In Response to Paul Davidson

Paul Davidson makes a most welcome appearance on this blog by commenting on an old post here. He summarizes his position by writing:

"But recent history shows that leaving it to the unfettered market almost always ends in a collapse of the system -- not necessarily immediate but potentially some time in the future."

I agree with this. In fact, Professor Davidson's work is directly responsbile for my belief in this statement. Markets are inherently unstable for the reason that "In order for a market to be efficient it must be possible to reliably predict the future (at least in a statistcally reliable sense)". This is true not because of the complexity of the relevant data, but of the ontologically uncertain environment in which entrepreneurs have to calculate prospective returns. I am on board here, and I think Mr. Davidson makes a mistake in writing of me that:

"I am afraid tht you neither understand my argument or the importance of the difference between ontological uncertainty and eistomological uncertainty, where complexity wopuld be an example of the latter."

My point was that I think Mr. Davidson makes a mistake in believing that government institutions can control and reduce uncertainty. There is no reason to believe that they can. Here is Mr. Davidson again:

"If the future is nonergodic and not reliably predictable without government institutions that constrain possible future outcomes, then market economies are inhenerently volatile, if market outcomes are not regulated or are under regulated. Government has a role in stablizing potentially unstable markets and therefore providing some market benefits to all the population."

Now markets are unstable. This is a fact we must learn to live with. But it does not follow that governments have a role in taming them through appropriate regulation. The first question that comes to mind would be: How would they go about doing this?

If the source of market instability comes from ontological uncertainty (and not complexity), then we should expect government to be helpless in this matter. And this is precisely what Professor Davidson's analysis implies, his theory of uncertainty being in this order: (1) ontological uncertainty; (2) a non-ergodic (i.e., unpredictable) choice environment; and (3) a transmutable reality. The inapplicability of "probabilities calculated from historical data" also supports my argument the economy is not sufficiently stable to provide scope for positive regulation.

For example, if aggregate demand falls on account of an increase in saving, Mr. Davidson believes that government should act to increase its expenditure on goods and services produced by the private sector in order to "increase the sales of industries, thereby, encouraging entrepreneurs to increase employment especially in the industries from which the government directly purchases, e.g., military hardware." For this reason he dismisses the view that there exists an "automatic mechanism that ensures that private spending on consumption will be just sufficient to assure full employment." The stimulus to employment provided by government expenditure must wipe out, according to Mr. Davidson, any unfavorable positions in industry and promptly establish a new equilibrium at full employment. Mr. Davidson thus regards government fiscal policy as a "balancing wheel" that is responsible for keeping aggregate demand at its full-employment level, in spite of the argument he advances in favor of a "transmutable reality."

Mr. Davidson's explanation is that the aggregate demand function, which represents total spending on producibles at every level of employment, will not have "the same values as the value of aggregate supply function at each level of employment, i.e., the aggregate supply function and the aggregate demand function would not be identical for all possible levels of employment." From this, in his view, all the rest follows. But this not correct.

Now, in so far as the direct effect of government expenditure is concerned, its influence depends largely on the ability of policy makers to determine the areas that have been most affected by falling demand. The government can, of course, consult the "historical data" gathered by the Bureau of Labor Statistics. This option is useful however only in an ergodic world, e.g., if "the outcome at any future date is the statistical shadow of past and current market data." The existence of a non-ergodic world, which was used by Mr. Davidson to demonstrate that "unregulated financial markets" cannot optimally allocate resources at the full-employment level, transforms every act into a crucial decision, e.g., that decision which is made "changes forever the economic environment so that the identical decision conditions are never to be repeated."

This is why I believe Professor Davidson has failed to extend the implications of his theory to government regulation. Take, for example, this passage: "For Keynes and the Post Keynesians, long-run uncertainty is associated with a nonergodic and transmutable reality concept. A fundamental tenet of Keynes's Revolution ... is that probabilistic risks must be distinguished from uncertainty where, for the latter, probabilities calculated from historical data are not reliable guides to future performance." The consequence of this is that the optimal level of income and activity is ontologically indeterminate.

-----

Paul Davidson makes a powerful concession here: "Keynes did not think everything occuring in the economy was uncertain in the nonergodic sense. But he emphasize that decisions that will result in possible payoffs five years or ten years from now were an act of faith -- relying on animal spirits -- gambles that could result, in an unfettered market, in the possibility of huge mistakes."

I would like to know what things are not uncertain in a non-ergodic sense? What of Shackle's crucial decision theory, or the idea that all knowledge claims are mere conventions that are designed to mask the reality of uncertainty? What of the transmutable reality of which Mr. Davidson spoke earlier?

Mr. Davidson concludes his post by suggesting that I have failed to distinguish non-ergodic uncertainty from complexity. In fact, however, I believe myself to be completeing Professor Davidson's non-ergodic uncertainty paradigm.

----

With all that being said, let me conclude by saying that I really admire Professor Davidson's work. He has been one of the most influential people on my thinking. Shackle was great on uncertainty, but Professor Davidson actually put this to use in his analysis of financial markets and industry. Everyone should read his 1972 book "Money and the Real World" to get an idea of what Post Keynesian Economics is all about.