Meet the Street: Sept. 11 Was Crucial to Recession Declaration

A member of the committee that last week announced the U.S. economy is in recession discusses that report.
Publish date:

Jeff Frankel, who sits on the committee that recently declared the U.S. economy to be in recession, says Sept. 11 was the deciding factor in that declaration.

Jeffrey A. Frankel,
Professor, Capital Formation and Growth,
Harvard University

Recent Meet the Streets

Evergreen Investment Management's
Prescott B. Crocker

WNBA New York Liberty's
Rebecca Lobo

Citigroup Private Bank's
Clare Costello

Suresh Gopalakrishnan

Ethan Allen's
Farooq Kathwari

Were it not for the effects of those attacks, says Frankel, the activity that began in March may have just been termed a slowdown rather than an outright recession.

Using a set of four indicators -- industrial production, employment, real income and wholesale-retail trade -- Frankel's group, the Business Cycle Dating Committee of the National Bureau of Economic Research, determined last week that the last business cycle peaked in March of this year and has been declining since.

Though the layman's definition of a recession is two consecutive quarters of negative economic growth, and almost every recession in the past 50 years has been accompanied by such a decline, the NBER prefers to base its calculations on these other indicators, because they are less subject to revision than the GDP data. spoke with Frankel, who is also a professor of capital formation and growth at Harvard University's John F. Kennedy School of Government, about the group's methodology, the differences between this recession and recessions past, and the outlook for the next economic expansion, among other things.

TSC: How long has the National Bureau of Economic Research been dating business peaks and troughs, and how did the organization come about?

Frankel: The NBER has existed since 1920, and initially its whole purpose was this: business-cycle economics, of which a lot of it was dating the peaks and troughs. In 1960, it was given official status, in the sense that the Bureau of Economic Analysis of the Department of Commerce, which compiles GDP, said that what the NBER says is a recession is a recession, and they use it for their own purposes.

TSC: And what purposes are those?

Frankel: Some things are probably legally triggered. In terms of policies, often Congress extends employment benefits such as the number of weeks you get during a recession, though they don't have to. The fact that we did this might put pressure on Congress to pass the fiscal-stimulus package. But that's not what we see as our purpose. Our purpose is to have everybody agreeing on the same dates.

TSC: On what basis did you determine that the downturn began in March? I know that you tend to concentrate on a set of four indicators -- industrial production, employment, real income, and wholesale-retail trade -- but it sounds like the data were inconclusive. What made March the best month?


There's a couple of things. Of the various indicators we looked at, some, at face value, would have indicated that a recession started a year ago -- those are the ones more related to manufacturing, which is not that large a share of the economy. Other indicators, such as the ones related to general income, would have said it began this fall. So one guiding principle is compromise. It's nice to pick a date in between. And in this case, March was particularly obvious because employment peaked, which is probably our most important cyclical indicator. It's available on a monthly basis; it doesn't tend to be revised a lot; and it covers the whole economy, unlike industrial production, which only covers manufacturing.

TSC: The NBER says that the indicators it uses are subject to less revision than the quarterly GDP figures, which some economists use to define recessions. Why are the indicators you use less subject to revision? What makes them more reliable?


Well, GDP is a pretty ambitious thing to measure. When the first estimates come out, there is a lot of incomplete information. The components of GDP, to simplify it, are consumption, investment, government spending and net exports. When the initial announcement comes out, that's more than a month after the quarter has ended. Even then, we have guesses for some of the components, like investment and trade.

So it's inevitable that a lot more information is available. They often go back and change the whole thing because of various index problems and things like that. Employment, on the other hand, is the number of people that have jobs, and while there are certainly measurement difficulties, it is a fairly well-defined number. There are all kinds of conceptual problems, for example, with the dollar value of goods and services produced by the economy. How do you compare the value of computers and wheat? So

GDP is inherently more subtle.

TSC: How do structural changes in the economy during the 1990s -- increased productivity and the reduced impact of manufacturing -- affect the NBER's methodology for dating business cycles?


There's nothing official about this, but in fact, out of the four measures we talk about, we now put less weight on the two that are related to manufacturing than we would have in the recent past -- those two being industrial production and retail trade.

TSC: What do you make of the fact that personal income hasn't declined as significantly as it has in past recessions?


That is a little bit of an anomaly, though there have been past recessions where that personal income indicator lagged behind. It's been sustained in part by some things like falling import prices. I think, even though we're not in the business of prediction, it is quite possible that it will turn down. Our job would be easier if all four variables turned down together, but they don't.

