Varoufakis & the Fallacy of Composition April 12, 2012Posted by larry in economics.
Yanis Varoufakis recently wrote a piece laying out errors often made by journalists, even those trying to do a good job. As in any profession, there exists a wide range of ability and performance. The complexity of the mainstream media sometimes prevents journalists from doing the job they are in fact attempting to do. Their efforts may be undermined by their editors or by these editors’ editors. There is, however, one problem in Varoufakis’s attempt to alert journalists to their errors. And this is that he calls one of the most famous errors in economics by a name different to that by which it is generally known. That is, he calls the fallacy of composition the fallacy of aggregation.
The fallacy of composition, which Keynes warned against, contends that it is a fallacy to generalize from what may be true of individual economic decisions to the economic system as a whole. That is, what is good for individuals may not be good for the group as a whole. Keynes is implicitly referring to a problem that some philosophers have worried about – that of emergent properties. And it may be that the fallacy of aggregation is different from the fallacy of composition. Let me use an extension of Wittgenstein’s example of the ladder to show what I mean.
Let us suppose that we have a number of pieces of wood lying all over a large piece of ground. Let us further suppose that there are two long pieces and a number of short pieces of the same length lying around higgledy-piggledy. Now these pieces of wood lying on the ground all over the place do not make up anything at all. They are nothing more than pieces of wood.
Let us now suppose that we “aggregate” the individual pieces of wood into a single pile. This aggregate still doesn’t make up anything other than a pile of wood. In particular, they don’t make a ladder. What is needed to turn this aggregate of pieces of wood into a ladder? To obtain a ladder, you need to add organization or structure and possibly some glue to hold the pieces together to get your ladder.
It could be reasonably argued that the ladder is more than just the aggregation of the parts, which gave us only an unorganized pile of wood. The parts needed to be composed together in some way in order to produce the ladder. The ladder could be said to be an emergent property arising from the individual pieces of wood. The additional ingredient needed to bring about the potential emergent property need not be physical, but could be conceptual, as it is in the case of the ladder (ignoring the glue for the moment). This would suggest that the fallacy of aggregation, which is a fallacy, can be distinguished from the fallacy of composition, and that they are somewhat distinct fallacies. Nevertheless, in the case of the ladder, it may seem that while the fallacy of aggregation can be seen to be a fallacy, the fallacy of composition might appear to be somewhat different and, therefore, not thereby necessarily in this case a fallacy. That this is only an illusion due to the simplicity of the example, let me show by means of the concept of society and the relation it has with the individuals associated with it.
Thatcher once famously said that there was no such thing as society, that there were only individuals. This involves a serious error that was not noticed by pundits at the time. Were this true, there could be no case made for a fallacy of composition here. But the statement can be easily shown not to be true.. A society isn’t just a bunch of people aggregated together simpliciter. Nor is it a bunch of isolated individuals arranged or composed in certain ways. A lot of other factors need to be added in order for the mix (composition) to add up to a society. The sorts of factors needed to obtain a society from a bunch of individual people are the social rituals that they tend to perform together and the cultural constructs that they use to interpret the world around them, for example. The social rituals, underwritten by cultural constructs, can be viewed as comprising part of the glue that holds a society together. If that is the case, then a mere aggregation of people does not and cannot by itself constitute a society nor would mere composition of the individuals into some sort of group unit on its own be sufficient to create a society or functioning socio-cultural group. While composition is distinct from (mere) aggregation, neither are themselves sufficient to produce a functioning group out of disparate individuals, however necessary they may be, as I think my previous discussion of Wittgenstein’s ladder shows. While both are fallacies, they are slightly different fallacies.
This is not the place to go further into the nature of this deep error other than to notice that many journalists and politicians at the time did not seem generally to understand how deeply erroneous Thatcher’s statement actually was.
One might argue that the journalists were just badly educated. While this is also true of politicians, some of these politicos know exactly what they are doing. I place Cameron in the badly educated group. Eton failed this chap, as did Oxford. But they succeeded in one respect. While they failed to truly educate Cameron, they did successfully inculcate at least part of the elite’s propaganda into his brain.
But, where do we put Osborne? Or the civil servants at the Treasury? In order to avoid committing the fallacy of composition, we perhaps shouldn’t lump all the individuals at the Treasury together. Or even all of the brain of a single person. After all, people’s brains contain all sorts of crap mingled among whatever jewels there may be. Brains and organizations of people are considerably complicated things.
One might further argue that it isn’t the individual brain’s fault that it contains a lot of crap, but that of the system into which it has been embedded since it came into the world, brought about by a great deal of social and cultural programming in order to turn an infant into a functioning social person . Of course, a brain doesn’t contain just crap. But it is the crap that, in the end, causes it problems, especially when the brain ‘sees’ the world in a way that fails to adequately reflect the ‘real thing’. Now, can a single brain, on its own, distinguish a piece of intellectual crap from an intellectual jewel? Indeed, can a ‘community’ of brains effectively accomplish this? Or does referring to a ‘community’ in such a scenario amount to nothing more than begging the question?
Sidestepping these questions, the ‘idea’ (conceptual construct, or model) that fails to successfully refer to the way the economy works is neoclassical economics and its many variants. Varoufakis contends that economics involves inherent errors, and while one might agree with this, some models are more erroneous than others. And since they involve social policies that can negatively affect real people, the degree of error is of great importance. One of the more erroneous ways of looking at the workings of the economic system is the neoclassical one. One of the less erroneous is the Keynesian. Another, perhaps even less erroneous, is the theory developed by Kalecki over a series of papers. In a review of Keynes’ General Theory, Kalecki corrected a serious error in Keynes’ theory of effective demand. It is unfortunate that he never felt able to write his own book on the subject — but he said that he became “ill” when he read Keynes’ treatise on the subject.
Getting back to the fallacy of composition, it is not only economists who commit this fallacy. For a long time, many psychologists took for granted that there was such a thing as general intelligence and that it was composed of myriad individual “intelligences”. They were all combined together into a single number referred to as the individual’s IQ. The “thing” that IQ measured was something Spearman called “general intelligence”, which has become known as Spearman’s g. This entity, g, does not exist and therefore can not be measured. If IQ is considered to measure anything, it is the ability to take certain culturally biased tests. If there is something that can be called general intelligence, and it is not entirely clear that there is consensus on what the term means, then it is something other than a simple (or not so simple) aggregation of intelligent parts.
Something similar can be claimed for certain traditional concepts in economics. For instance, there may well be no referent for the term GDP (or GNP) that has any real sociopolitical relevance, in the sense in which it is generally used by mainstream politicians and economists. A number of economists and related scientists have wondered over the years whether certain general notions such as GDP, conceived as collections of a number of individual measurements of different factors, actually ends up referring to anything meaningful. Indeed, the factors whose measures that make up, say, GDP can be seen as problematic themselves. A read of Myrdal’s Against the Stream should convince anyone of this who is not already convinced otherwise.
More could be said about this, for example, that the collection of individual measures into collective measures involves more than composition fallacies, but I will have to leave this matter for another time.