Wednesday, September 29, 2010

Why the free market is like quantum mechanics

If we were all omniscient and had infinitely fast and perfect computers, maybe we could use quantum mechanics to explain chemistry and biology. In reality, no amount of quantum mechanics can completely explain the chemistry that goes into treating a disease with a drug, baking a cake or making a baby. Why? Because the system is too complicated, and we don't even know all the factors that make it tick yet.

Now imagine someone who has started out with the honest and admirable goal of trying to apply quantum mechanics to understand the behavior of a biological system like a large protein. He knows for a fact that quantum mechanics can account for (a far better word than explain) all of chemistry- the great physicist Paul Dirac himself said that. He has complete confidence that quantum mechanics is really the best way to get the most accurate estimates of various properties for his system.

But as our brave protagonist actually starts working out the equations, he starts struggling. After all, the Schrodinger equation can be solved exactly only for the hydrogen atom. We are dealing with a system that’s infinitely more complex. The complexity forces our embattled savant to make cruel approximations at every stage. At some point, not only is he forced to commit the blasphemy of using classical mechanics for simulating his system, but he has to actually stoop to using empirical data for parameterizing many of his models. At one point he finds himself fighting against the Uncertainty Principle itself!

In the end our hero is chagrined. He started out with the lofty dream of using quantum mechanics to capture the essence of his beloved protein. He ended instead with a set of approximations, parameters from experiments, and classical mechanics-derived quantities just for explaining his system. Prediction was not even an option at this point.

But his colleagues were delighted. This patchwork model actually gave fairly useful answers. Like most models in chemistry, it had some explanatory and predictive value. Even though the model was imperfect and they did not completely understand why it worked, it worked well enough for practical purposes. But this modest degree of success held no sway for our bright young scientist. He stubbornly insisted that if, just if, we had a perfectly fast computer with unlimited accuracy and an infinite amount of time, quantum mechanics indeed would have been spectacularly successful at predicting every property of this system with one hundred percent accuracy. Maybe next time he should just wait until he gets a perfectly accurate computer and has an infinite amount of time.

The preceding parable was narrated to describe what I think is a rather unwarranted swathe of criticism coming from libertarians about the financial crisis during the last few years. For instance see Amit Varma's criticism here which sums up many of the major points. The reasons for the financial crisis are many, probably more complex than quantum mechanics, and society will surely keep on debating them for years. But one of the most common reasons cited by libertarians (usually in the form of a complaint) for the failure of the economy is that we should not blame the free market for what happened because we never got a chance to actually have a free market. If only we got a chance to have a perfect free market, things would be lovely.

Notwithstanding the fact that this argument inches uncomfortably close to arguments made by the most vocal proponents of socialism in the twentieth century (“There was nothing wrong with the system per se, only with its implementation”), I think it’s a little nutty. Maybe a perfect free market wouldn’t have led to the crisis, but that’s like our young chemist saying that infinitely accurate computers and quantum mechanics would not have led to the kind of imperfect models that we get. The problem is really that there are so many practical obstacles in the application of quantum mechanics to a real-life chemical system, that we are simply forced to abandon the dream of using it for explaining such systems. Unless we come up with a practical prescription for how quantum mechanics is going to address these real-life obstacles without making approximations, it seems futile to argue that it can really take us to heaven.

To me it seems that libertarians are ignoring similar obstacles in the way of implementing a perfect free market. What are these obstacles? Most of them are well known. There’s imperfect competition because of the existence of inherent inequalities, leading to monopolies. There’s all that special interest lobbying, encouraged by politicians, which discourages true competition and allows monopolies to get a head start. There’s information asymmetry, which simply keeps people from knowing all the facts.

But all these problems are really part of a great problem- human nature itself. All the obstacles described above are basically the consequence of ingrained, rather unseemly human qualities- greed, the lust for power, the temptation to deceive, and a relentless focus on short term goals at long term expense. I don’t see these qualities disappearing from our noble race anytime soon.

Now sure, I think we can completely agree that the free market was invented to curb some of the worst manifestations of these qualities, and it has worked remarkably well in this regard. Remarkably well, but not perfectly so. Maybe libertarians need to understand that the last vestiges of the dark side of humanity cannot be done away with, since they are an indelible part of what makes us human. So unless they come up with practical ways in which they can surmount these obstacles, in which they can solve the problem of human nature itself- a difficult goal to put it mildly- it’s rather futile to keep on chanting that all our problems would be solved if only we could somehow make these inherently human qualities disappear.

The final argument that libertarians usually make is; just because there are obstacles in the way of a goal (the perfect free market) that may seem even insurmountable, that does not mean we should not keep on striving towards the goal. I think that’s perfectly laudable. But the problem is, unless you come up with a practical solution for all the problems that you face on the way, your goal is just going to remain an abstract and unworkable ideal, not exactly the kind of solution that's desirable in the practical fields of politics and economics. More importantly, all this striving towards the goal may create problems of its own (the science analogy would be unimaginably expensive calculations, scientists laid-off because of the lack of results, overheating of the computers leading to fires etc.). We have all seen these problems. There’s the well-known problem of externalities, there’s the problem of unregulated firms getting ‘too big to fail’, there’s the problem of growing income inequality. Surely we have to admit that these are real problems too.

