Saturday, December 13, 2008


Robert Skidelsky, Keynes' biographer, writes in the Times magazine (excerpted below). Keynes had lived through the greatest of all bubbles and crashes, and saw through the convenient but deeply flawed idea of efficient markets.

As someone with a mathematical bent I was not initially drawn to Keynes' brand of economics -- my interests were in areas of modern finance like option pricing theory, volatility, stochastic models. But like Keynes I have seen a bubble up close -- first in Silicon Valley, and now, from a greater distance, the current credit crisis. What seemed to be reasonable rough approximations: efficient markets, no arbitrage conditions, stochastic processes, etc., have been revealed as terribly naive and dangerous. And so over time my views have come to resemble those described below. (See my talk on the financial crisis, and this Venn diagram.)

Although he is best known as an economist, Keynes' Treatise on Probability, written relatively early in his career, is quite good, and also stresses the idea of probability as a form of logic which goes beyond binary truth values. (See related post on E.T. Jaynes and Bayesian thinking.)

Note to commenters: I am not endorsing all "Keynsian" policy measures. I am endorsing Keynes' opinions on efficient markets, risk and the importance of psychological and sociological factors in economics -- i.e., what is discussed in the excerpt below.

NYTimes: Among the most astonishing statements to be made by any policymaker in recent years was Alan Greenspan’s admission this autumn that the regime of deregulation he oversaw as chairman of the Federal Reserve was based on a “flaw”: he had overestimated the ability of a free market to self-correct and had missed the self-destructive power of deregulated mortgage lending. The “whole intellectual edifice,” he said, “collapsed in the summer of last year.”

[Greenspan quote here.]

What was this “intellectual edifice”? As so often with policymakers, you need to tease out their beliefs from their policies. Greenspan must have believed something like the “efficient-market hypothesis,” which holds that financial markets always price assets correctly.

...By contrast, Keynes created an economics whose starting point was that not all future events could be reduced to measurable risk. There was a residue of genuine uncertainty, and this made disaster an ever-present possibility, not a once-in-a-lifetime “shock.” Investment was more an act of faith than a scientific calculation of probabilities. And in this fact lay the possibility of huge systemic mistakes.

The basic question Keynes asked was: How do rational people behave under conditions of uncertainty? The answer he gave was profound and extends far beyond economics. People fall back on “conventions,” which give them the assurance that they are doing the right thing. The chief of these are the assumptions that the future will be like the past (witness all the financial models that assumed housing prices wouldn’t fall) and that current prices correctly sum up “future prospects.” Above all, we run with the crowd. A master of aphorism, Keynes wrote that a “sound banker” is one who, “when he is ruined, is ruined in a conventional and orthodox way.” (Today, you might add a further convention — the belief that mathematics can conjure certainty out of uncertainty.)

But any view of the future based on what Keynes called “so flimsy a foundation” is liable to “sudden and violent changes” when the news changes. Investors do not process new information efficiently because they don’t know which information is relevant. Conventional behavior easily turns into herd behavior. Financial markets are punctuated by alternating currents of euphoria and panic.

Keynes’s prescriptions were guided by his conception of money, which plays a disturbing role in his economics. Most economists have seen money simply as a means of payment, an improvement on barter. Keynes emphasized its role as a “store of value.” Why, he asked, should anyone outside a lunatic asylum wish to “hold” money? The answer he gave was that “holding” money was a way of postponing transactions. The “desire to hold money as a store of wealth is a barometer of the degree of our distrust of our own calculations and conventions concerning the future. . . . The possession of actual money lulls our disquietude; and the premium we require to make us part with money is a measure of the degree of our disquietude.” The same reliance on “conventional” thinking that leads investors to spend profligately at certain times leads them to be highly cautious at others. Even a relatively weak dollar may, at moments of high uncertainty, seem more “secure” than any other asset, as we are currently seeing.

