« Retail loyalty card programs | Main | Good sense on food prices »
Bernanke's bubble laboratory
Manias can persist even though many smart people suspect a bubble, because no one of them has the firepower to successfully attack it. Only when skeptical investors act simultaneously -- a moment impossible to predict -- does the bubble pop.
...Mr. Bernanke hired finance experts who had broad interests and were eager to work with the university's deepening bench of theorists. He lured Dilip Abreu, known for work in game theory, back from Yale, to which he had earlier defected. Making a virtue of an institutional weakness, the absence of a business school, Princeton assimilated the finance scholars into the economics department and freed them to pursue research.
They are building on work done by the late Hyman Minsky, whose once-ignored ideas about investing manias are now in vogue, and the late economic historian Charles Kindleberger, whose 1978 "Manias, Panics and Crashes" is a classic. But compared with Mr. Minsky or another student of bubbles, Yale's Robert Shiller, the Princeton trio focuses less on mass psychology than on mathematical models. These they use to show how bubbles can be created even in markets that include rational, calculating investors.
Here is the full story, interesting throughout.
Posted by Tyler Cowen on May 16, 2008 at 12:14 PM in Economics | Permalink
Comments
I guess I didn't know that Hayek was an efficient market theorist. I already sent an email to the author complaining.
"Under the Hayek view, bubbles don't make sense. As soon as some group of traders irrationally pushes prices way up, more-rational traders should take advantage of the mispricing by selling -- bringing prices back down."
Posted by: John at May 16, 2008 12:20:47 PM
I have seen Minksy's ideas described as some sort of radical 'heterodox' departure from mainstream economics, but in this article bubbles don't seem to be that hard to understand (there are just some asymmetries at work). How far is Minsky from the mainstream?
Posted by: Luis Enrique at May 16, 2008 12:53:58 PM
"Making a virtue of an institutional weakness, the absence of a business school, Princeton assimilated the finance scholars into the economics department and freed them to pursue research."
I didn't realize you can't do research at a business school.
Posted by: Noah at May 16, 2008 1:03:15 PM
Bubble start when Jensen's inequality breaks down and end when the new Jensen's inequality establishes.
When the importance of the internet became know, then the old structure breaks down. We go through period of uncertainty when a new structure assumes form. During periods of reconfiguration, investors are purchasing the right to dominate the next production cycle. If the bubble investor makes the right bet, then his short term over purchase of shares will be converted into long term gains sue to monopoly effects.
So bubble investing, to a point, is rational. The investor is betting on the extra monopoly effects in value at the moment the new configuration is established.
Posted by: Matt at May 16, 2008 1:12:43 PM
I wonder if there is a very large fudge factor in reality. As a practicing investor, I buy stocks that I think are mispriced by a factor of 2 to 20 times (priced at 5 - 50% of "intrinsic value"). To the extent that I earn excess gains, I am rationalizing the market. But, in reality, unless I am getting something with at least a 50% discount, I am not buying, so even when I'm correct, my activity doesn't get the stock even close to a reasonable valuation. I have to depend on the corporation's operations or on some change in the market to attract another type of investor who will come in & bid the stock up to my price.
I think, essentially, in order to make a bet that you are right & the entire rest of the market is wrong, a very bold statement, you have to demand a very high risk premium. So, I am investing because I think the risk premium that the market is demanding on certain stocks is way out of line, but I am adding my own risk premium that is a payment for betting on my own hubris that I know better than the market.
I wonder if this sort of calculus results in contrarians working against a bubble, but never really counteracting a large amount of its error until the bubble participants themselves change direction.
Posted by: kebko at May 16, 2008 1:23:41 PM
It's nice to see Minsky getting some respect, but glossing him as a theorist of "mass psychology," as the article does, is mistaken. Minsky was a Post Keynesian with a particularly acute institutional understanding of how finance works. See Dymski and Pollin:
Dymski, G. and Pollin, R. "Hyman Minsky as Hedgehog: The Power of the Wall Street Paradigm" in _Financial Conditions and Macroeconomic Performance: Essays in Honor of Hyman P. Minsky_ edited by S. Fazzari and D.B. Papadimitriou. Armonk, NY: M.E. Sharpe, 1992, 27-62.
