A Mess of Misleading Indicators

By Daniel Benjamin

Statistics are the heavenly bodies of economics. Not only are they used for navigation by businesses, policy planners and researchers but they also exert a powerful pull over the tides of the economy. A high monthly figure for the trade deficit, for example, can send floods of money rushing out of the stock market in a sell-off. The sheer quantity of statistics available is immense. Nearly every business day the U.S. Government releases one indicator or another, from the Consumer Price Index and capacity utilization to retail sales and housing starts. Too often, however, the overall impact of the numbers is to generate confusion and anxiety. Some of the statistics are subject to repeated revisions. Other gauges fluctuate so wildly from month to month that they seem almost useless. More and more, the art of economic planning appears to be degenerating from astral navigation to something closer to astrology.

The most recent example of statistics gone awry involved the gross national product, the broad measure of the country's output of goods and services. On April 26 the Commerce Department announced that the GNP grew at a moderate annual rate of 2.3% in the first quarter of 1988. Experts interpreted the figure as proof that the economy was running smoothly. A month later, Government statisticians boosted first-quarter GNP growth to 3.9%, a change of nearly 70%. Suddenly, investors had reason to fear that the economy was overheating and that inflation was in danger of accelerating.

The reason for the sharp revision was simple: the first GNP figure was based on incomplete information. The Government reports a quarter's GNP after data for only the first two months are available; the statisticians make educated guesses about what happened during the third month. In this instance, an unexpected surge in exports in March led to the radical change in the calculation of the growth rate.

Such dramatic revisions in the GNP, retail sales figures and other important statistics occur with maddening regularity. "The numbers are not bad," says Sidney Jones, professor of public policy at Georgetown University. "They are just premature." But the Government is under pressure from anxious investors and executives to report economic data as soon as possible. Observes Robert Ortner, the Commerce Department's Under Secretary for Economic Affairs: "If you want something quickly, you give up something." In this case, accuracy.

Some of the Government's economic compasses may be poorly constructed as well. The index of leading economic indicators, a compilation of statistics designed to predict the direction of the economy, is frequently derided as the "index of misleading economic indicators" because of its uneven forecasting record. After last October's stock-market crash, the index declined for three consecutive months -- normally a strong sign that a recession is on the way. An upward revision in the December figure, however, broke the downward streak, and fears of a recession evaporated.

The most important statistic released last week was the unemployment rate. After dropping from 5.6% in March to 5.4% in April, it bounced back to 5.6% in May, leaving economists mystified about the trend. While the jobless rate is generally considered to be among the most accurate of the figures the Government puts out, economists argue over how to interpret the number. A 5.6% unemployment rate sounds fairly bad in an absolute sense, but some experts say that it overstates the degree of distress because a large number of those listed as out of work are people who are voluntarily moving from one job to another. According to this line of reasoning, 5.6% unemployment is actually close to "full employment," and any attempt to push the rate much lower will cause inflation to accelerate. Other analysts say the unemployment rate understates the problem because it does not include the so-called discouraged workers, who have given up looking for a job.

One of the most volatile statistics is the monthly report on the U.S. trade deficit. That figure jumped from $12.4 billion in January to $13.8 billion in | February, only to plunge to $9.7 billion in March. Part of the reason for such swings is that trade flows vary according to seasonal patterns. When the Commerce Department announces the April figure next week, the number will be "seasonally adjusted" in an attempt to smooth out temporary fluctuations.

While the Government is striving to improve statistical accuracy, the effort has been repeatedly undermined by budget constraints. Federal funding for the compiling of statistics has fallen from $1.7 billion in 1980 to $1.6 billion in 1987, even though the cost of gathering data has gone up. The Administration wants more money for the job, but as Congress struggles to shrink the budget deficit by cutting spending, the chances seem slim that something as unglamorous as statistics will survive the ax.

That disturbs the experts who rely so heavily on Government data. Says Thomas Juster, an economics professor at the University of Michigan: "This kind of activity doesn't involve a major cost. It's a small-potatoes operation in terms of what the Federal Government does. But it also doesn't have any political attraction to the general public." Complains Henry Kaufman, the famed financial forecaster and former chief economist of Salomon Brothers: "The Government has not given a high enough priority to improving the quality of compiling economic data. We really do not cherish people who are in the business of collecting statistics." Unless funding is increased, the Government may find it necessary to eliminate some of its statistical measures -- and business planners will have even fewer stars to fly by.

With reporting by Bernard Baumohl/New York and Gisela Bolte/Washington