Mark Watson on U.S. economy
Loungani: Is the U.S. economy out of the woods now?
Watson: I think so. There’s a very nice indicator of the U.S. economy that the Federal Reserve Bank of Chicago produces every month. It’s called the National Activity Index, and it’s an average of about 85 different measures of economic activity. That indicator is looking very positive and does suggest that we’re out of the woods. The past four or five months have looked good, including the employment picture.
Loungani: Could another shock—say, a sustained increase in oil prices—push the U.S. economy into another recession?
Watson: I think we have enough momentum to ride out even some very large shocks. Of course, in forecasting the economy, “never say never.” Certainly one can imagine shocks that can be devastating to confidence. Moreover, as you know, our econometric models explain, at best, 25 percent—or 35 percent if we really push them—of the variance of near-term economic activity. So there is a lot that drives the economy that we just don’t know about.
Loungani: Is inflation kicking up?
Watson: There’s a lot of inertia in inflation. The fact that it has been low for a while helps keep it down. And the Fed [U.S. Federal Reserve Board] is much more credible now than it was, say, 20 years ago. That keeps long-term inflation expectations pegged or nailed down at a low level. Now this doesn’t mean there cannot be short-run blips—up and down—in inflation. It also doesn’t mean inflation can’t get away from us again, but the likelihood of that is pretty low. We should, of course, expect some upward pressure on inflation from the pickup in real activity, but it can be contained by the Fed’s actions.
Loungani: What has the profession learned over the past two decades about forecasting economic activity, particularly recessions?
Watson: We ‘re much more successful at predicting where the economy is now than where it is going. Saying that we’ve gotten better at predicting the present sounds odd to people outside the forecasting profession. They say, “You should know where you are now; why do you have to predict it?”
We are better at “nowcasting” than at “forecasting.”
But it isn’t always easy to describe in real time what the state of the economy is; there are almost always conflicting signals. So even being able to provide a more accurate description of the present state of the economy is an improvement over where things stood 20 years ago. We’ve been less successful at saying where things will be in six months, and predicting a recession remains a difficult, if not impossible, task. Certainly the models that Jim Stock and I developed failed to predict the past two recessions. To use what I’m told is the new terminology, we are better at “nowcasting” than at “forecasting.”
Another thing we’ve learned is that to forecast the economy, it is better to use 70, or even 700, variables rather than 7. There’s more information in the economy than can be captured in a small number of variables. It’s better to average the information in a large number of variables than to select a few up front. And then, because the structural features of the economy may be changing, the statistical model that you use should have some ability to adapt. In the jargon, the model should have time-varying coefficients instead of coefficients that are fixed for all time.
Loungani: Having gone through another business cycle, what do we know now about what causes recessions—a few large shocks or many small ones?
Watson: My view is that recessions occur as a result of a sequence of small shocks from many sources that eventually accumulate. We get a run of bad luck, a bunch of negative shocks—there’s a big run-up in oil prices, the Fed tightens monetary policy around the same time there are other kinds of demand shocks. We just get, if you will, unlucky. The economy turns down. We know that even a simple linear statistical model, with some dynamics built in, generates cycles when perturbed by small shocks. This is what [Eugen] Slutsky told us ages ago. But it still seems to be a fair description of what causes cycles in the U.S. economy.
Loungani: You don’t agree then with James Hamilton, who argues that nearly every modern U.S. recession is due to an oil price increase associated with a supply disruption?
Watson: Oil is part of the story and one of the negatives in several of the past U.S. recessions. But in a couple of those episodes there were also monetary tightening and other kinds of demand shocks. All these things conspired against the U.S. economy; any one of them by itself wasn’t big enough to push the economy over into a recession.
Loungani: Have recessions become less frequent now, and is this behind what’s being called the “great moderation” in incomes?
Watson: In the 1970s and early 1980s, expansions were not too long and recessions quite frequent. But then the rest of the 1980s went by without a recession, and when one did come, in 1991, it was a pretty minor one. Then you had another decade of growth followed by our recent mild recession of 2001. This is certainly part of the story behind the great moderation in the volatility of U.S. aggregate incomes. But even if we look at times outside of the roller-coaster periods of recessions and recoveries, quarterly swings in real GDP—in our incomes, that is—are more moderate now than in the past.
Loungani: Many people will be surprised to be told that their incomes have gotten smoother.
Watson: As they well should be. What we’re talking about here is a moderation in aggregate incomes, not necessarily in individual incomes. Those remain, as I understand it, as volatile as ever. This can happen for good and bad reasons. For example, if your health suffers, your income will quite likely become more volatile. But there are plenty of good reasons why our individual incomes are volatile. In a dynamic economy, people are changing jobs, always looking for new ways to do things, and so on.
Loungani: So why should an individual care about what has happened to the aggregates?
Watson: It’s good news that the fluctuations in our individual incomes are less synchronized and less tied Watson: “It’s good news that the fluctuations in our individual incomes are less synchronized and less tied to the business cycle.” to the business cycle. If my income takes a hit the same time as yours, I’m going to be less able and less inclined to help you out. But if fluctuations in our incomes are on different cycles, we can each help the other out through bad phases. That’s why the moderation in aggregate incomes is good for the individual. In the jargon, we can—as a group—insure better against the idiosyncratic risks to our incomes.
Loungani: Why have we had this great moderation?
Watson: We don’t know yet. We have several candidates, but none takes us all the way in providing an explanation. One theory is that the Fed is doing a better job of conducting countercyclical policy, of leaning against the wind when economic times start getting turbulent. But that story turns out to be more useful in explaining what has brought inflation down over time than in explaining the great moderation. Another possibility is that the shocks have gotten less severe. But the shocks that we can identify, such as oil shocks, are just as variable as before.
Loungani: Hasn’t the switch from manufacturing to services helped stabilize incomes?
Watson: That’s not the whole story either, because even the volatility of aggregate manufacturing activity is lower now than it used to be. We’re left, then, with good luck as the possible explanation. But that’s an unsatisfactory explanation intellectually, and so the work will go on.
A student of mine, Austin Saypol, studied changes in financial institutions, such as in the way mortgages are issued and the way houses are financed. He found that those changes could explain a large part of the reduction in the volatility in residential construction, which is a sector that accounts for a fair chunk of the volatility in real GDP. You can extend that idea and look at changes in the way autos are financed, in the way washing machines are financed, and so on. Perhaps the great moderation will turn out to be due to the evolution in our financial institutions—that, at least, is my conjecture for the moment.
The IMF Survey (ISSN 0047-083X) is published in English, French, and Spanish by the IMF 22 times a year, plus an annual Supplement on the IMF and an annual index. Opinions and materials in the IMF Survey do not necessarily reflect official views of the IMF. Any maps used are for the convenience of readers, based on National Geographic’s Atlas of the World, Sixth Edition; the denominations used and the boundaries shown do not imply any judgment by the IMF on the legal status of any territory or any endorsement or acceptance of such boundaries. Text from the IMF Survey may be reprinted, with due credit given, but photographs and illustrations cannot be reproduced in any form. Address editorial correspondence to Current Publications Division, Room IS7-1100, IMF, Washington, DC 20431 U.S.A. Tel.: (202) 6238585; or e-mail any comments to firstname.lastname@example.org. The IMF Survey is mailed first class in Canada, Mexico, and the united States, and by airspeed elsewhere. Private firms and individuals are charged $79.00 annually. Apply for subscriptions to Publication Services, Box X2004, IMF, Washington, DC 20431 U.S.A. Tel.: (202) 6237430; fax: (202) 623-7201; email: email@example.com.