Monday, December 05, 2011

Walras's Law Is False

1.0 Introduction

I have asserted that the distribution of income is determined by political power, not by intertemporal utility-maximization. Interest rates are not the result of individual decisions trading off consumption at future dates against consumption now.

Is it not a consequence of this view that Walras's law does not apply to capitalist economies?

2.0 On the Derivation of Walras's Law

Walras's law states that the sum of excess demand across all goods is zero:

p z(p) = p1z1(p) + p2z2(p) + ... + pnzn(p) = 0
The excess demand for the jth good is the difference between the demand and supply of that good at the set of prices at which these functions are being evaluated:
zj(p) = dj(p1, p2, ..., pn) - sj(p1, p2, ..., pn)
And supply and demand functions for each market are found by summing over the supplies and demands for all individuals. But individual supply and demand functions are derived from the theory of utility-maximization, given the initial distribution of the endowments of all goods.

3.0 Macroeconomic Reasoning With Walras's Law

Don Patinkin considered an economy in which n - 1 commodities and "money" are traded.

"For the amount of excess demand for money equals the aggregate value of the amounts of excess supplies of commodities." -- Don Patinkin, Money, Interest, and Prices (2nd edition, Harper and Row, 1965)
I believe Patinkin's approach can be seen as building on J. R. Hicks's Value and Capital (2nd edition, Oxford University Press, 1946). Hicks included demands for both money and other securities in his General Equilibrium approach.

One can argue that the world economy is currently in a disequilibrium; both labor and produced commodities are in excess supply. Thus, by Walras's law, there must be an excess demand for money. And it is the job of monetary authorities, such as the United States's Federal Reserve, to meet that demand, by flooding the market with money while this disequilibrium exists. At least, this is the argument of prominent mainstream "Keynesians", such as Brad DeLong and Paul Krugman (suggestions for links to their blogs here are welcome).

4.0 Conclusion

Keynes's emphasis on fundamental uncertainty is arguably incompatible with the explanation of the demand for money by intertemporal utility-maximizing in a model of General Equilibrium. Thus, the above justification of "Keynesian" monetary policy, based on Walras's law, does not harmonize with Keynes's theory.

3 comments:

Blissex said...

«I have asserted that the distribution of income is determined by political power, not by intertemporal utility-maximization.»

I am against the central verity and truthiness of Serious Economics too, but that claim as simple as it stands seems too broad and we should be careful...

The distribution of income is determined by political power within wide bounds», at least in the long term. There is a sort of reality principle for the distribution of income too.

When I try to explain this to people familiar with the usual 2D graphs like demand-supply for a single good in a market, I try to make the point that in general the supply and demand schedules are not really 1D lines, and "equilibrium" point is not really a 0D point; that the schedules should be seen more as bands or strips, and that around the intersection there is an area of many similarly plausible

Blissex said...

«Don Patinkin considered an economy in which n - 1 commodities and "money" are traded. "For the amount of excess demand for money equals the aggregate value of the amounts of excess supplies of commodities."[ ... ] I believe Patinkin's approach can be seen as building on»

Patinkin's approach was not quite original, as it was building on standard non-neoclassical thinking. Both Marx and at the opposite side Keynes wrote quite a bit about inventories of "fictitious capital" or "liquidity" absorbing mistakes in the "tatonnement" process.

«Keynes's emphasis on fundamental uncertainty is arguably incompatible with the explanation of the demand for money by intertemporal utility-maximizing in a model of General Equilibrium.»

As you know there are many problems with General Equilibrium approaches, whether intertemporal or not, and the only reason why they are popular is their truthiness about the pareto-optimality of the distribution of income.

But the fundamental link with the underlying thinking of Keynes (and Marx and many others from the opposite or same side) is that "tatonnement" is wishful thinking, and true equilibrium never happens, and prey-predator modeling may be more productive.

I haven't seen this mentioned often, but this can also be expressed by saying that notional and actual demand and supply can be rather different, and that the notional is rather more important than the actual, and that in particular ex-ante intentions to consume or save are not the same as ex-post outcomes of consumption or saving, and the difference manifests in inventory changes (inventories of financial or physical goods/assets).

But even this very limited injection of sense has been refused with the Lucas trick, because it would invalidate the central truthiness. "Tatonnement" must always be error-free, forever.

My impression is that in Keynes' own thought intertemporal effect dominate, and the links between past present and future are uncertainty/animal spirits and capital (whether financial or means of production), and errors do happen (just consider his discussion of 3 levels of stock market pricing).

I tend to agree, and it is not happenstance that both uncertainty and capital are completely absent from Serious Economics.

Indeed Serious Economics entirely lacks a theory of capital that isn't a laugh, and this despite the Cambrige Capital Controversy 60 years ago. That nobody has felt like fixing that, and that Serious Economists don't ever talk about that (outside small applied studies), while capital and its formation and use is the very core of political economy, should tell you everything. And I guess it has.

BTW I think I already mentioned this, but one of Keynes' most interesting hints was the seemingly obvious one that almost all capital is long term.

If Serious Economics were not a propaganda tool this would have been the starting point of a lot of important research, because the really curious thing is not that there are cycles, but that oscillations are so small, short term, and around a pretty clear trend, so something must acting as ballast.

But overall: BAH!

Nick Rowe said...

Hmmm. I happen to agree with you here. I think Walras' Law is false (unless you redefine it in such a way that it becomes true, but trivial).

And the problem is that money, as a medium of exchange, is not like other goods. If there are (n-1) other (non-money) goods, there are (n-1) different excess demands for money - one for each of the n-1 markets.

A Walrasian monetary exchange economy is an oxymoron. A "Walrasian" economy is one where there is one big market where all n goods are traded simultaneously. A monetary economy has n-1 markets where each of the n-1 goods trade against money.

My thoughts here:

http://worthwhile.typepad.com/worthwhile_canadian_initi/2011/04/walras-law-vs-monetary-disequilibrium-theory.html