The Jerome Levy Institute website also has some of Minsky's papers (though its search capacities are poor) and much exegesis e.g.
http://www.levy.org/pubs/wp_452.pdf
http://www.levy.org/pubs/wp217.pdf
Posted by: Colin Danby at May 16, 2008 1:49:10 PM
"Because it's so much harder to bet on prices going down than up, the bullish investors dominate."
- is it really that hard to short stocks? To me this looks like an endogenous variable (lack of short selling) treated like an exogenous one.
Posted by: Johan Almenberg at May 16, 2008 1:53:42 PM
These researchers seem to be missing the obvious—the role of money. The article itself suggests that line of research in the following pertinent quotes:
“As a result of all that and more, the Princeton squad argues that the Fed can and should try to restrain bubbles…” (If the Feds can restrain bubbles, isn’t it reasonable that they can cause them?)
Fed Governor Frederic Mishkin: “if too-easy credit appeared to be fueling a mania…”
“When a lot of borrowed money is involved -- as it often is in a bubble -- once prices peak, the speed of their fall is intensified as investors sell urgently to pay down debt.”
Either the researchers or the writer of the article haven’t read Kindleberger or Hayek because the conclusions drawn from those works are just wrong. For example, the author writes “They are building on work done by … the late economic historian Charles Kindleberger, whose 1978 "Manias, Panics and Crashes" is a classic. But … the Princeton trio focuses less on mass psychology than on mathematical models.”
At the end of his book, Kindleberger writes that all bubbles share a common participant--credit expansion—and that credit expansion is the likely cause of bubbles.
“Under the Hayek view, bubbles don't make sense.” That’s clearly wrong. In “Pure Capital” Hayek discusses bubbles, without using the word. He states that bubbles are impossible with a fixed supply of money, but that an increase in credit that does not result from increased savings, (in other words, Fed pumping of the money supply) will cause bubbles without any doubt. What we call bubbles Hayek called an unsustainable boom that sets the stage for the next depression.
Posted by: fundamentalist at May 16, 2008 1:57:58 PM
Can you give us a page reference, fundamentalist, for Kindleberger writing that "credit expansion is the likely cause of bubbles"?
Posted by: Colin Danby at May 16, 2008 2:06:32 PM
Um, how bout the fact that one of the crack researchers investigating bubbles fell victim to one himself? Clearly his actions were rooted in behavioral psychology, not in any fancy math and models, yet the article says the Princeton crew is taking a different approach that the "soft" approach of Kindleberger (which is a must read) and Shiller (which was a must read, twice), one more focused on quantifying bubbles. I thought this article was poor and I think the idea that a bunch of eggheads whose only market experience appears to be buying at the top of a bubble will have anything revelatory to say about bubble formation is a joke.
"Mr. Hong, growing up in Sunnyvale, Calif., and teaching at Stanford, had a front-row seat to the technology boom. Recognizing a mania, he resisted investing in tech stocks himself -- until they were about to crest.
He recalls his thought process: "My sister's getting rich. My friends are getting rich....I think this is all crazy, but I feel so horrible about missing out, about being left out of the party." In 2000, "I finally caved in," he says. "I put in some money just as a hedge against other people getting richer than me and feeling better than me." But 2000, of course, was the year the bubble burst."
Posted by: joe at May 16, 2008 2:24:18 PM
Bubble need not occur with monetary expansion, but bubble can take liquidity from other sectors of the economy. The other sectors of the economy would depress their use of financing in favor of the economic sector that is bubbling.
