02 November 2008

Economics needs a scientific revolution: a comment

The essay Economics needs a scientific revolution by Jean-Philippe Bouchaud has been published in 30 October 2008 issue of Nature 455, 1181, is stimulating and an interesting read. It points out the inability of mainstream economics or financial engineering to either predict or avert a crisis.

This outcome is a result of the huge divide between normative (theory) and positive (empirical) economics. In the process, as Bouchaud indicates, assumption have taken the form of axioms: rational economic agents maximize profits, each agent's pursuit of profit is also best for society and markets are efficient that through prices reflect all known information. At best, these results are about some ideal situations, which do not exist in real life. Striving for a particular ideal situation is one thing, but considering that the features of this ideal system are a part of the real life situation is another. It takes us to a belief system.

Adhering to a belief system can make updating of knowledge difficult. For instance, Tycho Brahe, who despite being credited for his accurate and comprehensive astronomical and planetary observations, could not entirely got out of the Geocentric model that the Earth is at the centre of the universe. It was his assistant, Johannes Kepler, who used the observations and came out with his three laws of planetary motion that conclusively showed that the Copernican system, Sun is at the centre, is correct.

There are no two opinions that a scientific inquiry requires a questioning mind. If required, questioning the very premises by cross checking with empirical observations. The reign of empiricism is in a Bayesian world. A minimum requirement for this is that your prior cannot be certain - you cannot begin with a belief system. The reason is that if your prior is either zero or unity, then whatever be your observation, your posterior will be equal to your prior (see my working paper Understanding Fundamentalist Belief Through Bayesian Updating. All empirical observations become irrelevant.

True, the models used by contemporary mainstream economics fail to explain how small perturbations can have 'wild' impacts. It is not true that fringe attempts by econo-physicists (including Bouchaud himself) and behavioural economists are not being taken seriously. Paul Ormerod's Butterfly Economics that followed the East Asian Crisis is a best seller (aside - most Economists may not know of this). Daniel Kanheman, a behavioural economist, received the Nobel in 2002. There have been some hiccups, but then many prominent economists have done work on asymmetric information, uncertainty and missing market that cannot be brushed aside.

Having largely agreed with Bouchaud, I point out some differences. First, inability to predict and avert a crisis cannot be a basis to gauge a discipline's success or failure. Just as the space shuttle Columbia's disaster on 1 February 2003 cannot be construed to be a measure of judging Aeronautics and Space research, the current financial crisis cannot be a measure of judging Economics. Crisis or disaster by nature is rare and not predictable. Like the violation of some safety regulations in the Columbia disaster the current global financial crisis, to put it simply, is an outcome of greed by some players who manipulated the markets for their advantages. Modelling apart, an economic system should have effective regulation to safeguard us against possible disasters.

Second, understanding the market with more appropriate tools and techniques of modelling is necessary, but it is more important to acknowledge that market is a tool, albeit, an important one. It is a means, but not an end in itself. This is the approach of the larger human development paradigm, which is people-centric.

Third, which in a way is related to the previous point, is that the dominance of utilitarianism, and hence, of markets in economics, was first successfully challenged by a prominent moral philosopher of the 20th century, John Rawls. He argued against the monoconcentration of utilitarianism and its associated formulaic reductionism in favour of plural concerns. On his concerns of justice his focus was on the least advantaged while assigning priorities to equal liberties and equal opportunities (that is, against arbitrary privileges). This is made possible by invoking the original position, that is, by putting decision makers under a veil of ignorance - they do not know which group they belong to. This in essence does away with vested interests. Thus, the silent scientific revolution, which goes beyond Economics, is already under way.

1 comment:

  1. Response from Bouchaud
    I have no fundamental disagreement with what you say. I even like what you say about Tycho Brahe and the importance of non certain priors. My emphasis at the beginnig of the paper on the crisis is more of a catch than anything else -- although I think unwarranted modelling and faith in models had an impact that shouldn't be belittled.

    I also think that tools from complex systems offer some important insights on how collective effects appear,
    and can help improving markets as a tool -- beyond simple regulations.

    My response
    I agree with the two points that Bouchaud makes.