticized the use of equilibrium: capitalism is inherently in disequilibrium and the Sraffians' equilibrium didn't reconstruct real historical time, just an artificial logic-based version (without path-dependence, etc). Am I right that Jim's critique is not so methodological as this (i.e. the use of equilibrium per se)?
I totally agree with Robinson on this point, though the concept of economic equilibrium does play a role: I see equilibrium as only notional state. There are both forces that push the economy toward equilibrium and those that push it away from it. (In my dissertation, I present a growth model in which there is a stable short-term steady-state equilibrium, but in the medium term, it becomes unstable, a Harrodian knife-edge. In general, the macroeconomy oscillates in irregular ways rather than following any kind of steady state. Kaldor's business cycle model is an exemplar of this kind of thinking.) I really don't think Robinson changed her mind on these issues. (The big change, I guess, happened between her 1933 stuff on monopolistic competition and her Keynesian work.) Rather, she simply made her objections to the equilibrium-bound thinking of the Sraffians more explicit. Before that, for example in her ACCUMULATION OF CAPITAL, she was always clear that she was discussing parables and unreal cases. (In a 1941 article on micro theory, "Rising Supply Price" (in ECONOMICA), she started by saying "It may seem strange at this time of day to reopen the old familiar subject of diminishing returns and rising supply price. My purpose is frankly escapist, and what follows has no relevance to any problem of importance in the real world.")
I am hearing something like this (or have I got it wrong?): In looking at the choice of capital technique the Sraffians think of a capitalist who pages through a book of technical choices and picks the one that will maximize profits (which depends on the current prevailing wage). This is Jim's sharp "knife-edge" type choices. Neo-classicals think of capital techniques as "putty", the choices are continuously created as part of the process so the equilibrium changes appear to be more smooth. But I haven't followed how this makes the Sraffian view less appropriate for empirical work? Is it a practical problem (difficulty in modeling) or is it a theoretical flaw in the Sraffian view (such as Joan Robinson on equilibrium)?
The problem of Sraffian models isn't really the empirical part. Following the neoclassical tradition, you simply have to _assume_ that the model works and then stick in the numbers (from input-output studies). Ed Ochoa, among others, have done these studies. The problem is that on the theoretical level, the Sraffa model doesn't deal with transitions between equilibrium states well at all. If you can't make it from equilibrium A to equilibrium B without spinning into depression or hyperinflation, you don't have a very good theory. Sraffa's work is really a _critique_ of neoclassical work, as Michael Perelman points out (and as Robinson recognized). It's not really a replacement for the neoclassical model. That's one reason why the latter model persists in being used. (It's usually not enough to smash a theory; you need to provide an alternative.) The neoclassical model is absurd, but if you're willing to make the assumption that it's empirically valid, you can crank out results (as with the "new" growth theory). BTW, I think that one of the basic lessons of the Cambridge Controversy is that the only variables that can be truly aggregated are _monetary_ (nominal) ones. All of the other variables that macroeconomists use are simply first-approximation "as if" numbers. Using "real" variables is extremely dependent on the validity of the deflator used (and the deflator itself can be, and has been, a political football). (Maybe we should go back to Keynes' use of the average money wage as the deflator.) BTW2, as Samuelson noted while conceding the Sraffians' point, it's possible to produce "reswitching" results (one of the big conclusions of the Sraffa model) without using the Sraffa model. It can follow from simple net present value calculations where the stream of expected future benefits is irregular. Or alternatively, see a little article I wrote for the Wikipedia: http://en.wikipedia.org/wiki/Capital_controversy#Reswitching. The whole article still seems pretty good, not having been destroyed by interlopers.
Thanks to Jim for the post.
you're welcome! -- Jim Devine / New Quote Needed! Please submit one!