Posted by: Matt at May 16, 2008 3:04:50 PM
fundamentalist: That’s clearly wrong. In “Pure Capital” Hayek discusses bubbles, without using the word. He states that bubbles are impossible with a fixed supply of money, but that an increase in credit that does not result from increased savings, (in other words, Fed pumping of the money supply) will cause bubbles without any doubt. What we call bubbles Hayek called an unsustainable boom that sets the stage for the next depression.
This sounds like it's headed in the right direction, but did Hayek answer the question, why do bubble investors, with too much money burning their pockets, tend to choose one kind of asset and neglect others? Thus in the 1920s and 1990s stock prices were inflated and commodity prices deflated, but in the 1970s and now the reverse is true, and a year ago housing prices were bubbled but stocks were not. Why do bubble investors focus their overabundance of currency on a particular asset rather than diversifying?
One idea that occurs to me is that investors are treating the bubbled asset as money: as an intermediate good deemed to hold its value better than the dollar. Investors settle on one kind of asset for the same reason we settle on one normal currency, i.e. to conserve mental transaction costs.
Today, one could say that commodity index ETFs are the money of choice. I'm not really satisfied with this explanation, though. Mortgage-backed securities during their heyday, and commodity ETFs today, might have been seen as great ways to balance portfolios and hedge debt assets against inflation risk, but I it's hard for me to see the same role for stocks. But perhaps this is just my prejudice in favor of hard assets speaking. In the mergers & acquisitions business, for example, it's common to talk about using stock rather than cash as "currency" to buy other companies.
Posted by: nick at May 16, 2008 3:10:38 PM
Colin, I can't give a page reference because I read a library book and no longer have it. But it was in his last chapter, I believe, where he discussed the various theories for the causes of bubbles.
Nick: "...why do bubble investors, with too much money burning their pockets, tend to choose one kind of asset and neglect others?"
No, he didn't. That's a good question and I wish he had answered it. Of course, Pure Theory was published in 1941, so there weren't as many options for speculation back then. Machlup has a good book (available at mises.org) on the way that expanded credit goes to the stock market first, but that book is from the 30's.
For the answer to that question, I think you have to go to financial experts. What little I have learned from them is that speculators tend to be contrarian. Real Estate and commodities were big in the 80's, then collapsed, so in the 90's they went for stocks. Before the stock market peaked in 2000, the big speculators had sold and were already into real estate and commodities. It's usually the small and foreign investors who get stuck with collapsing markets, although many of the big ones (Bear Sterns) got hit in the real estate collapse recently, too. The big money tends to rotate between real estate, commodities and the stock market and it switches to whatever has been out of favor for a while. Kind of like value investing. Also, it seems to like innovative ideas. In the 80's it was junk bonds. Recently it was mortgage-backed securities. Innovative financial instruments are hard to price and assess the risk, so people underestimate the risk and overestimate the profits. Speculators tend to be a very optimistic bunch.
Posted by: fundamentalist at May 16, 2008 5:28:53 PM
I'd just point out, fundamentalist, that credit expansion is an obvious concomitant of bubbles, but that's logically distinct from "likely *cause* of bubbles." I remember Kindleberger's analysis in _Manias_ as being rather different from your representation of it, but I don't have a copy at hand to check either.
I can say that Kindleberger's 1982 _A Financial History of Western Europe_, (Oxford, 2nd ed.) which is at hand, goes over bubbles and crises in chapter 15 (264-280) with explicit reference to Minsky, and endorses (page 265) the Post Keynesian teaching that the money supply is endogenous. In that case it is logically impossible for an expansion of the money supply to cause a bubble.
Posted by: Colin Danby at May 16, 2008 6:13:34 PM
I can say that Kindleberger's 1982 _A Financial History of Western Europe_, (Oxford, 2nd ed.) which is at hand, goes over bubbles and crises in chapter 15 (264-280) with explicit reference to Minsky, and endorses (page 265) the Post Keynesian teaching that the money supply is endogenous. In that case it is logically impossible for an expansion of the money supply to cause a bubble.
