John Campbell is a British-American economist. He is a professor and chairman of the economics department at Harvard University and is known for his research in financial economics, macroeconomics, and econometrics.
Dan Richards interviewed Campbell at the American Economics Conference in late January.
This is a transcript of the interview. A video appears here.
In June of last year you published a paper that looked at the stock market decline from 2007 to 2009, and compared that to the tech bubble of 2000 to 2002. What were the similarities between those declines?
There certainly is a very basic similarity. In both cases the stock market went down a lot in a short period of time. So that is the starting point. Frankly the similarities end roughly there, because the decline at the beginning of the 2000s came off the heels of an unprecedented run-up in the market to valuation levels that were, as you might say, at "nosebleed levels." That caused people like [Yale economist] Bob Schiller and myself to sound the alarm about valuation. The downturn in 2007-2008, however, came from a background of high, but by no means unprecedented, stock prices.
In this paper I looked at this from a point of view that I have developed in my academic work over many years, which is to argue that there are really two sources of fluctuations in stock prices. You can think of these as cash flows and discount rates. Cash flows mean the profits of companies, reflected in the long-term profitability of companies that will be paid out to shareholders over time. That can vary, obviously, and if bad news comes in about profits, stock prices will go down.
Traditionally, finance academics thought of future cash flows as the only force moving stock prices, but in recent years we have come to realize that discount rates also matter. For example, when interest rates are very high, or when investors are feeling cautious, then they discount future cash flows at a high rate, and stock prices will fall even if profits don’t change. This is the type of price variation that we see in the bond market. Treasury bonds make fixed payments and their prices vary because their discount rates vary.
I've done a lot of work over the years trying to separate these two factors in the market. I applied these methods to these two downturns. The conclusion I reached was that the early 2000s were primarily a change in the discount rate that investors were applying to their profit forecasts. Essentially they became more cautious, and so they discounted profit expectations at a higher rate.
Whereas the downturn that started in 2007 and accelerated in 2008 was primarily due to bad news about profits. The reason that is important is that bad news about profits is permanent; there's no reason to expect it to turn around, although it may. A decline that's caused by a higher discount rate, on the other hand, essentially means that if you just hang on, things will come back just as they would if you were holding a bond.
What are the implications of this for investors?
For long-term investors, a decline in the market that is caused by a profit decline is a much more serious matter. For that reason, each percentage point of decline in 2007 and 2008 really had a much more serious effect on long-term investors than the same percentage point of decline in 2001 and 2002.
You see that, for example, in the reaction of universities, just to take one organization that I am familiar with. Universities cut back their spending much more aggressively in 2008 than they did in 2001 and 2002, despite the fact that in many cases their endowments were going down by comparable amounts. In my view the reason was that universities correctly saw that this was a different type of decline that would require a more serious and significant response.
You were one of 15 participants in the Squam Lake Group that just recently produced a book. How did those 15 academics from different universities and different points on the ideological spectrum come together in the fall of 2008?
First of all, many financial economists felt a great obligation to comment on what was happening, to offer diagnoses, and try to think through and come up with some policy prescriptions that could help prevent a repeat. So there was certainly a lot of pressure that many of us felt to stand up given the circumstances and try to contribute.
But I also want to give a lot of credit to Ken French from Dartmouth, who convened the group. Ken turned out to be an extremely talented convener and herder of cats, if you like.
It took place close to Dartmouth at the Squam Lake Resort in New Hampshire, which gave the group its name.
That's right. We were cognizant of the fact that resorts in New Hampshire have been named for policy contributions in the past, thinking of the Bretton Woods system after World War II, so we thought a lake in New Hampshire wouldn't be a bad name for the group.
We had an initial brainstorming session in November of 2008 when events were still unfolding. Through 2009 we put out white papers and memos, which we distributed online and through the Council on Foreign Relations in Washington, and we had our own website.
We realized we could have more impact if we pulled these together into a book. We were able to do that very rapidly, thanks in part to cooperation with the Princeton University Press, which expedited the production process. We gave up all royalties. We said just put the book out quickly, price it low, and distribute it.
We were able to wrap these individual policies suggestions – each of which became a chapter of the book – into a broader framework offering an overall diagnosis.
I believe there were nine or 10 broad policy recommendations. One of the hot issues related to capital requirements for financial institutions. What was the consensus among the group on that issue?
One of the things that the group agreed on very early was that capital requirements really should be set with an eye to the risks to the system as a whole. One of the big themes of the book is that, to be effective, regulation has to think about the health of the system, not just the health of individual financial institutions. It is very easy for regulations to have unintended consequences that actually damage the system even while shoring up a particular institution.
On this basis, we argued that capital requirements for banks should be set in relation to their systemic risk, not just to the individual risk of the bank.
