Financial Markets (2008)

ECON 252 (2008) - Lecture 25 - Okun Lecture: Learning from and Responding to Financial Crisis, Part I (Guest Lecture by Lawrence Summers)

Chapter 1. A Profile of Lawrence Summers, Memories of Art Okun [00:00:00]

Professor Robert Shiller: I’m Bob Shiller. I have the pleasure of introducing this Okun lecturer, Lawrence Summers. I wanted to say first that the Okun Lecture Series was funded by an anonymous donor in honor of Arthur Okun, who was Professor of Economics here at Yale, who later became Chairman of the Council of Economic Advisors, who founded with William Brainard the Brookings Papers on Economic Activity and was widely admired both here at Yale and around the world as an economist that had a lot of impact on policy — an economist with a purpose. It’s altogether fitting that we have Lawrence Summers here as our Okun lecturer for this year.

Larry Summers — I’ll go through his career briefly — he started out at age sixteen at MIT as a physics major. Fortunately for us, he switched into economics at a very — as a teenager. He got his PhD at Harvard. He got tenure at Harvard at one of the — he’s one of the earliest, at age twenty-eight. In 1993, he won the John Bates Clark Medal for the best economist under the age of forty. He joined the staff of the Council of Economic Advisors. He was Undersecretary of Treasury for International Affairs. He was Chief Economist at The World Bank. He was the U.S. Secretary of Treasury. He was President of Harvard University and now he is a University Professor at Harvard and also a Director at D.E. Shaw. His research has done truly stunning — I’ve always been a huge admirer of it — and it covers many fields — public finance, labor, macroeconomics, and finance, in particular.

I thought I would just conclude my introduction with a couple of famous Larry Summers quotes. These quotes are advice to graduate students but they might as well be advice to any researcher. Larry is quoted as saying, “You might as well work on big problems because it takes just as much time to write a paper on a little problem as on a big problem.” Another bit of advice to students: “Do your research on the economy not on the economic literature.” I think that’s a bit of wise advice. But, I think that — to me, that says something about the essence of his research style. He’s been extremely potent in his work on very important results that kind of shape our view of the world and I think that it’s this research style that is part of the reason why he’s been tapped for such important positions. I’m pleased to introduce Larry Summers.

Professor Larry Summers: Thank you, Bob, very much for that overly generous introduction. At least one thing you said wasn’t quite right. It may or may not have been economics’ good luck when I left physics, but it was surely the good luck of the physics profession when I left physics. I appreciated your generosity in reflecting on the different sectors in which I’ve had a chance to work. When I first moved from the University to Washington, people asked me, well what’s different about being a senior Treasury official from being a professor at Harvard? I realized the answer to that one was easy. As a professor at Harvard, the single worst thing you could do was to sign your name to something you had not written yourself. On the other hand, in the Treasury Department, it was a mark of effectiveness to do so as frequently as possible.

Then, I got back to Harvard and had the honor of serving as its president and people asked me, what was different about the University from Washington? In those first months I gave an answer that in retrospect seems slightly breathtaking in its naiveté. I said, Washington is so political; there’s so much opposition; University is very different from that; we’re all on the same side.

It is a real honor to be here at Yale and it is a particular honor to deliver the Arthur Okun lecture. I am told that I first met Art Okun when I was one week old, as he and my father were close friends during the time when they were assistant professors at Yale together. In fact, I grew up until a certain age calling him Uncle Art. I don’t have many vivid memories, but I have a quite vivid memory and I was able to go back and place the year. It was 1969 and Art delivered a set of lectures, probably a set of lectures like this one, at the University of Pennsylvania and he stayed at our house. He stayed up telling my parents a whole set of stories about life as an economist with Lyndon Johnson. Some of the stories are a little too raunchy to repeat right here, but it was the first inkling I had that this academic research stuff, like my parents did, really could have a very direct and real impact on the policy arena.

I remember, especially, somehow — I don’t know what determines what makes an impression on you when you’re fifteen, fourteen — but my father telling me this story of how there was a — and I have no idea whether this story is actually true. But, the story he told was that there was a publication that the Department of Commerce published called BCD, which stood for Business Cycle Digest, which compiled all the facts about the business cycle and whether the economy was up or whether the economy was down. Because of all this stuff that economists had figured out, like Art, we now had these practical policies that we could use so that when there was a recession, we’d stop the recession from happening and when there was an excessive boom, we’d smooth off the boom. So, we didn’t necessarily have to have business cycles anymore.

We liked calling the publication BCD. So, Art had had the clever idea of renaming the publication to The Business Conditions Digest, to celebrate the success of Keynesian economics. Time went on and the first great success, if there ever were any great successes of my career as a graduate student, came when Art Okun and George Perry invited Kim Clark and I to give what was the first significant paper that I had written as a graduate student, on labor market dynamics and unemployment, at the Brookings Papers on Economic Activity Conference. It was the thrill of our young lives and it was also a vastly, vastly better paper in the form in which it was published than it was in the form in which it was originally written, thanks to Art.