The money supply is endogenous only under free banking.
Under central banking, open market operations expand and contract the monetary base; these operations wouldn't exist in a free banking system.
Asset bubbles are caused by the kind of money and credit expansion that exists under central banking. Changes in money and credit in a free market are consistent with people's time preferences, the natural rate of interest, and the supply and demand for investible resources. Unsustainable, systematic increases in money and credit are impossible under such a system.
The post-Keynesians are wrong on this, as was Minsky.
Can't me as unimpressed by the Wall Street Journal article.
Posted by: Bill Stepp at May 16, 2008 7:44:24 PM
Bill Stepp,
Your latest post indicates that monopoly behavior causes bubbles. You haven't demonstrated that monopoly behavior in the financial industry in necessary, just sufficient.
Posted by: Matt at May 16, 2008 8:44:11 PM
Matt,
I'm not sure what you mean by "monopoly behavior in the financial industry." Certainly commercial banks aren't monopolies. The Fed has a monopoly of currency production in the U.S., if not of world currency production.
The ability of the Fed to cause the loan rate of interest charged by banks to sink below the natural rate of interest is a necessary condition for asset bubbles.
Under free banking, the money supply can't vary much from that which is consistent with people's time preferences. Economic shocks therefore can't emanate from the monetary sphere.
In his book _The Theory of Free Banking_, pp. 104-5, George Selgin shows how such a shock is more or less inevitable when central monetary planners manipulate interest rates. Of course, interest rate targeting has been the official policy of the Fed since the 1980s.
Selgin's book is one of the best books on money ever written.
Posted by: Bill Stepp at May 16, 2008 9:10:12 PM
"At the end of his book, Kindleberger writes that all bubbles share a common participant--credit expansion—and that credit expansion is the likely cause of bubbles."-fundamentalist
I think fundamentalist is basically right about Kindelberger, except that it wasn't at the end of the book. The end was about an international lender of last resort.
Chapter 4 ( pg. 49 in the fourth ed) begins:
Speculative manias gather speed through the expansion of money and credit or perhaps, in some cases, get started because of an initial expansion of money and credit.
So, according to Kindleberger, credit expansion is the fuel that allows the fire to spread and, in some cases, also the match that starts the fire in the first place. He goes on to give some examples and criticize monetarism (though his depiction of monetarism strikes me as caricature).
Posted by: rwe at May 16, 2008 10:29:42 PM
Thanks to rwe for an actual quotation from CPK, which you will notice is finely nuanced. I would caution that in a framework of money endogeneity, "fuel that allows the fire to spread" is not an apt metaphor.
CPK may be mistaken in his money endogeneity or his anti-monetarism. I'm not asking you all to agree with him, I'm just trying to avoid needlessly conflating him with positions he opposed.
Money endogeneity is one of those things that elicits dogmatism on both sides, so I don't want to get embroiled. I would just say that there's a range of institutional factors that may make money endogenous, and that it should also be recognized that in a particular place and time money may be endogenous for some actors and not others, a point I have written a little bit on.
Here's one recent discussion with excellent references:
http://cas.umkc.edu/econ/economics/faculty/Fontana/Winter%202005/ROPE.pdf
Posted by: Colin Danby at May 16, 2008 11:03:28 PM
"I can say that Kindleberger's 1982 _A Financial History of Western Europe_, (Oxford, 2nd ed.) which is at hand, goes over bubbles and crises in chapter 15 (264-280) with explicit reference to Minsky, and endorses (page 265) the Post Keynesian teaching that the money supply is endogenous. In that case it is logically impossible for an expansion of the money supply to cause a bubble."-Colin Danby
I think that's also mostly correct. Kindelberger criticizes both the monetarist and Keynesian views of the Depression, saying:
No quantity theory of money or autonomous shift in spending, with or without a decline in the stock market, can account for these precipitous movements (in industrial and automobile production). They require an old-fashioned theory of the instability of the credit system.