So what should libertarians do? Well, didn’t our intrepid quantum mechanic grudgingly accept the intervention of approximations and parametrizations? These seemed ugly, but he had no option but to use them, since quantum mechanics simply could not solve all the obstacles in his way. Similarly, perhaps free marketers could realize that at least in some cases, government intervention, no matter how ugly it may seem, may be the only way to reach a workable goal. Sure, it may not be the best of all goods, but it could be the least of all evils. What would have happened if our bright young scientist had kept on insisting that he wouldn’t budge an inch if he is forced to use anything other than quantum mechanics? He would have ended up with nothing.

And in economics even more than in chemistry, a model that partly works is better than a model that does not exist. “Sometimes it’s not enough to do our best; we need to do what’s necessary” (W. Churchill)...

Labels: , ,

Friday, September 10, 2010

The price of teaching

I have always been wary of evaluating faculty members based on the amount of money they bring in. One of the casualties of American academic science in the latter half of the twentieth century was that it commodified research, and money became a much bigger part of the equation. Research groups started to bear a striking resemblance to corporate outfits. Undoubtedly there were benefits to this practice since it brought in valuable funding, but it also tended to put a price on the generation of knowledge, which seems inherently wrong.

Now it seems that Texas A & M is thinking of turning this kind of valuation into official policy. As Chemical & Engineering News reports, TAMU is planning to rate its faculty based on their "net worth". This would be calculated based on the faculty member's salary, the funding that he or she can generate, and teaching (how on earth are they going to financially evaluate that?)

Sorry, but I think this is hogwash, and others seem to agree with me. The "worth" of faculty members goes way beyond the funding they can procure. There may be professors who bring in modest amounts of money but who inspire generations of students through their teaching, who significantly contribute to the public perception of science through science communication, and who generally contribute to the academic environment in a department simply through their passion and strong advocacy of science. Even from the point of view of research, there are faculty members who publish relatively less, do research on the cheap, and yet steer their respective fields in new directions simply by generating interesting ideas. Very few of these qualities lend themselves to spreadsheet analysis.

In fact, I will go a step further. If a faculty member does little more than inspire generations of students to pursue careers in science research, education and policy, there is no metric that can financially measure the worth of such contributions. Simply put, such contributions may well be priceless. That should easily satisfy Texas A & M's criteria for high-value "assets".

Labels: , ,

Wednesday, September 08, 2010

Louisa Gilder and Robert Oppenheimer

Louisa Gilder's book "The Age of Entanglement" is a rather unique and thoroughly engrossing book which tells the story of quantum mechanics and especially the bizarre quantum phenomenon called entanglement through a unique device- recreations of conversations between famous physicists. Although Gilder does take considerable liberty in fictionalizing the conversations, they are based on real events and for the most part the device works.

Gilder's research seems quite exhaustive and well-referenced, which was why the following observation jumped out of the pages and bothered me even more.

On pg. 189, Gilder describes a paragraph from a very controversial and largely discredited book by Jerrold and Leona Schecter. The book which created a furor extensively quotes a Soviet KGB agent named Pavel Sudoplatov who claimed that, among others, Niels Bohr and Robert Oppenheimer were working for the Soviet Union and that Oppenheimer knew that Klaus Fuchs was a Soviet spy (who knew!). No evidence for these fantastic allegations has ever turned up. In spite of this, Gilder refers to the book and essentially quotes a Soviet handler named Merkulov who says that a KGB agent in California named Grigory Kheifets thought that Oppenheimer was willing to transmit secret information to the Soviets. Gilder says nothing more after this and moves on to a different topic.

Now take a look at the footnotes on pg. 190-191 of Kai Bird and Martin Sherwin's authoritative biography of Oppenheimer. B & S also quote exactly the same paragraph, but then emphatically add how there is not a shred of evidence to support what was said and how the whole thing was probably fabricated by Merkulov to save Kheifets's life (since Kheifets had otherwise turned up empty-handed on potential recruits).

What is troubling is that Gilder quotes the paragraph and simply ends it there, leaving the question of Oppenheimer's loyalty dangling and tantalizingly open-ended. She does not quote the clear conclusion drawn by B & S that there is no evidence to support this insinuation. She also must surely be aware of several other works on Oppenheimer and the Manhattan Project, none of which give the slightest credence to such allegations.

You would expect more from an otherwise meticulous author like Gilder. I have no idea why she gives credence to the canard about Oppenheimer. But in an interview with her which I saw, she said that she was first fascinated by Oppenheimer (as most people were and still are) but was then repulsed by his treatment of his student David Bohm who dominates the second half of her book. Bohm was a great physicist and philosopher (his still-in-print textbook on quantum theory is unmatched for its logical and clear exposition), a dedicated left-wing thinker who was Oppenheimer's student at Berkeley in the 1930s. After the War, he was suspected of being a communist and stripped of his faculty position at Princeton which was then very much an establishment institution. After this unfortunate incident, Bohm lived a peripatetic life in Brazil and Israel before settling down at Birkbeck College in England. Oppenheimer essentially distanced himself from Bohm after the war, had no trouble detailing Bohm's left-wing associations to security agents and generally did not try to save Bohm from McCarthy's onslaught.

This is well-known; Robert Oppenheimer was a complex and flawed character. But did Gilder's personal dislike of Oppenheimer in the context of Bohm color her attitude toward him and cause her to casually toss out a tantalizing allegation which she must have known is not substantiated? I sure hope not.

Labels: ,