I have a final memory that pertains to Art that also says something about Art. I remember sometime in the early 1980s sitting at the lunch table in the MIT faculty club in a way that some of you who know me won’t find surprising. I was declaiming on various subjects with a certain lack of reticence that perhaps was not commensurate with my lack of knowledge. I asserted some proposition and Paul Samuelson turned to me and he said, Larry I just wrote a eulogy for Art Okun, who died recently, and I said that in all the years that I had known Art Okun, Art Okun had never said a stupid thing. Well Larry, now I won’t be able to say that about you.

One of the things that gives me particular pleasure in, along with Greg Mankiw, taking over the editorship of the Brookings Papers on Economic Activity from Bill Brainard and George Perry, who had done such a stellar job for the last twenty-five years, is the knowledge that it was a project that had been initiated by Art. It was a project that had been initiated by Art out of a conviction that I have tried to live by in my career and that I hope in some small way to illustrate with my remarks here today — that serious economic thinking can help one understand the world much better; that with the aid of the metaphors and the models and the measurements that come from serious economic thinking, one can direct economic policies in better and in more sensible ways. As a consequence, the lives of very large numbers of people, most of whom who will never hear of economics and certainly will not study its details, will in some way be improved by having higher incomes, by facing less risk, by being more secure as they go through their livelihoods.

Chapter 2. Okun’s Concerns on Stable Growth, Inflation, and Cyclical Fluctuations [00:12:48]

It’s in that spirit that I want to reflect today on some aspects of macroeconomic thinking and experience and then some aspects of financial thinking and experience and how they bear on financial crisis. Then tomorrow, on the particular financial crisis that we face right now, what has been done, what may happen, what should be done, and what can be done over the longer term to make the return of such a crisis more important — less likely.

Art was preoccupied through most of his career with the central macroeconomic issue of assuring stable growth. To use language that’s perhaps more congenial to older than younger ears — avoiding underemployment and output gaps. It’s revealing that the journal he founded was called the Brookings Papers on Economic Activity, suggesting a concern with the cyclical performance of the economy. He surely, as much as anyone, would have subscribed to Jim Tobin’s famous dictum — that it takes a heap of Harberger triangles to fill an Okun gap — or that the macroeconomic issues of recession and depression are terribly, terribly important relative to the microeconomic efficiencies that are a preoccupation of a very large number of economists.

This idea, this first idea that Okun gaps exist, and are large, and are very important relative to microeconomic inefficiencies is one that is increasingly unfashionable today. Discussing them within the context of what has been called, although I suppose it can’t be called that anymore, the new classical macroeconomics is almost impossible; there’s no role for any kind of gap. It’s much like discussing epicycles with a modern astronomer. Even within a Keynesian tradition, the accepted versions of the Keynesian model greatly circumscribe the scope for what was Art’s major concern: underemployment and the output gap.

If one thinks about the centerpiece of the supply side of any macroeconomic model, almost any macroeconomic model post 1980, in some form or other it has the implication that the change in inflation is in some way dependent on the output gap. Perhaps it’s the change in inflation; perhaps even more problematically, it’s the unexpected component of inflation. Perhaps the causation runs the other way and the output gap is related to the — is caused by the change in inflation or is caused by the level of unexpected inflation. Any equation, any theory of output that has that property has a simple consequence if you aggregate it over a long period of time. Namely, that if you look at the difference between inflation at the end of the period and inflation at the beginning of the period, or equivalently, if you look at the average level of unexpected inflation, which has to be zero, it will relate to the size of the output gap. Therefore, if one thinks of a steady state as being a state in which unexpected inflation averages zero or a state in which the accelerant, the change in inflation, averages zero, you know what the output gap has got to be; it has got to average zero.

This doesn’t mean there’s no role for cyclical policy, but it has the very strong implication that what you gain in the booms you lose in the recessions. If you have deeper recessions, you will be able to have larger booms; that if you have larger booms, you will be able to have — you will be forced to have deeper recessions. In a sense, if one accepts this idea, the subject of cyclical fluctuations in macroeconomics becomes an order of magnitude less important than it was if one maintained some older idea in which it’s possible to have output gaps that are not offset by subsequent gains in output. To be sure, much has been made of the point at certain moments; there may be some impact of non-linearity in the relationship between the gap and the change in inflation or unexpected inflation. If this relationship is non-linear, then it’s possible that by running a smoother economy, one can run a somewhat higher-pressure economy.

In a world where econometricians struggle to estimate any kind of Phillips Curve, the idea that one can measure with any precision a non-linearity of this kind is, at the very least, problematic. Ironically, the Keynesian model as it came to evolve — the Keynesian model that — and the Keynesian discussion that had as its center piece the Phillips Curve in some form, in almost all of the forms in which the Phillips Curve came, was quite skeptical of the role of macroeconomic policy in affecting average levels of output through time. Very little room for Art Okun’s dedication to closing output gaps or maintaining a higher pressure economy as a way of spurring economic growth or as a way of promoting a more equal income distribution.