So, if I understand him correctly, he is arguing that money supply is largely endogenous, though the Federal Reserve exerts some exogenous control. But since the credit system is inherently unstable, it is prone to wide swings that central bankers will not be able to dampen completely.
He sharply criticizes Friedman for ignoring the instability of credit and argues that appropriate monetary policy can dampen the swings in credit but not eliminate them entirely:
"The weight of historical evidence strongly favors the case that while monetary policy might have moderated the booms leading to bust, it would not have eliminated them all.
Again, this is because the Fed controls the money supply only in a very narrowly defined way. But money cannot be so narrowly defined (he notes, laughing, the various measures of the money supply from M1 up). Since the Fed's control is limited, it is easy for bankers to fuel credit expansion regardless of what the Fed does. He endorses the view he ascribes to Minsky and Simons that:
...the Great Depression was caused by changes in business confidence, leading through an unstable credit system to changes in liquidity and effects on the money supply.
So, again there seems to be a lot of endogeneity involved, as Colin says. In a boom business confidence soars (sometimes due to poor monetary policy, but often not), banks start lending money more freely, this greatly expands the money supply, and the Fed has only a limited power to control that expansion. And there seems to be a kind of reciprocal causality in Kindelberger's scheme, so that the credit expansion, which is an effect of the boom, is also itself a cause of an even greater boom.
That's it as I see it, anyway. Kindelberger's ideas are interesting, but his shots at Friedman and Hayek irritate me. I don't think he does justice to their ideas.
Posted by: rwe at May 16, 2008 11:08:49 PM
I forgot to give page numbers, but all of those quotes came from Chapter 4 as well, which is not that long. So anyone should be able to find them easily enough... I agree with Colin that Kindleberger's ideas are quite nuanced. And his claims about the instability of credit have the ring of truth (especially in light of recent events in the credit markets).
Posted by: rwe at May 16, 2008 11:47:49 PM
I don't see any evidence that the monetary system is at fault, though it is affected. One would have to speculate that money is somehow a special psychological entity, I doubt it.
First, one has to explain why money flows to one sector, even one market in the sector. Even with excess liquidity there is nothing in these posts that would indicate why one sector is favored. And how does the bubble flow?
In fact, efficient market theory would say that no single sector should be subject to bubbles at any time, assuming that transaction costs are proportional across.
We can prove that monopoly behavior would be involved if we can show monopoly price changes coincide with bubbles. When a monopoly large enough changes prices, as a result of reconfiguration, then one would expect a period when resources are volatile.
The most plausible theory is large cyclic behavior in monopolies cause some of us to plan production in fixed periods. During the end of a particular period there is a tendency to release resources, in time, so to speak, so the change we want occurs with efficiency.
In Silicon Valley we called it the "next big thing", and all the venture capital managers got their money in fixed time periods, and they purposefully synchronized.
The deal is, we change state. We have a period of rearrangement when money is very well spent to get through that phase, for in that phase we are very inefficient.
We break our Jensen's inequalities, releasing a few tier of resources at once. There is an energy needed to break the aggregation in production systems.
The resources go through a Ramsey re-allignment, and begins to compute the trajectory, with plans purposefully synchronized.
So, the bubble, if done efficiently serves a purpose. Flood the chaos with money to cover the released resources, let everyone scramble to re-allign to the most dominate mode. Rebuild the Ramsey graph. Resest the Kalman filter to match the newly designed production function. Then, start diagonalizing that production function using exchanges.
How often do we do this? Over long time periods, we do this whenever the noise of estimating the next step in current production is equal to the noise in estimating the cost of reconfiguration.
Because we regraph along the current, longest modes in the economy; large monopolies would have an extraordinary effect. If ee timed our economic moves according the the lock step of monopoly entitlements, very dangerous.
Central Banks would be stuck if they could not simulate a competitive monetary system.
