I asked Gil what Veneziani's specific results were, but he didn't seem to have time to respond. Luckily, I found my copy of Roberto Veneziani's "Exploitation and Time," (London School of Economics: Centre for Philosophy of Natural and Social Science, discussion paper 47/00). [BTW, he doesn't just cite Dymski and myself; he also cites pen-l's Justin Schwartz.] His analysis of Roemer's theory of exploitation produces the following results. If you want to skip the quote, the start of my summary and commentary are marked by two asterisks.
"(1)" Veneziani first discusses the importance of time in terms of whether we should look at "within period" exploitation [static cross-section or synchronic] or "complete lives" [dynamic, over the life-cycle] exploitation. I agree with him that "the former emphasizes the importance of the structure of economic relations in a capitalist economy" which is "not only more in line with Marx's theory, but also it gives the opportunity to evaluate the dynamics of exploitation deriving from the structure of the competitive economy." [The assumption of a competitive economy doesn't hurt this analysis, since the existence of monopoly doesn't seem to contradict his results. BTW, I've corrected some typos below and reformatted a bit (adding some punctuation) for readability. It is not quite a literal quotation.] "(2) In particular, focusing on within period exploitation, it is possible to analyse another problem that emerges once time is introduced, i.e., the issue of the persistence of exploitation. If agents' lives are divided into several productive periods, Roemer's models are not able to provide microfoundations to [explain] exploitation as a persistent phenomenon. This fact has two theoretical implications: "firstly, Roemer's models are not suitable to discuss Marx's concepts of exploitation and class. According to Marx, exploitation and class are inherent phenomena of a capitalist economy and they relate to the very structure of capital/labor relations. Instead, the intertemporal generalisation of Roemer's model shows that even without a capital market, the competitive economy displays and inherent tendency toward labour value pricing, i.e., a just structure. [Veneziani assumes that price = value represents fairness or justice in exchange, which doesn't fit with Marx's views.{*}] [BTW, this fits with the standard Marxian result that if the rate of surplus-value is zero (here the result of the end of Roemerian exploitation), values and prices of production correspond.] "Thus, in Roemer's framework, exploitation is _a transitory accident_ [my emphasis] related to an initial unjustified differential ownership of property assets that has nothing to do with the functioning of a capitalist economy [as modeled by Roemer or Veneziani]. And [Veneziani's Proposition 3] suggests the very counterintuitive result (from a marxian perspective) that a laissez-faire policy could be justified in a competitive economy [because exploitation tends to evaporate and the inequality of wealth ownership to moderate: trickle-down works in Roemer's model when Veneziani adds time to the mix]. "Secondly, Roemer's models provide microfoundations to exploitation and class as persistent phenomena [only] thanks to a very restrictive set of assumptions, and in particular the impossibility of accumulation. In the one-period model, agents optimally choose their exploitation and class status given the constraints, but they cannot change it. Accumulation is an essential ingredient to analyse the last issue and when it is included, Roemer's models do not provide microfoundations to [the existence of] exploitation and classes. ... "(3) According to [Proposition 3], the tendency of exploitation to disappear and the persistence of DOPA [the differential ownership of productive assets] are coexisting features of the economy and this fact has two consequences. First of all, the persistence of inequalities in the ownership of productive assets is not a sufficient [indicator] of the fairness of the labour/capital relations (and more generally, of the society) from a marxist perspective. In the intertemporal context, DOPA remains probably necessary to have exploitation, but it is by no means sufficient for it to persist. In this sense, DOPA does not contain all the relevant moral information, as maintained ... by Roemer -- and his game-theoretic definition of exploitation based exclusively on DOPA cannot be considered a generalization of Marx's theory. "Secondly, given that the analysis of DOPA may be seen as reflecting a different moral concern from exploitation, if the results of the intertemporal model are taken into account, Roemer's reasoning about the relative importance of DOPA and exploitation can be reversed -- and it seems legitimate to ask ...: should anybody be interested in DOPA? [Proposition 3] shows that DOPA can be a persisting feature of an exploitation-free economy and that the competitive economy tends to this allocation even if there is a persistent wealth inequality. "Thus there seems to be no reason to care about inequalities in ownership of productive assets, since they have no effect on the fairness of the competitive economy. Roemer's emphasis on inequalities in the ownership of productive assets seems misplaced: DOPA is relevant for its consequence in terms of relative welfare or because it can contribute to the existence of exploitation, but it is not negative _in itself_" [his emphasis]. "(4) Similarly, since exploitation tends to disappear while class division is a persistent feature of the economy, Roemer's definition of classes in terms of the chosen vector of labour purchased and performed tends to lose its normative content in the intertemporal framework. Since the economy moves toward labour value pricing (i.e., fair pricing in [Veneziani's interpretation of] Marx's theory) the fact of being a proletarian rather than a capitalist -- according to Roemer's definitions -- becomes increasingly less important. However, the disappearance of exploitation together with the persistence of class division is a quite counterintuitive result that deserves further investigation and raises further doubts about the opportunity of discussing marxian concepts by means of neoclassical competitive models." (pp. 27-28). Veneziani then goes on to discuss some limitations of these conclusions, but my impression is that Gil didn't think of them as important. ** What all this says is one major point that Gary Dymski and I emphasized in our paper: that Roemer's story of exploitation self-destructs, as exploitation disappears and wealth differences moderate due to accumulation, including workers' accumulation. Unlike our paper, Veneziani shows this result rigorously within the orthodox mathematical framework that Roemer uses. (We used logic and our knowledge of economics instead.) If I understand Gil's point correctly, he's saying that Veneziani's analysis also undermines Marx's vol. I/ch. 25 analysis. I guess that for him, capitalists can invest all of the surplus-value they appropriate without producing the kinds of results that Dymski/Devine and Veneziani find for Roemer, i.e., the end of the scarcity of the means of production and thus the end of Roemerian exploitation as a social phenomenon. But throwing a little bit of workers' saving into the mix undermines not only Roemer's story (as Dymski/Devine argue) but also Marx's. I presented an argument about why this isn't so (see http://csf.colorado.edu/mail/pen-l/2002I/msg02491.html and http://csf.colorado.edu/mail/pen-l/2002I/msg02502.html) and won't repeat those arguments here. (By the way, my argument applies once capitalism has already been established, e.g., nowadays, after the period of primitive accumulation. It's possible that Gil's argument would apply when capitalism was in its infancy. Of course, the state applied all sorts of force to make sure that workers didn't accumulate. See Michael Perelman's THE INVENTION OF CAPITALISM.) A major reason why there's a disagreement here is because of the different nature of the theories that Roemer and Marx present. On the one hand, Roemer presents an abstract model in the Walrasian tradition, a total idealization of real world markets that totally abstracts from the most important aspects of capitalism, including the production process (it's a purely exchange economy). In the anti-rigorous rigor of Walrasianism, it's most common not even to acknowledge deviations from reality. The assumption that money doesn't exist in the model, for example, is almost never mentioned. The tradition is deductive in approach and idealist in philsophy, following the Platonic tradition of dealing only idealized forms. As is well known, and as seen in such theorems as "the general theory of the second best," it's possible to introduce one or maybe two "imperfection" (i.e., element of the real world) into the Walrasian tradition but that as the model becomes more and more realistic, it becomes less and less meaningful, producing fewer and fewer clear results. That is, introducing realistic markets, economies of large-scale production, production itself, historical time, a serious theory of money are all possible, as is getting rid of the silly Walrasian system of equilibration (a god-like Auctioneer who determines the prices for all products at the same time, before anyone's allowed to trade). But the more you do this type of thing, the less the model says. That is, Roemer's self-chosen theoretical framework is extremely brittle and breaks as it moves toward realism. One little snake can destroy the Roemer's idealized garden of Eden, while having more than one snake guarantees this result. On the other hand, Marx's theory is institutionalist and inductive, realist in philosophical orientation rather than idealist like that of Roemer. It's extremely complex in its presentation, as it starts out with a very simple and abstract analysis and slowly moves toward a more realistic picture of the capitalist world. The different parts of Marx's analysis are also dependent on each other, since he's trying to describe a complex totality. What this means is that one little part (such as KI/25's theory of accumulation) is dependent on the rest of the book, all three volumes. As Sweezy pointed out, Marx's definition of what he meant by capitalism took an entire volume. His understanding of accumulation takes even more (e.g., the last part of vol. III). This is tragic, of course, since Marx never finished volumes II and III or the book on wage-labor. In other words, when Gil says that Marx's theory of accumulation in KI/25 is shown to be wrong when you bring in workers' saving, in the Marxian framework that's saying that "Marx's theory in KI/25 is too abstract and we need to bring in more real-world elements," which was a good idea anyway. That is, we need to read and understand the other two volumes and bring other-real world elements into the overall framework. Unlike the Walrasian framework, the Marxian framework aims to correspond with empirical reality, by bringing in the heterogeneity and complexity. Of course, this doesn't mean that Marx was right about everything. Far from it. (The standard "orthodox" theory of the tendentially falling rate of profit springs to mind, though that may be a misinterpretation of Marx.) {*} In Marx's analysis, for example, the price of credit (the interest rate) does not correspond to the value contributed by the banker's labor, but he didn't see that as a form of exploitation. Jim Devine [EMAIL PROTECTED] & http://bellarmine.lmu.edu/~jdevine