Tuesday, April 5, 2011

The GWAS of economics?

Why don't economists understand economies?
To try to understand why economics didn't predict or prevent the recent economic crash, BBC Radio4 has done a series on the history of the discipline.  The first of the three programs was about economics as a story of morality, the second about economics as a science, and the third about the psychology of economics.  The central question of the programs is, in a nutshell, why economists can't predict the economy.  Because we write a lot about why geneticists or epidemiologists can't predict disease at MT, this seems like an appropriate topic to expand a bit upon here. The details are all different....but the phenomenon may be similar.

Economics as morality
Economics as a discipline began in Greece, as philosophers like Aristotle thought about the market and how it shaped and reflected morality. Wealth should be used for the Good Life. And indeed many people think of the recent global economic crash as a giant morality tale, with the market forces of greed and evil resulting in harm to millions of innocent victims.  The conclusions people draw about morality, however, seem to vary according to the observer.  To some, the victims weren't all that innocent, or the government should have intervened earlier, or shouldn't have intervened at all, or the bankers should either be stoned or given bonuses, and so forth.  A bit like assessments of success or failure in genetics.

The program discussed the origins of economics as an institution that evolved to build trust among humans, humans being the primates that build trust via "psychological interactions and formal institutions," while, the guest informed us, chimps build trust by coalition management, and bonobos by having sex.

But this seems not only simplistic but a blatantly specious argument because, if economies exist to build trust, why do we need so many laws to regulate economic behavior?  Anti-trust laws, laws against insider trading, laws about how financial institutions work, and on and on, and why is there so much white collar crime?  And didn't the crash happen because there wasn't enough regulation, which is needed because so many economic actors can't be trusted? 

And, if economics is about morality, it's a shifting morality.  Self-interest and selfishness have become acceptable, or even applauded in the last 30 years or so, whereas it used to be shameful to do things for personal profit, and one certainly couldn't boast about it. So clearly much that is real is not hard-wired into human behavior or societal structures.  Of course this is an acute, but largely unheeded lesson for those who yearn for genetic determinism in human behavior, built into our nature by natural selection.

To Adam Smith, how the economy works had to do with what makes humans different from dogs.  We have an inbuilt propensity to barter and exchange, he said, while dogs just fight over bones.  Ours aren't narrow self-interests, but based on our ability to reason. Of course, we know that any human universals--a target of Darwin as well as western economic theorists--are subtle and elusive, if they are even known.

But, don't even lions share?  And ants?  Ok, maybe they don't barter or exchange value for value in human terms, but they do redistribute resources. Some would argue that that's because lions and ants that share are sharing among kin, but even so, this does put a damper on the idea of human exceptionalism.

And still, why did economics get it so wrong about the crash?  Are markets emergent properties that cannot be forecast by or are chaotic relative to, imperfect measurements of their component activities?

Scientific economics
The second program explored whether or not economics is a science, whether it should be, and why as a science it still can't predict economic behavior. At the end of the 19th century, economics books begin to look like mathematics books, and economists began to talk about laws, forces, and mechanisms. If economics is a science, it's a science of observations, like astronomy, not a laboratory science.

To illustrate the law-like nature of economics, an inventor, Bill Phillips, proposed to a Cambridge economist in the 1940s that he could build a machine that could predict the economy.  "I don't understand economics," he said, "but I do understand plumbing."  So, using water flow, he built a model of the Keynesian theory of economics, the Phillips machine.  And indeed it solves Keynesian models, but does it model economies?  Here's a video of the machine that may (or may not) clarify the issues.  What such  devices (and similar kinds of computer simulations) is build in some assumptions and work out the consequences.  But if the assumptions are wrong, or the system is sensitive to conditions at any time, the predictions will follow from the assumptions but won't generate what happens in the real world---which, presumably, is what we care about.


But, just like defenders of the genetic model of disease, many economists insist that some day their models will be more precise, even though we can't estimate with precision yet, nor predict what the net effect of market forces will be.  Economics, even scientific economics, can't yet predict growth or explain or predict the business cycle.  But perhaps that's to do with the human factor.  Again, just like genetics--risk of disease, even if there's a genetic component, seems to have much to do with how we live our lives, which tends to make disease outcomes rather unpredictable.

Homo Economicus
Because economics is so poor at prediction, many economists have decided that this may be because of the human factor, and this is what was discussed in the third program.  The answer must come from understanding humans and what drives their economic behavior.  And yes, to this branch of economics, the answer could be genetic.

So, human behavior is either noble, with implications for economic behavior, or economic behavior is at the mercy of human impulse, and if we can just understand that, we'll understand economics.  Do people have sound economic judgment? Unfortunately, the possibility that we do not has been demonstrated numerous times.

According to one expert, the problem is that economics was established by apes who evolved on the savanna, in small bands of related individuals.  The psychology those apes brought to the task then simply was not up to the complexity of the system they came up with, and that explains why economies run away from us now.

But, assuming that because some human trait began when our ancestors were on the savanna means we can't adapt to current circumstances now that we're no longer on the savanna is just wrong.  We can do calculus, can't we?  We invented rockets and penicillin and nuclear bombs long after our brains evolved to be as able as they are, with presumably no idea of going to the moon.  Successful organisms are nothing if not flexible, adaptable, able to change with changing circumstances, and humans of course are no exception.  And by what reasoning do we go back only to other primates (i.e., other primates alive today)?  The thought processes didn't originate with primates.

John Maynard Keynes believed there was something beastly about our behavior.  He wrote about "animal spirits", referring to the driving force that gets us going in the economy.  Economic models can't explain what makes economies fall into recession, and then what makes them rebound so Keynes said that maybe moods are fundamental to economics.  Populations change their thinking in unpredictable ways, with unpredictable economic results.  The human factor.  This is why Game Theory and even genetics are big in economics today.

The upshot
So, basically, we don't understand what drives economies, nor what drives economic behavior.  Rather akin to how we don't understand disease causation in so many cases, nor know how to predict who will get what.  We do understand how to act, as experts, to continue to ask the public to shed resources on us because of our expertise--that's an undoubted skill.  However, as one economist pointed out in the program, if economists knew anything, planned economies should be more efficient and predictable than the free markets that triumphed over the Soviet Union in the 1980s, but they are not. Statistical approaches to genetic diseases should tell us more than they do. Perhaps, as another guest on the third episode of this series said, we'll eventually understand 9/10ths of the forces that drive markets, but we'll never understand them all.  We'd like to see more geneticists be so accepting of genetic realities.

There is a problem here that is more than the clearly empirical fact of the lack of predictive power in many areas of life, including many areas of genetics.  It is that the 'experts' have a lot of knowledge, but a limited ability to actually predict what we are paid to predict.  In that sense, why should our jobs not be taken away and given to people who actually give us what we'd like: sex, music, other entertainment, new kinds of fast foods, video games, and the like?  Since we still have jobs, clearly experts do provide something that society feels is useful!  Expertise is hard-won and clearly real in many ways.

Yet, experts are the priests of secret knowledge to whom we still turn even knowing that their knowledge, while real, is often not sufficient for accurate prediction.  Even in the age of science, we live on future promise, ignoring past records.  This is very curious!

3 comments:

Joe Terwilliger said...

http://people.few.eur.nl/thurik/Research/Articles/Genome-wide%20association%20studies.pdf

Anne Buchanan said...

Thanks, Joe. That explains it!

Ken Weiss said...

Great use of taxpayer's money! The professor welfare state at work!