Art’s last–the last decade of Art’s professional work was directed at questions that in one way or another relate to these equations. His final book was entitled Prices and Quantities. His dominant professional preoccupation was, in his last years, with how the rate of inflation could be brought down with a minimum level of the output gap with, as he would have put it, a minimum sacrifice ratio. It’s not hard to understand why these questions relating to inflation and output were preoccupations of Art’s during his professional lifetime. It’s interesting to think about U.S. business cycle history. When people like me gave talks about the economic outlook to non-academic audiences fifteen or twenty years ago, it was considered a cleverness to observe that American economic expansions do not die of old age; they’re murdered by the Federal Reserve with inflation control as the motive.

If one thought about — thinks about — a universe of experience, the three Eisenhower recessions that were formative for Art, although I believe the second of them coincided with my birth — the 1966-1967 economic downturn — more conspicuously the major downturns in ‘74-‘75, in the wake of the oil supply shock, and in 1980-‘82, the Volcker disinflation; these were most of the experience we had with the business cycle in a substantial sloth of economic history. The economy expanded in one way or another, inflation got out of control, people got nervous about inflation, the Fed hit the brakes, interest rates spiked, the yield curve inverted, the economy slowed.

To be sure, there were great controversies about how — what exactly the mechanism was. What was the role of credit rationing in the housing sector versus higher interest rates? To what extent did the transmission mechanism take place through Tobin’s Q. If one asked, what was the cause of downturns, the cause was monetary. The accounts of business cycle historians who tracked events were couched in those terms. The enormously influential monetary history of the United States emphasized the role of monetary contractions and of Federal Reserve policies. The academic issues of the day, which focused critically on the questions of price rigidity, therefore led to a natural focus as the canonical experiment of — suppose you reduce M, the money stock, what happens to the level of output in the short-run and why does that happen?

The prevailing view about business cycles was that they were determined by monetary policy and what led to contractionary monetary policy was either a mistake by the Fed or a necessary action by the Fed as inflation accelerated. It is interesting to contemplate that if one looks before or after what one might think of as the Okun period and tries to give a simple account of what was behind the business cycles that took place, one offers a rather different kind of story. 1907 is linked to a financial panic; it couldn’t have been caused by the Feds, since the Fed didn’t exist yet. In 1929, the actions of the Federal Reserve in 1928 or 1929 may or may not have been a crucial causative factor, but there’s a reason they talk about the stock market crash as a prelude to The Depression. The dominant influences were endogenous to the workings of the financial system.

If one thinks about the three most recent instances of cyclical fluctuation in the United States, the 1990, 1992 case can be debated. There was an oil shock; there had been some tightening of monetary policy, but I think most observers would assign dominant influence to the problems in the financial system associated with the S&L’s and the overextension of banks in real estate lending. There’s no question that almost any close watcher, if asked to say, what was the 2001 recession about, they would say bursting of the technology bubble. I’ll have more to say about the 2008 recession a little later and tomorrow, including some speculations on the probability that there is one.

No one would say of the current fluctuation that it reflected the Federal Reserve’s excessive or necessary effort to curb inflation in product price markets. Rather, they would attribute it in one way or another to something endogenous to the financial system. The debates would turn on whether it was best thought of in terms of a housing price bubble; whether it was best thought of as a matter having to do with the excessive extension of subprime credit; whether it was a reflection of a generalized excessive complacency in credit markets that led to the extension of financial institutions. All of those questions could be debated, but no serious dispute about the proposition that it came out of the workings of the financial system.

Chapter 3. The Interconnectedness of Modern Financial Crises Worldwide [00:29:05]

With this history before us, in thinking about macroeconomic policy and in thinking about cyclical fluctuations, seems to me appropriate that we at least consider redirecting some macroeconomic attention from traditional preoccupations with the price setting process — the Phillips Curve — and disinflation to more traditional Keynesian preoccupations with the workings of the financial system with animal spirits — with, if you like, shocks to the IS Curve that come out of the workings of the financial system. This view, I would suggest, is certainly confirmed by less detailed reflection on international experience.

One thinks about the two most dramatic OECD macroeconomic events of the last twenty years. They are almost always — excuse me, President Zedillo, [former Mexican President Ernesto Zedillo, who was among the people seated in the classroom for Summers’ lecture.] permanent OECD member events, so I exclude the drama of which you were a — you were so central, so central a part. They were, almost certainly, the Japanese financial crisis followed by depression in the 1990s and the dramatic declines in output associated with the failure of the banking system, to the point where it needed to be nationalized in the Nordic countries in the early 1990s. Indeed, it is interesting to note that in recent years financial crises have become pervasive. One always remembers the comments one makes about economic fluctuations that prove to be prescient and forgets the ones that prove to have been completely wrong. I am no exception.

