I see, now that the numbers favor GWB, we have a conspiracy.
Darn convenient.
When the economy hits a turnaround, conspiracy theories abound.
The New York Times
August 22, 2002, Thursday, Late Edition - Final
BYLINE: By Alan B. Krueger; This column appears here every Thursday. Alan B. Krueger is the Bendheim Professor of Economics and Public Affairs at Princeton University and editor of The Journal of Economic Perspectives.
"Prediction is very difficult," Niels Bohr once remarked, "especially about the future." Lately, predicting the recent past of the economy has proved difficult too. Actually, determining last quarter’s or last year’s G.D.P. growth has always been difficult, and subject to error and revisions.
The daunting task of computing G.D.P – that is, the total value of economic output – for an ever-evolving $10 trillion economy falls on the Bureau of Economic Analysis, a nonpartisan agency in the Commerce Department. The Bureau bases its estimates on information from a diverse set of sources, not all of which are instantly available. Final tallies of corporate profits, for example, are not available until three years after the fact.
To produce timely information, the Bureau announces an “advance” estimate of G.D.P. within a month of the end of each quarter. Missing components are extrapolated from recent trends. As more complete information becomes available, the advance estimate is revised. Every year, usually in July, the Bureau updates the preceding three years’ worth of data.
Late last month, as part of its regular revision cycle, the Bureau announced eye catching revisions to the preceding three years’ G.D.P. figures. Most strikingly, the economy began to slow more than previously realized in 2000, and was shrinking in the first half of 2001.
Had this been known sooner, policy very well could have been different. For example, the backend of the $1.4 trillion tax cut would have been much harder to justify if it was apparent that the economic contraction underway in early 2001 was cutting into the budget surplus.
Why does the B.E.A. periodically rewrite history? Why not wait until the final estimates are available to tell us? This is like asking why people look at a weather forecast: why not just wait to see if it rains over the weekend before planning a picnic?
The preliminary estimates are imprecise, but they give an indication of where the economy stands. Surely, no one would want to wait three years to find out if economic growth was slowing down.
Besides, the B.E.A. has a reasonably good track record, especially in comparison to other countries.
Economists judge the accuracy of the initial estimates of G.D.P. by whether the future revisions can be predicted from currently available information, and whether the revisions tend to be in a positive or negative direction. On this score, revisions are “efficient,” meaning they are hard to predict and not too far from zero on average -- even though the average absolute revision to quarterly G.D.P. growth due to subsequent revisions is around 1.4 percentage points (on an annual basis).
A careful study by Jon Faust, John Rogers and Jonathan Wright, three economists at the Federal Reserve Board, found that revisions to G.D.P. growth in the United States are the smallest among the G7 countries (consisting of Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States). In addition, the direction and magnitude of revisions to the initial G.D.P. estimates are harder to predict from available information like stock returns and interest rates in the United States than elsewhere. They also find that in the United States revisions tend to be no different in an election year than in other years.
This record is even more impressive in view of the fact that, on average, the delay between the end of a quarter and when G.D.P. is announced is twice as long in the rest of the G7 than in the United States: 30 days in the United States versus 60 days abroad.
Nevertheless, it is generally the case that early estimates of G.D.P. growth are particularly noisy around turning points – that is, times when the economy enters or exits a recession -- because the extrapolations used for incomplete data no longer provide good approximations.
Consequently, revisions that encompass turning points regularly expose the Bureau to sniping and unfounded accusations of partisanship. Writing in the Washington Post, for example, Robert Novak accused the career professionals at the B.E.A. of “incompetence” and suggested they cooked the books under pressure from the Clinton administration to make 2000 look better than it was. Rush Limbaugh accused the B.E.A. of deliberately lying. And Martin Hutchinson of the U.P.I. questioned whether the B.E.A. and Alan Greenspan, “heavily influenced by his liberal Democrat wife,” engineered the statistics to give the illusion of a productivity miracle.
If they were not taken seriously by some people and amplified in the press, these attacks would be amusing. But, instead, they damage the credibility of government statistical agencies and erode morale among their employees.
Had the B.E.A. really cooked the books, why would it bother to correct them? And why would the B.E.A. deliberately understate the contraction early in 2001 if it was part of a left-wing-Mitchell-Greenspan conspiracy?
Most telling, the recent revisions to G.D.P. are not out of line with history, especially around turning points. From 1978 to 2000, two thirds of eventual revisions to G.D.P. growth were between –1.1 and 2.0 percentage points. Revisions to the G.D.P. growth in the last three years are not statistically inconsistent with this range, though some components are somewhat more variable than usual. The average absolute revision in this period to date is 1.4 points, right on the long-term average.
“The last time we went through a debacle like this,” Steve Landefeld, director of B.E.A. said, “was when the economy was recovering from the recession in the early 1990s.” Then the agency was accused of overstating growth to make the previous Bush administration look good.
Rather than respond to politically motivated critiques, the B.E.A. has worked on a plan to minimize the size of revisions by improving measurement of services and software, which account for more than half of recent revisions. But it is disappointing that the White House has not jumped to defend the agency against these attacks. A statement of support by someone on the administration’s economic team would help shore up morale and maintain confidence in economic data.
While Monday morning quarterbacking is a popular sport, it is always difficult to predict revisions to G.D.P. Clearly the Bush administration misjudged how the economy performed in 2001. In March, Treasury Secretary Paul O’Neil said, “It seems quite clear now that our economy never suffered a recession."
Prediction is indeed difficult, but it is predictable that anyone who makes too much of noisy preliminary data or alleges a conspiracy without any basis is likely to look silly when the final results are tallied.
Source: The New York Times
August 22, 2002, Thursday, Late Edition - Final |
