Commentary: The Productivity Boom Is Still A Mystery

Long before he dreamed of studying economics, Alan Greenspan studied box scores. Growing up in New York, the future Federal Reserve chairman honed his statistical skills on the slugging percentages of the Dodger greats. These days, Greenspan may wish economics was more like baseball. The game has changed--but no one revises 1939's batting averages. In economics, the stats get rewritten every year.

In mid-July, Greenspan went out on a limb. He suggested an upcoming revision of gross domestic product data for 1993-96 would back up his belief that productivity is rising faster than official statistics show. He relied partly on an alternate measure of the economy, gross domestic income, or gdi, which since 1993 has grown much faster than GDP. If the official growth rate were revised to match income growth, the healthy '90s would look even more robust--a clear sign productivity is surging.

FLAWED STATS? Greenspan was not alone on his limb. Weeks before the Fed chief's statement, I predicted in these pages that a large GDP revision would show an acceleration in output sparked by corporate restructuring and technology investment (BW--July 7). And Wall Street seems to accept a new era of growing productivity to support currently high stock valuations.

But the data didn't settle the productivity debate. Statisticians bumped up the GDP's 1993-96 growth rate only slightly--from 2.6% a year to 2.7%. They revised down the income measure a bit but still left a major gap: gdi was $60 billion higher than GDP in 1996, and it has grown at a substantially stronger 3.9% rate since 1993.

New Economy proponents argue that the next GDP revision will support their thesis. And it's still possible that the stats are flawed. But "wait 'til next year" isn't going to work. The July 31 figures increase the odds that official economic data will never capture the productivity growth that Greenspan and Wall Street so fervently believe in.

The revisions actually bolster the GDP numbers while leaving income figures shaky. The GDP report is missing only final figures on 1996 services spending, while the income report relies on tax data trickling in for 1995. The income side also uses larger and more dubious fudge factors. Take proprietors' income, a $520 billion element of national income in '96. Internal Revenue Service audits in the 1980s showed that farmers and small-business owners report only half their earnings, so Commerce Dept. statisticians double the figure they get from the irs. But if the agency's increased ability to match income reports and returns is prompting more honesty, the income numbers may be overstated.

DISCREPANCIES. Another example: Many companies have shifted from writing off software the year it's purchased to depreciating costs over several years. That's sound accounting, but it boosts pretax profits and income. "That could explain billions of dollars of discrepancy," says J. Steven Landefeld, director of the Bureau of Economic Analysis.

Details buried in the revisions also undercut some circumstantial evidence for the productivity boom. The earlier GDP numbers didn't show spending keeping up with gains from the surging stock market. This missing "wealth effect" had made the income measure seem more credible. But in the new report, Commerce raised household spending within GDP by $110 billion over 1993-96--which is exactly in line with wealth-effect predictions.

But the biggest puzzle--the productivity paradox--remains. Corporations have poured billions into technology, with no sign of a return on that investment in the productivity stats. It's still likely their bet will pay off someday. But it may be difficult to measure the results with today's flawed data, especially if the productivity boom eventually turns out to be a long, hard slog. In the Greenspan economy, it's singles and bunts--not homers--that will bring in the runs.

Before it's here, it's on the Bloomberg Terminal.