I will recall that I was fond of observing in 2005 and 2006 that if one looked at the last twenty years, there had been six or seven — six major financial crises: the 1987 stock market crash; the 1990 savings and loan real estate crash; the 1995 Mexican crash — some would argue that if this is a U.S. list, that Mexico doesn’t belong and one could have that argument — the 1997 Asian financial crisis; the events associated with LTCM and Russia that President Clinton and Secretary Rubin, at the time, labeled the most serious financial crisis since the Second World War; and the 2001 financial crisis associated with the bursting of the tech bubble, problems in the high yield sector, the subsequent problems in capital markets — exemplified by what happened at Enron. It is interesting to ask the question, whether there is some relation between all these events. Because what I’ve suggested so far is that if one distinguishes the last generation from the preceding generation — in the preceding generation, financial crises were relatively rare; business cycles were caused by the Federal Reserve eliminating — Federal Reserve’s concern about inflation. For the last generation, business cycles have been associated with financial crisis and the Federal Reserve has not caused downturns through its concern with inflation.

How are these events linked? I’m not entirely sure. Sitting in the Brookings Papers meetings for all these years, I’ve learned how Bill Nordhaus is hard to satisfy. Sometimes people have t-statistics of three and Bill pronounces them to be quote-unquote Darwinian, reflecting the data mining of the author. On other occasions, they have only a few observations, in which case they’re deemed to have committed one-observation econometrics. I’m going to commit about four-observation econometrics and offer the hypothesis that the right way to understand this pattern is that we achieved — that Art’s generation and then Paul Volcker — achieved a hugely important thing. They achieved a measure of credibility and commitment to non-inflationary monetary policies that enabled inflation expectations to be kept in check. That, just as success in curing infectious disease leads there to be far more cancer than there was before infectious disease was brought in check, so too the achievement of disinflation leads to an economy in which expansions are enabled to continue until they lead to the complacency and confidence associated with financial crisis.

Perhaps we will have 6% inflation two years from now and the Fed will cause — Fed will find it necessary to engineer a recession and this theory will fall completely by the wayside. Perhaps there is an alternative, more compelling explanation for this pattern. But for now, I would adduce the hypothesis that the increased incidence of financial crisis and the closer relationship between financial crisis and economic downturns is a by-product of our success in achieving low and credible inflation and the associated great moderation. Whether or not we can accept that idea as to why financial crisis seems to have been the cause of previous recessions where it was not previously, the empirical observation stands that financial crises are frequent and that not infrequently, when there is a financial crisis, it leads to some significant disruption in real economic activity. It follows that the causes of financial crisis and the policy response to financial crisis need to occupy a larger part of our consideration of macroeconomics than they have traditionally.

I should suggest also that this is — this conclusion is reinforced by the observation that, whereas disinflation recession lends itself very much to the unfortunate conservation argument I presented at the beginning that suggested that you couldn’t really alter the output gap over long periods of time if you did not want to accept changes in inflation. It is easy to understand financial crisis and the associated downturns in output as a reduction in aggregate supply. And, therefore, [it is] highly plausible to believe that the better management of financial crisis or the prevention of financial crisis will lead not just to a smoother path of output, but a path that, by avoiding downturns, has a higher average over long periods of time. And, therefore, [it] is more successful in achieving the goal that Art Okun and Jim Tobin and Yale, for many, many years, set for macroeconomic policy of raising the average level of output.

Chapter 4. The Bank Run Metaphor in Non-Bank Financial Crises [00:40:05]

Now, discussions of financial crisis tend to have the character of accident investigations. There are many different suspects; great numbers of details are mustered to tell the story. There are always particular words associated with particular financial crises. One talks about portfolio insurance in the context of the 1987 crash. One talks about Tesobonos in the context of the Mexican experience. One talks about bubble psychology in the context of the NASDAQ experience. One uses the — one speaks about the viscidities of the separation between the originators and the owners of mortgages in talking about the current subprime mortgage crisis. It all brings to mind Tolstoy’s observation that every happy family is the same and every unhappy family is miserable in its own way. It is all correct — not all of it is correct — but in each of the crises there are important elements of institutional detail of particular practice that explain how the story played out in the way that it did.

I want to suggest that there is some common structure that lies behind the various phenomena that are associated with financial crisis. The understanding of which can at least help us to think about the mechanism of crisis and perhaps the appropriate policy responses. For thinking about financial crises, economics has a classic metaphor that captures many elements of the problem. It’s a set of ideas that go back at least to Bagehot [British businessman and writer Walter Bagehot, 1826-1877] economic history and are known to current economic theorists as Diamond-Dybvig phenomena. [Douglas Diamond and Philip Dybvig, “Bank Runs, Deposit Insurance and Liquidity,” Journal of Political Economy, 91(3), 1983]The idea is simple, powerful, and incomplete. Imagine a bank; the bank takes deposits; the deposits can be withdrawn at any moment or at least can be withdrawn at high frequency. It takes the deposits and it lends the money for projects. The projects have a higher rate of return than the deposits do, but the projects take some significant time to complete and there is no market for a half-finished project. Think about that financial institution; perhaps it also chooses to keep a portion of its deposits on hand in cash.

