Gross domestic product, job growth, productivity: These numbers have a reassuring solidity. But when it comes to economic data, the first pass can be misleading. The initial numbers are often revised repeatedly over the course of years, sometimes leading to a very different view of the health of the economy.
Look at the last recession. That downturn is now deemed officially to have started in July, 1990. But back then, the data showed an economy that was expanding, not contracting. The Commerce Dept. reported that GDP was growing at a 1.6% annual rate in the third quarter of 1990. Meanwhile, the Bureau of Labor Statistics announced that private-sector employment was up by 240,000 in the six months ended in October, 1990. With that apparent strength, it's little wonder that Fed Chairman Alan Greenspan waited until December to cut the discount rate.
But these gains evaporated as the data were revised over the next couple of years. It turned out that the BLS had greatly overstated the number of jobs created. In particular, the "bias adjustment factor," which accounted for job growth at new businesses, was far too high. It turned out that the number of jobs actually fell by 240,000 over that six-month stretch. And the GDP growth number for the third quarter of 1990 was steadily adjusted downward until, by 1992, it was revised into negative territory.
The most striking change, though, affected productivity data. In the months leading up to the 1990 recession, measured productivity growth appeared to be basically flat. Indeed, Greenspan repeatedly stressed what he called "the weak productivity performance" of the economy in his July, 1990, Humphrey-Hawkins report to Congress. And productivity did not appear to pick up as growth slowed. Output per hour at nonfinancial corporations--one of Greenspan's favorite measures--apparently fell over the course of the recession, according to data released in late 1991 (chart).
But after a decade of revisions to the data, a completely different productivity picture has been revealed: In fact, productivity dramatically accelerated over the course of the 1990-'91 recession as companies chopped jobs. It is reasonable to believe that if Greenspan had had access to the correct numbers rather than the original ones, he would have acted more aggressively.
No one knows how much--or in which direction--the current numbers will be revised. Yet the lesson is clear: Treating economic statistics as gospel is precisely the wrong thing to do.