Monday, 8 November 2010

Evolution of bagatelles

John Neville Keynes, in "The scope and method of political economy", 1891, highlighted two schools of political economy.  He prefers to regard them as differences of emphasis and notes that are both present in the great economists.  Adam Smith, he claims, didn't dwell on methodological issues explicitly but proponents of the two approaches both claim Smith for their side.  You could fancifully imagine that the original Smithian approach fractured into two, embodied in Ricardo and Malthus:

1. Is-based, theoretical, abstract, deductive, ahistoric , a priori, logical essences of human nature, Ricardo
2. Ought-based, ethical, realistic, inductive, historical, a posteriori , contingent evolved economic states of affairs, Malthus

Keynes, like J.S. Mill and Senior, are on the side of political economy as logical science.  One of the claims of this school is the ability to isolate more or less pure economic lines of reasoning which are distinct from general human social reasoning.  This empowers the scientist of economies to postulate models of behaviour which are exclusively economic in nature.  And no doubt Auguste Comte's views on the classification of each of the social sciences, and the possibility that they are allowed their own scientific models (grounded in mathematics, which he sees at the root of all science) dealing distinctly with phenomena from that subject have a large role to play in philosophically underpinning this school.

I think I see in this Comtian idea the beginnings of the intrinsic/extrinsic, modelled vs externally shocked dichotomy which permeates economic thinking and language to this day.  Keynes says:
Economic facts are, it is allowed, influenced by social facts of very various kinds, and in their turn influence them; but it is nevertheless held to be possible up to a certain point to isolate the study of the phenomena of society ... Economic science constitutes, therefore, a distinct, though not entirely indpendent, department of sociological speculation.
Keynes gives two more prosaic reasons to favour the deductive theoretical school - (i) we can't carry out economic experiments, hence undermining the more empirical of the two approaches; (ii) given the 'leakage' of causal explanations into general sociological departments of human behaviour, inducing a coherent and pure economic theory is very difficult.  Point (i) seems less relevant to us today, given the new subject of experimental economics, and (ii), I would have thought, has a similar flip-side w.r.t. the deductive approach - namely the artificiality of the 'purely economic' simplifying model could render theories inspired by this approach intellectual bagatelles with no serious value.  I think this is probably too strong, given the subsequent productive directions economics has moved in following Ricardo, J.S. Mill and Cournot, but nonetheless it can't be taken for granted that in general any theoretical economic model will necessarily be useful.

The axioms of the deductive approach are essentially axioms about human nature - for example,  all men would prefer the outcome of any economic behaviour they engage in to increase their wealth.  All the other idiosyncratic motivations can, by definition, be assumed to be random noise around the basic tenets, and hence can be ignored in models as a simplifying first assumption.  But nobody at the time sees this as anything more than a model of a hypothetical man.  Rational expectations assumptions were still a long way off.  Still, it is indeed not a bad approximation to identify these axioms and consider them static for the purpose of the model.  Yet of course, there's always a historically embedded, institutional version of even these axioms of human nature - how our cognitive facilities developed, no doubt as our material environment and maturing culture evolved.

Keynes points out that in reality, neither of these two schools were as different from each other as they claimed (except in the positions of their own extremists).  This sounds right to me.  Take, for example, the efficient markets hyppothesis.  Markets are clearly historically situated; and they have been moving towards a position of achieving efficiency through the construction of historically embedded institutions of trading, (exchanges, electronic exchanges, S.E.C. protections, corporate govenrances, etc.)  Yet the mathematician who jumps straight to the matehmatical asymptote of this essentially cultural evolution may also be justified in constructing his model with these assumptions.  They are two sides of the same analytical coin, for me. 

It is interesting to see how Keynes distinguishes the positive science (IS) of political economy with the ethics (OUGHT) of political economy and the art of political economy.  His son would contribute to all three.  Also, it isn't clear to me how the OUGHT school got to be associated with an empirical approach.  Aren't there just as good a set of reasons to associate IS with the empirical approach?  Conversely the OUGHT school could be considered theoretical and ahistoric - after all, a moral system is often considered timeless and theory-based.

Keynes is an advocate in positive economic science of the principle which resists any final vocabulary, since the phenomena of economic activity are driven not only by human beings' self-interested motivations, but also by their other-regarding motives.  And the particular mix of these two must be historically contingent and subject to change.  This position is akin to adding a historical element to Smith's "Theory of moral sentiments".  Shouldn't the key distinction be not between self-interest versus altruism but long-established pieces of cultural-biological context versus more recent cultural context?