Another interesting indicator is the aggregate hours of work -- which is my own personal favorite. It's the number of people employed times the average work of their workweek. It's not one of the official four, but we do look at other variables.

TSC: If the Sept. 11 attacks were such an important factor in tipping the economy into recession, as you say in your release, how do you reconcile that with dating the recession back to March?


Right. Well, a lot of people have asked that, and one answer is, we make a two-part decision. One is, has the downturn been sufficiently severe, broad and well-entrenched to say that there is a recession? And it's a separate question when it started.

In this case, what's happened to the economy over the last few months, and presumably the effects of Sept. 11, is what it took before we became convinced to a degree of certainty that we were in recession. And then you look back at when the recession started. I would say that if Sept. 11 never happened, and the economy had leveled out or improved, then we would just be saying this was a slowdown that started in March, but that the slowdown wasn't severe enough to merit the title of a recession. Sports analogies are overdone. But it's like, which of the various runs that were scored actually won the game?

TSC: It seems like there has been some criticism of the NBER's timing in making these announcements. Why does it take so long to determine where the peaks and troughs are?


I don't know if it's really criticism. I think a lot of people just have an instinctual reaction. Since it's always well after the fact, and those who haven't heard of us, or think that a recession is defined by two quarters of negative growth, they'll say, oh, these geniuses have finally figured out what the rest of us always knew. So I think there's a certain automatic reaction like that, which was especially true 10 years ago. It's been a little less so this time around.

There are certain lags in gathering the statistics, but unlike most economic commentary, or most of what you read in the newspaper, or most of what we economists do a lot of the time, the job of this committee is to speak only when it's virtually certain. If you had asked us as individuals, we all would have said earlier that we were in recession. But the point is to wait until we're certain, because we don't want to have to change our minds if the statistics get revised or the economy suddenly rebounds. People thought we were going into recession in fall of 1998, and that was just way premature.

TSC: What past recession does the current one resemble most?


I would say that because of the Sept. 11 component, there's some resemblance to the last recession in 1990-91. There's a conflict in the Middle East taking an economy that had been just beginning to slip into recession and pushing it over the edge. In both cases, a Bush was president and Greenspan was



But basically I would emphasize the differences and look at the whole cycle of the 10-year expansion of the '90s followed by this recession. For me, the biggest difference is that the previous expansions in the '60s, '70s and '80s were to a large extent driven by the government, by easy monetary and fiscal policy. For example, a lot of the increase in employment was government-related. They were fairly long-lived expansions, but by the sixth or eighth year of the expansion, inflation was rising and the budget deficit was growing.

Indeed, the causes of the subsequent recessions, in part, were efforts to address those longer-term problems. In the early '80s, you had the Fed sharply raising interest rates deliberately in response to inflation. And in 1990 you had Bush and Congress agreeing to raise taxes in response to the rising budget deficit. So those things contributed to those recessions, and the fact that inflation was high and budget deficits were high sharply constrained the government's ability to use expansionary monetary and fiscal policy in response.

This time it's completely different. The '90s expansion was based on private firms undertaking investment, and households undertaking consumption, while employment was in the private sector. By the 10th year, inflation was still low, and the budget was in surplus. It means that other things caused the recession, and that the tools of monetary and fiscal policy are available to be able to combat it.

TSC: Do you think the economy has responded to the Fed's stimulus yet? Are there specific signs, or is it still too soon?


I think it's too soon. Maybe the slide would have been even steeper if we hadn't had it. But it's really too soon to see much.

TSC: Does fiscal stimulus tends to have a big impact on pulling an economy out of recession?


It usually isn't a big impact, but it can help. The problem typically is not on the economic side, but on the political side, and I don't knock politicians. It's inevitable that if you're spending people's money, you need the political process trying to reconcile priorities. But the result is usually that by the time the fiscal stimulus comes, it's too late and was badly designed. That's evident currently as well.

TSC: Will the U.S. economy begin to see longer and longer periods of expansion in the future? Is there any kind of certainty about that at this point, or is it totally unpredictable?


Well, during the whole postwar period, the recessions have been shorter and the expansions have been longer than was true previously in our history, and I think that will probably continue. But will we improve on the expansion of the '90s? No, I don't think so. That was an unusual experience, and part of the problem is we kind of got spoiled by both the length of the expansion and the magnitude of the growth. I don't think we'll go further in that direction. We'd be lucky to do that well again.