In standard neoclassical theory, rational actors must have reliable probability forecasts::
“To make statistically reliable forecasts of the future, agents need to obtain and analyze sample data from the future. Since that is impossible, the assumption of a predetermined-ergodic-reality permits the modeler to assert that sampling from past and present market data is the same thing as obtaining a sample from the future. Ergodicity implies that future outcomes are merely the statistical shadow of past and current market signals. Presuming ergodic conditions reduces the modeler's problem to explaining how and at what cost agents obtain and process existing data (in the form of ‘price signals’).Davidson proposes the following classification of the way in which mainstream and heterodox economic theories treat economic reality and human knowledge of the future:
Unlike the old classical economists, rational expectations theorists do not claim that the agents in their models obtain complete knowledge of reality. Rational expectations models only require agents to use existing market price signals to calculate subjective probabilities that are statistically reliable estimates of the objective probability function describing the reality that governs future events. Subjective probabilities calculated from current and/or past market data can provide these statistically reliable estimates if, and only if, the economic system is ergodic. Hence, all rational expectations models are based on the ergodic axiom.” (Davidson 1996: 480).
“Concepts of External Economic RealityThe Type 2 models assume that in the short-run economic agents are ignorant about the immutable reality, and have very incomplete knowledge, because there are serious limitations on the human ability to collect and analyse the time series data necessary to obtain reliable knowledge (Davidson 1996: 484). In type 2 models, economic agents are therefore subject to a type of epistemological uncertainty.
A. Immutable realityType 1. In both the short run and the long run, the future is known or at least knowable. Examples are:B. Transmutable or creative reality: Some aspects of the economic future will be created by human action today and/or in the future. Examples of theories using this postulate are:
a. Classical perfect certainty models.
b. Actuarial certainty equivalents, such as rational expectations models.
c. New Classical models.
d. Some New Keynesian theories.
Type 2. In the short run, the future is not completely known due to some limitation in human information processing and computing power. Examples are:
a. Bounded rationality theory
b. Knight’s theory of uncertainty
c. Savage’s expected utility theory.
d. Some Austrian theories.
e. Some New Keynesian models (e.g., coordination failure).
f. Chaos, sunspot, and bubble theories.a. Keynes’ General Theory and Post Keynesian monetary theory.
b. Post-1974 writings of Sir John Hicks.
c. G.L.S. Shackle’s crucial experiment analysis.
d. Old Institutionalist theory.” (Davidson 1996: 485).
Regarding decision making in the Type 2 theories, Davidson notes:
“Type 2 immutable reality models typically employ a subjectivist orientation. Agents form subjective expectations (usually, but not necessarily in the form of Bayesian subjective probabilities). In the short run, subjective probabilities need not coincide with the presumed immutable objective probabilities. Today’s decision makers, therefore, can make short-run errors regarding the uncertain (i.e., probabilistic risky) future. Agents ‘learn’ from these short-run mistakes so that subjective probabilities or decision weights tend to converge onto an accurate description of the programmed external reality.” (Davidson 1996: 486).Davidson then draws attention to the distinction between risk and uncertainty in Frank Knight’s work:
“the practical difference between the two categories, risk and uncertainty, is that in the former the distribution of the outcome in a group of instances is known (either through calculation a priori or from the statistics of past experience), while in the case of uncertainty, this is not true, the reason being in general that it is impossible to form a group of instances, because the situation dealt with is in a high degree unique” (Knight 1921: 233).Finally the Post Keynesian view, as derived from Keynes himself:
“For Keynes and the Post Keynesians, long-run uncertainty is associated with a nonergodic and transmutable reality concept. A fundamental tenet of Keynes’ revolution ... is that probabilistic risks must be distinguished from uncertainty where existing probabilities are not reliable guides to future performance. Probabilistic risk may characterize routine, repeatable economic decisions where it is reasonable to presume an immutable (ergodic) reality. Keynes ..., however, rejected the ergodic axiom as applicable to all economic expectations when he insisted that the ‘state of long term expectations’ involving non routine matters that are ‘very uncertain’ form the basis for important economic decisions involving investment, the accumulation of wealth, and finance. In these areas, agents ‘know’ they are dealing with an uncertain, nonprobabilistic creative economic external reality.” (Davidson 1996: 492–493).Shackle’s concept of the “crucial choice” is also relevant here. A “crucial choice” decision is one that has a fundamental influence on the economic environment, and the conditions under which the decision is made are not repeated (Davidson 1996: 495). The transmutable economic future is created by such decisions, but often contrary to what agents intended. As Davidson notes, for Shackle, the
“future is not discovered through the Bayes-LaPlace theorem regarding relative frequencies or any error learning model.” (Davidson 1996: 499).Furthermore,
“[s]ome economic processes may appear to be ergodic, at least for short subperiods of calendar time, while others are not. The epistemological problem facing every economic decision maker is to determine whether (a) the phenomena involved are currently governed by probabilities that can be presumed ergodic – at least for the relevant future, or (b) nonergodic circumstances are involved.” (Davidson 1996: 501).Where economic phenomena are non-ergodic, “discovered empirical regularities in past data cannot be used to predict the future” (Davidson 1996: 502).
Davidson, Paul. 1996. “Reality and Economic Theory,” Journal of Post Keynesian Economics 18.4: 479–508.