How Not To Measure The U.S. Economy

When Congress and the White House finally put aside their bickering over the budget and reopen the government, one of the first things the Commerce Dept. will do is issue sweeping changes to its flagship measure of economic performance, gross domestic product. The new data attempt to give a more accurate picture of GDP, and in the new Information Age, we applaud any government effort to provide better information. Problem is, the new GDP benchmark, at least for now, lowers economic growth rates, moving us further away from the truth instead of closer to it.

Commerce's Bureau of Economic Analysis is trying to remove a long-recognized flaw in the way we compute GDP. It's something economists arcanely refer to as "substitution bias," which arises because the current fixed-weight GDP overstates the growth impact of sectors where prices are falling, such as computers. The revision, called the chain-weighted GDP, reflects price changes more quickly and continuously, lessening the influence of these rapidly expanding sectors on growth. However, the effort to eliminate that distortion is only one of many projects under way at the BEA with the intent of moving toward a better GDP measure. The correction for substitution bias, taken by itself, yields dramatically slower growth figures for recent years. For example, growth in the first three quarters of 1995, now on the books at 2.7%, drops to a meager 1.8%.

To aggravate matters, the Kansas City Federal Reserve now says the economy's new noninflationary growth limit has dropped from the generally accepted 2.5% down to only 2%. Can these numbers really be true? The world seemed dismal enough when economists calculated that the U.S. economy could grow at only 2.5% without generating inflation.

Worse still, the new chain-weighted index of GDP nearly eliminates all of the previously reported strong productivity gains in the 1990s. Yet in many respects, today's economy is reminiscent of the high productivity of the 1960s, when the economy grew at a 4.4% annual rate using the chain-weighted method. Simple observation of today's low inflation, strong corporate profits, low unemployment, and nascent gains in real wages clearly indicate that productivity growth is vigorous. No wonder the latest statistics will leave policymakers and corporate chieftains more confused than ever about the real state of the economy.

The BEA is giving us only half a measure. What's still missing from a better tracking of GDP is an upward adjustment for changes in the quality of various high-tech sectors other than computers, such as telecommunications and semiconductors. Moreover, the BEA's measures of output in many service industries are still woefully inadequate. Most economists agree that modifications in these areas will push up the GDP growth rate. The BEA plans to incorporate some of these adjustments into its next GDP overhaul, which is scheduled for mid-1996. So why not wait and give us the whole pie instead of giving us a measure that eliminates one distortion while creating others?

U.S. economic data are unquestionably the best in the world, but this GDP fiasco needlessly tarnishes that hard-earned reputation. True, budget constraints have limited the ability of government statisticians to upgrade their product. Yet budget limits cannot justify poor decision-making. We all want better data, but don't jerk us around in the process.

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