1
We need to distinguish between two aspects of the discussion on the theory of value that get conflated often. (This distinction is from Rubin, but Sweezy popularized it.) In the narrower sense, the theory of value is about what explains the rate at which commodities exchange or -- perhaps more precisely said -- what determines the *change* of commodity exchange rates (e.g. price changes). Plainly said, in this sense, the theory of value is about what explains price fluctuations. This is the sense in which Stigler interpreted Ricardo's as a "93% labor theory of value."
Now, when you ask an economist what the theory of value is about, they will tell you this is it. And most economists today will claim that you do not need to pin down an (ontological or epistemological) "substance" of value (e.g. labor or "utility") to develop a practically useful guide to understanding price changes. This is -- of course -- correct. In the same sense that it is "correct" to eat food without giving a damn about how the digestive process really works or driving a car without knowing of Newton's laws.
So, yes, we can have a reasonably good guide to how prices fluctuate with no reference whatever to what underlies commodity exchange. One may "predict" price changes (with some degree of certainty) without having to view exchange in its totality as the metabolism of social *labor,* as total social labor being shuffled and re-shufled around so that a society can reproduce itself.
But, it is also true that a robust understanding of price fluctuations does not exclude an explicit admission of the fundamental fact that the ultimate resource societies have in its "trade" with nature (what we call "production") is societies' purposeful, conscious activity time: labor. Learning physiology won't make food taste better. Knowing mechanics won't make us better drivers. But that's because the practical needs considered are limited to enhancing the flavor of food and improving driving skills. There are, however, other practical needs that do require knowing mechanics and human physiology -- namely manufacturing cars and improving one's digestive health. Similarly, there are practical political tasks that do require the understanding of commodity exchange as the metabolism of social labor, namely building a workable communist society.
The emphasis on the (historically-materialistic) aspects of the labor theory of value (the historicity of value, of the duality of labor as producer of commodities, of commodity exchange as being fundamentally a specifically social form in which total social labor gets reallocated) was uniquely Marx's. They are all missing in Smith and Ricardo. A hint of this is still missing in most modern economic theorists (with few exceptions, Trygve Haavelmo and -- more recently -- William Sharpe, to name a pair). I say this, because the way I read Mike's claim is not as saying that Marx was pointing out of labor theory of value in this general sense, but rather pointing out of the *Ricardian* apparatus to grasp price fluctuations. In the fundamental sense, Marx's theory of value remains a 100% labor theory of value.
2
Mike tells the story of Marx picking up on Ricardo's notion that, even if you account for the fact of profits proportional to capital stock, the bulk of price fluctuations result from fluctuations in relative labor requirements. I have a trivia story on this. In the mid 1990s, I was Anwar Shaikh's research assistant at the New School, and one of the things I did was test this Ricardian proposition. At the time, the BLS had an old but large data set with detailed annual estimates of direct labor requirements in various industries. The BLS had recently discountinued the set, but I remember it covered a fairly long period of time (from the 1950s to the 1980s, I think). I spliced this data set to another one that had (deflated) prices for the same industries over the same period of time.
Anwar was not very comfortable with econometrics, but we did agree on a simple procedure to see whether the annual changes in prices covaried with the annual changes in direct labor requirements. To ascertain cointegration (avoiding spurious time-series covariation), I "de-trended" the data using several conventional procedures. So, basically, we had cointegrated residuals of both price changes and direct labor requirement changes. The correlation coefficients were around 93%! If the data was any good, the changes in the capital composition of industries had little effect on prices. No doubt, Ricardo was a smart man. (The paper was left unpublished.)
3
Mike Beggs wrote:
"It becomes clear that supply and demand - as modern economics conceives them - actually do play a role in establishing long-run values and prices-of-production, not merely the fluctuations around them."
Absolutely. Most Marxists today have a very superficial understanding of modern economics. I remain hopeful that new generations of social fighters will take the trouble of digesting modern economics -- rather than just, pre-rationally, declaring it erroneous or irrelevant.
With regards to Mike's attempt to clarify the implications of Marx's notion of socially *necessary* labor time, I'll just say that seriously grappling with the issue is very unlikely to leave Marx's notions intact. Marx's discursive (pre-analytical) treatment of the issue is lacking, and for understandable reasons. Marx wrote the final version of his chapter 3, volume 1 after he wrote the drafts that Engels would later compile into parts 1 and 2 of volume 3. He intended to rework the latter, but he died too soon. So, no wonder, connecting these two segments of Marx's work is not straightforward. I think that one of the benefits of engaging modern economic theory (at its highest level) is that whatever results from the exercise will have a greater analytical punch, and wrinkles like these will be easier to straighten out.
Finally, I'll say that, at its highest level, today's economic theory views demand and supply are not viewed as "schedules" or "curves" but as *mappings* or -- more restrictively -- as (multivariate) *functions* of "parameters" reflecting some "fundamentals" or (in the dynamic specification) some "initial conditions." Schedules and curves are merely used as didactic devices. But the main point that (I think) Mike makes stands, namely that in modern economic theory, supply and demand are extremely general concepts that don't exclude the labor theory of value in its most general (historical-materialistic) sense. On the contrary, at least to me, they make much more sense when you view them in the framework of Marx's labor theory of value.