It is this flight into cash that makes interest-rate policy such an uncertain agent of recovery. If the managers of banks and companies hold pessimistic views about the future, they will raise the price they charge for “giving up liquidity,” even though the central bank might be flooding the economy with cash. That is why Keynes did not think that cutting the central bank’s interest rate would necessarily — and certainly not quickly — lower the interest rates charged on different types of loans. This was his main argument for the use of government stimulus to fight a depression. There was only one sure way to get an increase in spending in the face of an extreme private-sector reluctance to spend, and that was for the government to spend the money itself. Spend on pyramids, spend on hospitals, but spend it must.

This, in a nutshell, was Keynes’s economics. His purpose, as he saw it, was not to destroy capitalism but to save it from itself. He thought that the work of rescue had to start with economic theory itself. Now that Greenspan’s intellectual edifice has collapsed, the moment has come to build a new structure on the foundations that Keynes laid.


STS said...

Keynes was arguably the last of the great political economists. (Well, there's John Kenneth Galbraith, but why let that get in the way of an amusing generalization?) He didn't let efforts to construct useful mathematical models completely obstruct his view of the deeply political and psychological issues involved.

Krugman's "rules of research" are a good path to recovering a bit of that legacy. Don't get bogged down in spurious technicalities ("dare to be silly" as Krugman puts it) and try to relate simple models to accessible intuitions ("listen to the gentiles"). For example, people (non-economist gentiles) say markets "freeze" for a good reason. But because no economists understand phase transitions, none of them try to apply that kind of math. The tribe of the econ are very tradition-minded and averse to letting common sense interfere with ritual. The law of publish or perish has that effect on most tribes.

Anonymous said...

If the market is a non-equilibrium phenomenon as it is manifested in the debt boom and debt deflation what we are witnessing right now. Can we simply jump to the conclusion that the market is always like that?

What about the fractional reserve system, in which debt multiplies astronomically during boom periods? Is it any wonder it will collapse disorderly?

What if the current economic system is built in with distrust and abuse of the power of money supply?

Anonymous said...

Wonder if Keynes was thinking about "uncertainty" of the Knightian variety.

hcl said...

"Efficient markets" is an assumption helpful to economic modelling. It is not an economic principle.

Can society get past that, please?

Now the relevant theory's simple enough, but for sanity's sake, I went through a stack of econ books some years ago to confirm that I had not remembered incorrectly, gone nuts, whatever, and that it's just the general public mislead by the mass media.

Sure enough, the intro macro and microecon books each state that markets fail, and governments ought intervene to correct said market failures. So not only is it not theory, this ought to be widely known.

hcl said...

Keynes emphasized its role as a “store of value.” Why, he asked, should anyone outside a lunatic asylum wish to “hold” money? The answer he gave was that “holding” money was a way of postponing transactions. The “desire to hold money as a store of wealth is a barometer of the degree of our distrust of our own calculations and conventions concerning the future

If someone expects to not die in the next moment he wants cash to fund *future consumption*, to explain that more simply.

It's not necessarily about distrust or uncertainty. I know what I will eat for breakfast tomorrow, but do not want the eggs and Coca Cola now, no thanks.

Anonymous said...

higher math has no utility. according to the us dept of labor there are only 3000 mathematicians working outside academia.

as i've said before it's real purpose is similar to that of Latin for the pre-vatican II RC church. that is, to conceal the most trivial and useless in science.

this time math's inscrutability has done damage.

David said...

this time math's inscrutability has done damage.

Heehee. That made my day.

Anonymous said...

economics, Keynesian or otherwise is, has been, and cannot be any more than ideology. the number of phd economists who have read and what's more understood Capital rather than dismiss it out of hand can be counted on one hand. that economics is ideology is disguised with game theory, Ito calculus, etc.

the "rocket scientists", the mathematics priests, transsubstantiated what would have at best been original issue junk into tranches which include AAA, and people around the world believed. why? did they trust Moody's etc.? part of the reason they believed is they were told, "we've got these math guys on the tenth floor ..."

all you have to say is math and nearly everyone says "ooooo, ahhhhh, ..."

smekhovo said...

As Whitehead said about philosophy and Plato, macroeconomics is just a series of footnotes to Keynes.

Anonymous said...

Or in a word: alchemy.

hcl said...