But, this theory explains why bubbles migrate. They migrate for the same reason matrix diagonalizing algorithms cycle when too few eigenvectors are specified, naive algorithms cycle the excess innovation noise through the diagonal, looking for a fit.
But, it is quantum, we time this stuff. We time this stuff for 50 70 years, seriously, look at public finance across the globe. Look how long those cycles are, many preplanned.
Read Will Smith, Berkeley for some good models
Posted by: Matt at May 17, 2008 12:30:24 AM
The scholars quoted in the original article sound like they are just engaging in a fancy form of technical analysis, which purports to capture in mathematics the psychology of the crowd, but in almost all cases turns out to be worthless numerology.
The fancy math doesn't explain bubbles any better than Schiller's simple driving analogy: there is a big event at an obscure location (let's call it Burning Man, out in the middle of the desert somewhere). People driving to the event are very uncertain of the directions. At a certain intersection, on average 60% of them correctly believe they must turn right and 40% incorrectly believe they must turn left. But they can usually reduce their uncertainty by observing the behavior of others, who are somewhat more likely to be correct than wrong, and altering their Bayesian priors accordingly. Normally this works, but on rare occasions it goes wrong: for example, if the first three cars happen to turn left, the fourth, who had believed with 60% confidence that right was the proper direction, will rationally change his mind and go left. Thus a string of bad luck can make all the cars start going off in the wrong direction, except for those handful of drivers that are strongly confident in their knowledge that one should go right.
Where this analogy goes off the rails as public policy analysis is with the tacit hubris that certain academics from sufficiently elite schools are flying above the whole event in a helicopter and can direct traffic, if only their mathematical analysis is fancy enough. Rather academics and policy makers are in the traffic themselves, generally seeing information biased in ways similar to or even more extreme than the information investors see and act on. In many cases mispriced markets create arbitrage opportunities for truly knowledgeable investors, but analogs to such arbitrage opportunities, i.e. the ability to be rewarded for correctly actual misinformation, are much less prevalent in academic and political policy circles. We should thus expect political policy to be much more prone to biased political fads and herd-following than markets are. Both markets and governments may often take wrong directions, but for politics the inability to correct wrong directions may be endemic. Markets tend to correct themselves, usually in the short run and practically always in the long run, depending on the costs of arbitrage, but there is often no easy way to recognize or correct a political bubble.
Thus, applying the same rational uncertainty assumptions to politics as we do to markets, if we give political decisionmakers the power to "pop" bubbles they think they recognize, they will probably tend to make genuine market bubbles worse, will prevent markets from sending genuine supply and demand signals, and will introduce other extra transaction costs.
Posted by: nick at May 17, 2008 3:15:53 PM
I think there is some confusion of terminology in the discussion with the distinguishing between the money supply and credit expansion. Some people see the money supply as currency only, but it should include currency and demand deposits, at least. So when credit expands, the money supply expands also. I realize there is a lot of disagreement over what constitutes money, but currency and demand deposits should be the minimum considered.
If you consider the money supply to be endogenous, then all of us have no choice but to become amateur psychologists, like the authors in the article. And when we do that, we'll have 10,000 explanations for bubbles, because when you go delving into the motivations of other people, you're limited only by your imagination, and no one can prove you wrong.
What's really striking to me about the article and the attempts to explain bubbles is that more than two centuries have past since Adam Smith and mainstream economics still has no clue as to the causes of business cycles and bubbles. Yet when offered an explanation by Austrian economists that logic and history support, mainstream economists prefer their current mass confusion.
Posted by: fundamentalist at May 17, 2008 7:14:28 PM
More clearly, Keynesian and mainstream econ remove the problem of bubbles from the field and economics and deposit it in psychiatry. Austrian econ keeps it squarely within the field of economics. The mainstream approach to bubbles is abandoment of the question to another field of study.
Posted by: fundamentalist at May 18, 2008 2:33:32 PM