So bigger banks, and banks engaging in more risky activities, would have different capital requirements.
That's correct. Larger banks and banks with shorter-term funding models also pose more systemic risk. Think of Northern Rock, for example. That was not a bank with an unusually risky asset side of its balance sheet. Its problem was that it relied almost entirely on wholesale funding. So it had a risky liability side, and that is what made it systemically risky. These things should be taken into consideration in setting capital requirements.
Another recommendation dealt with the issue of contingent capital. Can you talk about that?
One of our strongly held views was that, to get the financial system working properly, it was really vital to put the banks’ creditors at risk, so that they enforce some market discipline on the banks and the taxpayer is left with less of the liability when the bank goes down.
It has been very disturbing, watching the crisis unfold, how often the senior creditors of banks have been fully protected. For example, that famously happened in Ireland, which is one of the reasons why the Irish fiscal crisis is now so severe. So we thought it was important that the creditors of banks and not just the initial equity holders should be at risk.
The difficulty is that when banks finance themselves entirely with short-term debt, the short-term creditors have the ability to run the bank and essentially get out of harm's way before they can be made to share in the losses of the bank. So long-term debt is an important part of the capital structure, but it is necessary to have a mechanism that will simulate the effects of a traditional bankruptcy on those creditors.
In a traditional bankruptcy the business is kept going, the equity holders are wiped out, and the long-term creditors become the new equity holders. That is an efficient means to essentially take the losses and put them into the new owners of the firm, while keeping the firm running if it is a viable business going forward.
The difficulty is that the traditional bankruptcy code operates too slowly to handle the needs of complex modern financial institutions. We need something that can happen quickly and simulate the effects of bankruptcy but more rapidly.
Our proposal was for what we called hybrid-regulatory convertible debt, with provisions in banks’ long-term debt contracts that will under certain circumstances trigger conversion of debt to equity, thereby diluting the previous equity holders and converting the bondholders into equity holders.
A very topical issue today around the world is the regulatory framework that we need given the evolution of the financial system. What did your group conclude on that issue?
Once again, our overarching concern was getting a sufficient focus on the health of the system. So we argued that what is needed is a systemic regulator who is concerned with the health of the system.
Our preference would have been to see the central bank in the US, the Fed, take that role. Ultimately the central bank is the institution that has the economic expertise, the interactions with the markets through the regular conduct of monetary policy, and the ultimate responsibility of stepping in during the crisis.
Politically in this country it is very hard, particularly at present, to provide more power to the central bank. So the Dodd-Frank legislation has instead created a systemic counsel of regulators. That's better than nothing. It wouldn't have been my first choice, but it is likely to be workable going forward.
Very importantly, we feel that this systemic regulator, or counsel of regulators, needs to take a very broad view of the financial system. Part of the problem that became clear in the crisis is that modern finance had evolved to give a more and more important role to the so-called "shadow" banking system that was outside the traditional regulatory framework yet was performing vitally important functions. When that system went down, it turned out society couldn’t tolerate that. We had to step in. We had to bail it out, yet there was no regulation in place before hand.
The last area I would like to talk about is one that perhaps isn't at the top of the list in terms of importance from a policy perspective, but it is a hot topic in the media. That relates to executive compensation at financial institutions, which is an area, obviously, of some debate. People like Mervyn King, governor of the Bank of England, have come forth with some fairly strongly held views on this issue. Where did the group net out in terms of a framework for thinking about compensation?
This is an incentive problem. Compensation needs to align incentives correctly. That has been a theme of academic economists for some decades now.
But academics missed one very important trick. In the past, we have focused almost exclusively on aligning the interests of managers with those of shareholders. Now that creates problems when shareholders have the wrong incentives, which is the case when an organization becomes distressed and when it has the potential to require a bailout.
The compensation system should be designed to align the interests of managers with those of the creditors of banks as well as the shareholders of banks, the long-term bondholders, and ultimately the taxpayers, the people who are going to be on the hook if problems occur.
The mechanism for doing that is to take part of managers' compensation and defer it in such a way that managers themselves are unsecured creditors of the bank, because their compensation is sitting in an account and it won't be paid if the bank goes under. That will give managers the incentive to think long-term, think about downside risk, and manage these institutions more prudently.
Did the group reach a consensus on the percentage of managers' compensation that it would take to create that effect?
We stopped short of becoming too specific about that, but we feel it should be a significant chunk of compensation that is held back.
I should mention that I have been involved over the last few years with the management of the Harvard Management Company (HMC), which invests Harvard's endowment. HMC has had a compensation system like this for some time, including withholding provisions that ensure people aren't paid until some time after good performance, and subsequent bad results will undermine pay. That has worked very well in the Harvard context, and it can also work well for large banks.
Read more articles by Dan Richards