The comments section on this site could a good anarchists' section, beginning with "school a waste of tuition", "higher math has no utility" (gee, tell that to the physicist blog owner...), and so on.

Anonymous said...

- to Anonymous who posted at 9:36 PM

When you say "Capital", are you referring to Karl Marx's "Das Kapital"?

Anonymous said...

"Capital" is my translation of "das Kapital"

you don't have to be an anarchist to find that higher math has no utility. just ask people who actually do something for a living.

Anonymous said...

The Austrian school of economics (ie. von Mises, von Hayek, etc ...) almost completely dismisses any use of mathematics, in formulating their theories. Attempting to quantitatively verify or debunk these sorts of non-mathematical economic theories becomes quite difficult.

Anonymous said...

Non-quantitative and non-falsifiable economic/political theories aren't much better than ideological tautologies.

Perhaps the intentions of theories like Marxism, the Austrian School, Keynesian economics, monetarism, etc ... are largely ideology related, and very little to do with empirical reality.

silkop said...

I second hcl's comment. Do not confuse a descriptive model (market efficiency), which has yielded quite good predictions over long periods of time, with an underlying causal model supposedly representing some "first principles". If you see that markets conform to EMH, it is not necessarily because individuals are rational optimizing agents. It may also be in spite of individuals not being rational optimizing agents.

If you want to see another good example of this same sort of confusion in physics, read up Jaynes on the normal distribution, which can work well in modelling errors not because of some inherent properties of natural processes and inherent stochastic randomness, but simply because any other distribution is even a worse candidate for accurately describing what we know (which is often: nothing).

BTW, "this time it's different" has been a universal motto in all bubbles so far. I suspect it also applies to recovery processes.

Law of the Bad Premise said...

I am not sure you were explicitly supporting deficit spending in the current environment or just reviewing what Keynes was likely to support. Economics, like many "non-hard sciences", can produce nice sounding logical arguments which are very difficult to test, let alone demonstrate as useful. Macroeconomics almost still needs to demonstrate it can be a science, as its main issues of study are so much in contention.

For example, Greg Mankiew on his blog has pointed to a variety of studies (for example: Blanchard and Perotti; Mountford and Uhlig; Alesina, Ardagna, Perotti, and Schiantarelli)critical of even the short term effects of fiscal policy of the tax cutting or spending variety.

This, of course, does not default to monetary policy as the solution one way or the other. Markets may not be efficient, whatever that really means, but markets will hack their way out of jungles eventually when given more rather than less freedom to do so. I point to the remarkable and even bizarre activity of the California housing markets as indicative. Volumes in the cities which have had the biggest bubble and the biggest crash are now experiencing record volumes. This all began happening before Paulson/Bush/Obama/Bernanke/Bair were able to get their two cents in.

Anonymous said...

Marxism isn't a theory, despite what Marxists including Marx may claim. But it is a description.

Economics seeks to explain, and thereby justify, why things are the way they are. That's fine if man has no control over how things are. Man cannot change the laws of physics, but he can change all the phenomena that economics seeks to explain.

Economics cannot be any more scientific than political science or sociology, history or anthropology. All of these can only be descriptive. That economics has pulled over the mathematical purity mantle, is a function of ideology.

silkop said...

"Economics cannot be any more scientific than political science or sociology, history or anthropology. All of these can only be descriptive."

All of these can and preferably should be predictive as well, regardless of what fake scientists from these disciplines wish to claim. Nobody forbids a social scientist or historian to make predictions and to verify them as new data comes in. To be respectable, the verification criteria must be set out up-front, of course. We can and do have predictive models (= science) concerning man-made systems even if we lack the ability to perform controlled experimentation (although it's more difficult).

Anonymous said...

right. it is more difficult, more specifically, it is impossible.

the succes of nat. sci., it's explosion, was so great even by 1800 it led Kant to ask why philosophy had not progressed at all in comparison.

Kant and social scientists make the mistake of insisting that the success of nat. sci. can be duplicated in all fields, and that nat. sci. is a pardigm for all intellectual activity. "nothing succeeds like success".

these have not understood that the great success, the explosion, of nat. sci. has nothing to do with habits of mind or its professinal method, and everything to do with the subject.

