Fitzgibbons’ Foundations of Keynes’ economics
Athol Fitzgibbons The microeconomic foundations of Keynesian economics, in Dow and Hillard.
This rejects the ‘decision-by numbers’ version of rationality, advocating ‘judgment’ where the numbers are not soundly based, and highlights the importance of Keynes’ concept of uncertainty. Unfortunately, it also misrepresents the potential role of mathematics beyond numbers.
II RATIONALITY, UNCERTAINTY AND EXPECTATIONS
If all decisions could be based on a mathematical calculation, the narrowest form of rational behaviour would dominate the markets, since the greatest fool could hire a rational thinker to make decisions on his behalf. When we speak in a shorthand way of irrational behaviour we do not usually mean a wilful insistence on being foolish, but behaviour arising out of some misconception in a complex and diﬃcult situation.
Here ‘mathematical calculation appears to mean ‘arithmetical calculation’, as in the conventional theory of rationality. It does not appear to be the intention to criticise Keynes’ broader use of mathematics.
In many strategic games there is no such thing as an optimal decision, because there is no uniquely best strategy when the other players are free to generate surprises.
[O]ptimization theory … does impose stringent restrictions on behaviour that do not apply in general. Optimization implies the existence of a uniquely best decision … . However, rational motives alone, in the context of uncertain information, cannot ensure the existence of a uniquely best decision… .
III CLASSICAL OPTIMIZATION
[E]conomic decision making must often respond to scientific, technological, moral and political propositions, which are not susceptible, even in principle, to quantification.
IV THE LIMITS OF ‘AS IF’ THEORIZING
… Even when it does not apply strictly, the theory of the consumer might still provide a useful template for investor behaviour. Decision makers can act as if they knew the future and were maximizing profits, just as – this example was given by Milton Friedman – professional billiards players can act as if they understand the laws of mechanics.
It is clear that to live intelligently in our world – that is, to adapt our conduct to future facts – we must use the principle that things similar in some respects will behave similarly in some other respects even when they are very different in still other respects. (Knight, 1921: 206).
‘Economics’, Keynes (CW XIV: 296) said, ‘is a science of thinking in terms of models joined to the art of choosing models which are relevant to the contemporary world’;
All decision makers, not just economists, need to practise ‘vigilant observation’; that is, they need to select the leading features of a situation before they choose which theoretical model will be the most reliable. The same creative powers of judgment that are required of the strategic game player will be required of the creative scientist or the lone mountain climber, and as many economists have observed they are also required of investors and entrepreneurs.
If decision makers were to change their perception of the future, and vigilant observation sometimes suggests that they should do so, they would change their preferred mental model.
Here a ‘pragmatic’ approach, of selecting a single model, is advocated. But not justified.
[M]athematical logic does not determine which analogy with reality is the best unless there is full knowledge. Analogies will have to be drawn between reality and the theoretical assumptions, and a decision maker may revise the theory altogether if it reaches implausible conclusions.
It might also be noted that Jack Good’s probabilistic logic, a development of Keynes’ work, could be often be used to identify a model which is ‘best’ in some useful sense. But would this be sensible?
It might seem hyperbole to describe the inexact process of judgment as rational, because (unlike optimization) there is no demonstrable connection between the facts and the decision, the decision is revocable, and it might turn out to be wrong, or it might be impossible to know whether the decision was wrong. To dramatize his argument that only perfect knowledge is consistent with rational decision making, G.L.S. Shackle once said that no one can pass over a bridge with a broken span. However, the reply is that ignoring the available information because it is incomplete would be self-defeating, and if a decision must be made then part of a bridge can be better than nothing. A commitment to think rationally means that we go as far as the formal logical bridge extends, and then rely on metaphors and analogies to carry us as far as they can. There is no other way, and that is how decisions are made and policy is formed, in both the public and the private sectors.
As Keynes pointed out in his work on induction, the connection between ‘facts’ and decision is at best conventional, and subject to group-think, rather than being purely logical. We always rely on metaphors and analogies. We should at least strive to use the best possible metaphors and test them logically, rather than rely on intuition, no matter how deeply ingrained such intuitions might be in our culture or ideology.