A moment’s reflection on that financial institution will cause you to recognize that it is subject to multiple equilibria. If everyone believes that the bank will remain healthy, the bank will in fact remain healthy. Depositors all can be paid back; the project will be completed; the bank shareholders are better off; the depositors are better off; the entrepreneurs who have been lent money for their projects are better off. It is a Nash Equilibrium and everyone is better off. It is not, however, the only Nash Equilibrium. A moment’s thought will also cause you to realize that if everyone else is taking their money out of the bank, then you are well advised to take your money out of the bank as soon as you can because the bank is not going to be able to pay off all of the claims on it. A solvent bank is subject to self-fulfilling prophecy. It can either remain solvent or it can fail depending upon the state not just of expectations, but in a manner much like Keynes’s Beauty Contest — expectations of others’ expectations. If I expect that you expect that TN [Professor T. N. Srinivasan, Yale University] expects that the bank is going to fail and that he is going to withdraw money, then the safer thing for me to do is to withdraw money.

If one wants a simple theory of why if you go to any American town and you look for the most solidly built building, with the most impressive façade, it will be a building that was used as a bank, particularly if it was constructed before the era of deposit insurance. It is precisely this notion of multiple equilibrium that explains why banks, more than other institutions, are at such pains to project credibility and to project permanence. Notice that this metaphor is very powerful and is very suggestive of a case for activism because it suggests that there are multiple possibilities to which the free market will be subject. It suggests that if coordination can be achieved, there are possibilities that are unambiguously better for everyone and possibilities that are worse for everyone. It suggests a natural role for government as a coordinator or as a guarantor in favor of the better equilibrium.

Now to be sure, the more rigorous and skeptical among you will have noticed that I slipped an important issue by in telling this story; I said quickly that there was no market for unfinished projects. It’s hard to understand why, if it’s a certainty that all unfinished projects are going to succeed and it’s a certainty that when they do succeed the bank will have more money — have enough money — to pay off its depositors; why there will not be some profit-maximizing actor who is prepared to come in and buy those unfinished projects and, in the process, make the depositors whole. In the language that is always used in talking — frequently used in talking — about financial crisis, it’s hard to understand how you can have a liquidity problem without at least the possibility of a solvency problem. And once you have the possibility of a solvency problem, it becomes less clear just how much you’re making everybody better off by lending the money or by allowing the projects to be partially completed.

Nonetheless, this liquidity — nonetheless, this Diamond-Dybvig bank run metaphor has been very powerful in influencing financial policy in a wide range of situations. It was this idea that was the justification, albeit without mentioning Diamond or Dybvig’s name, that Secretary Rubin [Treasury Secretary Robert Rubin] and I used in explaining to President Clinton why it was important that we support Mexico in 1995 and in countless other efforts of what their supporters call support programs and their opponents call bailouts in financial history. When institutions or countries or individuals in financial distress were provided with support, the concept of it’s a liquidity problem, so by providing support we can coordinate around the positive equilibrium, was invoked. This idea of a bank run or a set of bank runs and the provision of support provides a powerful way of thinking about a destructive, gap-increasing financial crisis.

It certainly fits a variety of situations — the situation that Mexico faced in 1985; perhaps the situation that Bear Stearns faced several weeks ago; certainly the situation that the Fed was concerned about that has led it to provide financial support in the form of back-stop facilities of many kinds to many financial institutions. It does not really address or seem in an obvious way to apply to many of the financial crises that we observe. Many of the financial crises that we observe — think of the 1987 crash or think about events in the housing market or mortgage market today — involve freely traded assets; they don’t involve banks; they don’t seem to involve bank runs. What I’d like to suggest — and I can only suggest this in an imprecise way — is that in an important sense there is a broad class of behavior that goes on in speculative markets that is in its character very similar to the character of a bank run.

One way of thinking about that bank run is to think about the value that you associate with a deposit claim on the bank in question. There’s this troubled — there’s this bank; in normal times, I have a one dollar deposit. The deposit is worth one dollar — have a one dollar deposit that’s worth one dollar — that’s easy. Now, imagine that the price of that deposit falls because there comes to be a question whether that bank is going to succeed or whether that bank is going to fail or, for that matter, that the price of that claim falls because of doubts about the solvency of the bank. Do I, as a depositor, wish to put more deposits into the bank because the price has fallen? Or do I wish to sell my claim as rapidly as I can? Think about it for a minute. The situations in which the values of deposits are falling are the very situations in which you are most eager to sell those deposits back to the bank.

To go back to economics’ standby, the most basic idea that we teach freshmen in economics is the law of supply and demand — the stabilizing nature of markets — markets as a negative feedback mechanism. We tell the story in lots of ways; there’s a surplus of apples; when there’s a surplus of apples, the price falls, and then demand increases, then there isn’t a surplus of apples anymore. There’s a shortage of apples, the price rises and then there isn’t a shortage of apples anymore. Our basic image of the economic system is as a self-equilibrating system with a single equilibrium that is determined by that law of supply and demand. If the demand curve slopes in the wrong way — if a falling price leads to less demand or, equivalently, to more supply — then none of our normal intuitions work. Price falls in response to excess demand and rises in response to excess supply, magnifying the excess. If there’s a single equilibrium, it can be unstable. If the demand curve has the more exotic shape that goes with some people responding to price over some range by reducing demand and others responding by increasing demand; and there are other possibilities.