CW said...

Silkop: The previous commenter correctly observed: "Man cannot change the laws of physics, but he can change all the phenomena that economics seeks to explain." What does prediction mean under this circumstance?

Evidently, it's like predicting the outcome of football games, where you assume that the participants will continue playing by the rules of the game, to some good approximation enforced by the referees. Suppose they decide not to do so, for whatever reason, the referees are bribed, etc? If you know those things are happening you can attempt to model them as well, but this makes the analogy with physics ring a bit hollow, to say the least. The analogy with recent events in the economy is obvious.

From Keynes (quoted in a photo caption accompanying Skidelsky's piece): “Economists set themselves too easy, too useless a task if in tempestuous seasons they can only tell us that when the storm is long past the ocean is flat again.”

The unrepentant free market advocate's response to the current situation appears to be essentially this:

"Yeah, participants in a free market will f**k things up occasionally, and do it big-time. That's just tough; deal with it, and let the free market clean it up. Government should just stay the hell out of the way."

(Andrew Mellon's version: "Values will be adjusted, and enterprising people will pick up the wrecks from less competent people.")

silkop said...

Anonymous, substantiate your claims. If a phenomenon - any phenomenon - is stable enough to be observed more than once, it can be modelled and predicted. Yes, there are vastly more free parameters and less possibility for control in social sciences, psychology etc. But then, there are enough stable phenomena to become the subject of study. For example, the human ways of thinking have not changed throughout the entire history of humankind. That's more than enough time to arrive at some reasonable conclusions. Philosophy, on the other hand, is NOT science, but also irrelevant to our current discussion about economics.

CW: prediction are of course always "given that". Newton's theory of gravity makes very good predictions... up to a point.

And yes, taking note of past events is the only way of arriving at any model at all. If you do this at too coarse granularity - ignoring some relevant causal factors, such as the possibility of man-made changes in man-made systems - the future is surely going to surprise you and the past won't "repeat itself". However, this is not some grand qualitative distinction between physics and other sciences. This is just a matter of how careful you are and what kind of claims you are willing to entertain. Biological and social phenomena are, after all, the result of the very same laws that physics examines, which doesn't mean that a bottom-up approach to understanding them is reasonable at all, mainly due to the sheer computational complexity of such naive models.

We've had some marked progress in quality of predictions in a field as daunting as meteorology in the past decades (no controlled experiments!), so why not economics?

silkop said...

CW: As for the free market advocates' response to the situation, the opinions which I am aware of are not quite those that you quoted ("free markets f**k up from time to time, get over it"). These wicked people even dare say that the current state of affairs had been caused by - gasp - misguided regulatory efforts (Fannie and Freddie being big government-backed institutions and all that). "If only we have had truly free markets, this would not have happened!"

I don't know what to think about this sort of argument (though it sounds a bit like conspiracy theory along with "the FED did it on purpose"). Nevertheless, I predict that the American economy will not be shrinking for decades to come. I vaguely and non-scientifically predict a surprising recovery in scale and time beyond the expectations of most analysts, based on my perception of their credibility.

Anonymous said...


"However, this is not some grand qualitative distinction between physics and other sciences... We've had some marked progress in quality of predictions in a field as daunting as meteorology in the past decades (no controlled experiments!), so why not economics?"

i admit i'm prejudiced, but meteorolgy is much less daunting than economics. the brain trust at renaissance capital/david shaw has done very well, but their predictions/trading is the shortest of short-term and involves only securities prices not real life economics.

psychology is not quite the same as
political science, sociology, etc. but it is close. psychiatry for example is, as crazy ol' tom cruise said, a pseudoscience, but the discovery by Binet that all tests of mental ability correlate with each other is very important.

Anonymous said...

While we're on the subject of economics and human behavior, see this little profile of auction site Bernie Madoff has nothing on these guys (well, okay, except a few tens of billions USD).

Anonymous said...

the dicovery by Spearman not Binet

Blog Archive