V HOW UNCERTAINTY AFFECTS DECISIONS
Sometimes exact relationships do hold over long periods of time, even when there is no apparent reason why they should, but at other times they break down in the most spectacular way. They have broken down in the product market (rising Phillips curves), the stock exchange (bubbles and crashes), the money markets (secular and cyclical changes in velocity) and in the markets for foreign exchange. Yet at other times prices have remained constant over long periods of time, despite major changes in the parameters underlying the market.
Often what we witness is not random market behaviour, but an irregular and erratic lurching that seems to suggest the operation of invisible causes.
Intransitive behaviour can be rational whenever there is no unique best choice, or … whenever it is impossible to insure against a bad decision.
To move with the herd, following closely or trying to keep just one step ahead, is never an optimal strategy, and sometimes it can lead to disaster. However, if no optimal strategy exists, following the herd is one way of benefiting from pooled judgements and buying asset insurance.
This applies to the selection of ideologies and models.
Workers, firms and asset owners do not react in the full unison that would occur if they had a common evaluation of the future.
Thus diversity improves stability.
[U]nquantifiable probabilities can also impose a conservatism on decision making, which causes the economic system to move discontinuously when it moves at all.
Unquantifiable probabilities … can also turn the economy into a cataclysm system – although the economy is more stable than classical theory would suggest, every now and again a cataclysm occurs.
[U]nquantifiable probabilities elasticize the causal link.
These are all supported by Keynes, and can be demonstrated by appropriate mathematical modelling.
VI RADICAL UNCERTAINTY
As the degree of uncertainty deepens, value judgments tend to outweigh judgments of fact because the former become more reliable.
Note the dichotomy.
VII CONCLUDING COMMENT
Keynes once noted that we live in the transition.
This is a good insight, that merits development. Clearly, rational decision theory leaves no room for free will, and hence ‘life’, so life is evident in those areas where rational decision theory does not apply, such as strategizing.
What we need at present is not a superior theory, but humility to recognize, in a very systematic way, the limits of economic decision making.
Or perhaps we could follow Keynes in improving the theory?
This is full of good insights, contra rational decision theory, but I am not clear what I am intended to understand by terms such as ‘comprehension’ and knowledge’, since no theory is given.
It seems to me that:
- There is no recognition of mathematics beyond the numerical.
- A dichotomy is being drawn between arithmetical calculation and human judgement, presumably based on experience.
This would leave no recognition of or room for the type of mathematics that Keynes and his associates (Whitehead, Russell) developed and applied, yet this mathematics seems central to Keynes’ work on economics. To draw on Shackle’s analogy of a bridge, could we not use the mathematics of Keynes or that developed from Shackle’s work to help build the bridge?
Further, the paper seems to advocate a form of pragmatism, in the sense of identifying a preferred model an using that, while adopting ‘vigilant observation’ for contrary indications. It may be that in practice this is adequate, but Keynes’ own approach seems to me to be more relevant to periods such as 2006-2012, at least.
For example, Keynes’ theory of induction shows that the longer a model is consistent with the evidence, the more probable it is that it will continue to do so – as long as there is no substantive change. Thus if multiple models are consistent with the evidence then they will tend to be consistent with other, and differ only in details and inessentials. Thus one could –pragmatically – combine them into a single model with the desired level of detail. But the validity of this preferred model is still subject to there being no substantive change, as will happen from time to time in complex adaptive systems. When the preferred model is violated it follows from Whitehead that there will often be some more general model that remains valid, and which can continue to be used. It may even by that the system has ‘modes’ for which we have models that can be dusted, adapted off and re-used when the mode changes. Thus even where one is mainly acting on a single preferred model it could pay to maintain alternatives. Indeed, from a logical point of view it does not matter how vigilant observation is unless it is ‘in effect’ informed by some such modelling of potential transitions. Some of Keynes’ work relating to causality would also seem relevant.