I draw here a diagram that’s been produced many times, in many different contexts, where you engage in the intellectual exercise of — this is complicated; this is wild; this is really crazy. Let’s think of a reason why the demand curve could have both slopes. Let’s suppose that sometimes there’s the normal slope and sometimes it has the other. Then, if it fits together in the right way, there can be two equilibria — one good and one bad. One could clearly proliferate examples of this kind. If there is a mechanism present more generally in financial markets that causes the price of — the demand for financial assets to increase with price or to decrease with supply or to — or causes demand to decrease as price falls, then one has the possibility of instability that has very much the character of the bank run. Is this a reasonable hypothesis to maintain?

Chapter 5. Behavioral Finance: Reasons for Positive Feedback [00:58:38]

This brings us, and I will not dwell on this at any length, to a broad set of issues associated with behavioral finance. Like Bob Shiller, I was very struck when I encountered the assertion of Mike Jensen and others in the late 1970s that the efficient market hypothesis was the best-established fact in the social sciences. By the way, that didn’t necessarily mean that it was a well established fact, but as Jensen used the term that was its meaning. I have to say that as I make that remark about social sciences, I’m reminded of the observation somebody once offered — that if you have to call it science, it isn’t — cognitive science, Christian science, political science, social science — and so I have, ever since hearing that, never used the term economic science to refer to any phenomenon.

These issues were crystallized for me. I had followed Bob’s work on excess volatility and tried to contribute some of my own on the power of statistical tests. I was struck by the bitterness of the controversies that surrounded Bob’s work on excess volatility. In particular, where so many in finance seemed to regard attacks on the efficient market hypothesis in rather the way they would regard attacks on their spouses or children, and called for not just a rational rejoinder, but some effort at the destruction of the assailant. I remember that my sense of the issue and the merits, where my sympathies had always been with Bob, crystallized in the summer of 1987. In the summer of 1987, there was a certain set of facts that was very clear. It was very clear in the summer of 1987 that if you divided stock prices by any interesting denominator, they were at an extraordinarily high level — earnings, dividends, sales, GNP. If you were at Yale, the value — the capital — you added the debt and you divided by the value of the capital stock. It didn’t matter, all denominators; stock prices were extraordinarily high.

Then, the question was how to interpret that observation and there were two interpretations on offering. One was the conventional interpretation that would have seemed the only plausible interpretation to anyone who read the newspaper, which was that stocks had been going great and a lot of people wanted to buy stock. Therefore, they all tried to buy stock and there only was so much stock and the price went up. As a consequence of the price going up, it was probably true that the return going forward was going to be lower than it had been before the price went up. That was one interpretation. That seemed the obvious and natural interpretation.

The alternative interpretation was that, for some reason, people who used to demand 8% expected returns to buy stocks were now demand — had decided for some reason that they’d really be just — they used to demand 8% to buy stocks given the risk, but all of a sudden for whatever reason — and the reason was never spelled out with any clarity — they now were happy to accept a 6% rate of return. Therefore, all the future dividends and all the future earnings were discounted at a lower level. There was, in a way, a simple intuition for deciding what you believed. Did you believe that returns were low because stock prices were high? Did you believe that stock prices were high because people had decided they were perfectly content with low returns? Well, it was an easy way to distinguish those hypotheses. Bob did it at the time, which was to ask people, what returns they were expecting — what returns they were settling for. The answer suggested that, if anything, they favored high returns rather than — they expected higher returns than they had previously, not lower returns than they had previously.

Why do I dwell on all of this? Because it is a first example of what I actually think are a substantial variety of mechanisms, all of which can produce this kind of positive feedback behavior in financial markets. The first and easiest mechanism for understanding why there will be positive feedback behavior or, equivalently, a decline in demand as price falls and an increase in price in demand as price rises, is the extrapolative expectations that come from recent experiences. One of the things I learned painfully, being involved with managing the dollar as Secretary of the Treasury, was there is absolutely no way, no matter how powerful you are and no matter how clever you are, that the dollar can be lower without falling. It cannot happen no matter how much you want to have a falling dollar without — a low dollar without a falling dollar. In just the same way, a rising stock price leads to a higher stock price. Therefore, if people demand more after increases in price, they are demanding more as price rises. That’s one mechanism through which there can be positive feedback.

A second mechanism, which Bob has put considerable emphasis on in his writing, is social contagion. It’s people persuade other people; people look at other people. Basically, the bubble psychology takes hold as a social dynamic not necessarily related to any rational assessment of the future behavior of markets.

A third mechanism is one that can easily be given rational explanation. I am the risk manager; John is a trader. I don’t actually know whether John knows anything or not; he says he does, but he’s got a complicated story and I can’t quite tell how much he knows. He just lost a lot of money. Should I have more confidence in John or should I have less confidence in John, knowing that he lost a lot — that he just lost a lot of money? Probably, I should have less confidence in John. As an empirical proposition, his risk manager will have less confidence in John. So, positions that do not succeed have less capital associated with them than positions that do succeed, again creating this kind of positive feedback.

A fourth mechanism that will create this kind of positive feedback derives from leverage. Anyone who buys on margin or any financial institution that has a capital requirement — when they lose money, there is a margin call or a need to generate capital, which can lead to — that can lead to the liquidation of positions; again, falling price, more selling.

A fifth example and a fifth source of positive feedback comes from what one might think of as model uncertainty. There are different ways to tell the story and I suspect it could be given much more rigorous formalization than I’m going to attempt here. A prominent macroeconomist, early in his career, developed on the basis of some macroeconomic model that he had an extremely strong view that the British pound was going to depreciate. With great excitement, he found an older colleague who had a broker; called the broker and explained that he wished to place an order in the futures market to sell the pound short. The broker said, okay I’ve got your order. What would you like your stop-loss to be? The young macroeconomist said, what’s that? The broker explained that it was normal that when you placed a position, in case the position moved against you, since you didn’t want to lose too much money and if the position moved against you, it might signal that you didn’t understand this — things weren’t playing out according to your theory of the case — that you might want to sell. The macroeconomist said, hell no. If it goes up 5%, I want to buy a lot more.

The young macroeconomist did not have — went on to have a very distinguished career as a macroeconomist, but a rather less distinguished career as a speculator. If I asked you, in what way would you like to place a — you have some conviction. You’ve got some idea that the Yen is going to up — that the Yen is going to go down — something, let’s suppose it’s that the Yen is going to go up. I say, would you rather buy Yen, take your chances that the Yen might go up or the Yen might go down, or would you rather buy an option on the Yen? If you buy an option on the Yen, then there’s no way, no matter what happens, that you can lose more than the amount that you paid for the Yen. You can get very substantial leverage on the option if you buy an out-of-the-money option; you can pay only a very little bit of money. If the Yen goes up, you will multiply your initial stake by a very substantial amount.

Most people presented with that choice, if they believe the option is priced fairly, choose to buy the option. But if you think about what that means, that is a decision to pursue a positive feedback trading strategy. That is a decision that the more the position moves in your favor, the more your option is in-the-money, and the larger is your exposure to the underlying stock or the underlying currency. Anyone who chooses to use options as a trading vehicle is choosing an approach in which they are going to increase the magnitude of their position as the price rises or decrease the magnitude of their position as the price falls — again, exactly the opposite of the natural, stabilizing response.

A final source of this kind of positive feedback is the prospect that increases in movement for any of these reasons are jumped on by those who simply anticipate the positive feedback behavior of others. Therefore, trends follow in an effort to front run others who are generating these trends. I would suggest to you that consideration of these mechanisms, along with consideration of the experience in markets — a basic pattern that we tend to observe in markets — positive serial correlation in the short run, mean reversion in the long run — suggests the pervasiveness of this kind of positive feedback behavior and, therefore, suggests the likelihood that financial markets will be subject to very substantial instability and multiple equilibrium. Just as the positive feedback associated with the bank run makes a case for — or least opens up a case — for activist policy, so also, in the case of financial markets, the possibility of downdrafts caused by falling prices that lead to selling — that lead to more selling — lead to falling prices — or updrafts caused by the opposite mechanisms, make a potentially strong case for government action to contain financial crisis.

Chapter 6. Summary and Questions on Government Interventions and Moral Hazard [01:15:37]

So far, I’ve tried to make the case that I want to make for today — that the macroeconomics of the next generation will be much more about finance and financial crisis than the macroeconomics of the generations that shaped our textbooks; that the social gains from preventing and containing financial crises are likely to be significant. And offer the prospect of achieving the deep-seated goal of every Keynesian macroeconomist: reducing the average output gap over a long period of time and thereby producing more total output. That a realistic theory of financial crisis has to be based on something other than the idea that the market is efficient. And that the notion that bank run-like behavior can be generated out of the behavior of speculative participants in financial markets who respond perversely, from an economist’s point of view, to price signals, is a highly plausible one. Tomorrow, I hope to illustrate that these broad ideas have connection with the current financial crisis that we face and to suggest — to evaluate — some of the policy measures that have been undertaken and to suggest some policy actions for the future. Thank you very much. Bob, I don’t know what the format here is, but I’d be happy to answer a few questions. Yes, sir?

Student: If the government seeks to stabilize this plan of avoiding these perverse positive feedback consequences, doesn’t that very government action itself make for this very possibility? Does the government continue doing that and, therefore, it gets to a point where it says no, it will have created its own positive feedback mechanism reverse itself?

Professor Larry Summers: I had a conclusion addressed to the word moral hazard that I decided to defer to the first part of what I was going to say tomorrow. But, since you ask essentially the question, I’ll address it right now. I chose, in summarizing the main points, to use the phrase — this possibility of positive feedback — open space for constructive government action precisely so as to not judge with absolutism that in every case there would be desirable things for governments to do, precisely because of the expectational effects that you describe. It seems to me that before embracing moral hazard fundamentalism and the idea that if the government gets involved, then there will be expectations of the government doing things, and then if their expectations that the government’s doing things, people will respond to those expectations and then it’ll surely be bad. Before embracing that idea, one needs to pause over three points. First, moral hazard is not a reason why there should not be insurance or a reason why problems should not be addressed by collective action.

The standard wise crack on the subject is: the fact that I smoke in bed is not usually considered a good reason for the fire department not to come to my house to put out the fire, especially if my house borders very closely on many other people’s houses. The first point is, yes there may be adverse effects that come from the expectations and there may be beneficial effects that come from the direct action; those have to be traded off.

Second, there is a certain sloppiness in supposing that all induced behavioral changes are to be thought of as undesirable. Consider a standard bit of public policy — the placing of guard rails beside highways. The government chooses, as part of the provision of a public good, to place guard rails beside highways and to place good guard rails beside highways. It is a documented, empirical fact that as a consequence of the presence of those guard rails, people drive faster than they otherwise would. How should one think about, as an economist, the fact that they drive faster? As in doing a welfare evaluation of guard rails, I think a moment’s thought should convince you that since they had the option of driving at the same speed that they did before, in the presence of the guard rails they adjusted their behavior. The welfare gain from building the guard rails is greater than it would have been if people — is magnified by the behavioral response that takes place.

If people decide, for example, that because we have deposit insurance they are not going to check on the expressions of the bank tellers at their bank everyday, that I would argue is an induced behavioral response from a public intervention, but it is one that magnifies the gain from deposit insurance rather than reducing the gain from deposit insurance. The third point to be considered, and it’s getting a bit ahead of our story tomorrow, is that there is no — is that one has to very carefully consider the cost to the government of contemplated actions. People talk about bailouts in a rather indiscriminate way, but it is a very different thing for me to loan my friend, John, $1,000 than for me to give my friend, John, $1,000. If I loan my friend, John, $1,000 at a premium interest rate that exceeds the cost at which I borrow, it may even be profitable for me to lend money to my friend, John. If I am making a profit, the question of in what sense I am subsidizing him becomes a rather — becomes a rather subtle one.

So, take as a concrete example that poses the issue: the loan that the United States made to Mexico in 1995. The loan that the United States made to Mexico came at a premium of, let’s say, on average about 200 basis points over the Treasury’s borrowing costs. At the time the United States made it, Mexico would have probably been unable to borrow at any interest rate below 1,000 basis points and perhaps would have been unable — more realistically, probably would have been unable to borrow at all. So, how does one think about that 200 basis-point loan? One way to think about it is, we made a profit; 200 is more than 0. Another way to think about is, well wait, the market price of a loan to Mexico was 1,000 basis points; we made the loan at 200 basis points, therefore the loan is under water in the first instant. Another way to think about it is, well the market price is 200 basis — the market price is 1,000 basis points, but that’s assuming that it’s Citibank trying to collect from Mexico. Probably the United States is going to be much better at collecting from Mexico for a variety of reasons than Citigroup, just as the Fed is going to be much better at collecting from Merrill Lynch than most others.

There’s considerable subtlety involved in judging whether lending activity at premium interest rates does or does not constitute a subsidy. Almost always, it will be profitable if paid back relative to the lender’s borrowing costs. Almost always, it will be at an interest rate that is lower than the borrower could borrow at on the markets. And almost always, in the case of public or quasi-public activity, there will be a reasonable argument that the lender is better at collecting than the private sector would have been. So, the question of evaluating subsidy is difficult. The position that I always took as Secretary of the Treasury, with respect to the IMF, was, look we’ve got a long experience and over time our loans have been paid back and we’ve made money; so, you can’t really think of this as bailout and subsidy. I think it was a pretty strong argument.

I would want to resist in the strongest terms the idea that because government involvement will affect private sector behavior, there is to be some presumption that it is to be avoided. Yes?

Student: Why didn’t people see these defaults coming?

Professor Larry Summers: The world is a very uncertain place and default risks are very, very difficult to judge. It’s far from clear that, as a general matter, investors understate default risks. Indeed, I think most people who look closely at pricing patterns in bond markets would tell you that, if anything, investors systematically overstate default risks. So, on a very consistent basis, if you buy riskier bonds and sell safer bonds, that on average proves to be a profitable strategy and even proves to be a profitable strategy adjusting for the degree of risk that you’re taking. The premise of policy has tended to be a great deal of emphasis on transparency in the hope that with more transparency and better information, investors will make more rational decisions and will make more rational decisions ex ante. There are also sets of issues that are involved here — and I’ll say something about this tomorrow — having to do with the fact that the optimal number of auto accidents is not zero. The optimal number of plane crashes is not zero. There is a trade off to be made between accepting a certain possibility of accidents and a variety of advantages that come from the conservation of capital and the making available of finance in a variety of ways. Thank you very much.

[end of transcript]

Back to Top