By Robert J. Samuelson THE GREAT BOOM of the 1990s was barely anticipated. The same was true of other upheavals: sporadic “energy crises,” the sharp rise of inflation in the 1970s, its dramatic fall in the 1980s and various shifts in productivity growth. At present, most forecasts are bravado and bluff. Hardly a day passes without the government or private industry disgorging some economic statistic that is instantly seized upon to demonstrate that (a) “the worst has passed” or (b) “the economy is tanking.” In the first quarter the economy grew at a 2 percent annual rate—better than expected. A good omen. But in April unemployment rose to 4.5 percent—worse than expected. A bad omen. Also, productivity (output per hour worked) dropped for the first time in six years—another bad omen. On the other hand, stocks have recovered from recent lows—a good omen. The truth is that no one knows. Even the brightest mortals cannot peer far into the future. (Alexis de Tocqueville once wrote of the French Revolution: “Never was such a great event ... so well prepared and so little foreseen.”) As recently as last October the “consensus forecast” of 52 economists surveyed by the Blue Chip Economic Indicators—a newsletter—was for a strong 3.5 percent growth in 2001. The “consensus” is now down to 2 percent, and though few economists yet predict an outright recession (usually defined as two consecutive quarters of falling output), this clearly is a possibility. The errors have not reduced economists’ eagerness to prognosticate—only their credibility. Their picture of the future usually reflects the recent past, because that’s what they know. They underpredicted inflation in the 1970s because it had been low in the ’60s and they expected it to stay low. They overpredicted inflation in the 1980s because it had been high in the ’70s and they expected it to stay high. This backward vision also explains the optimistic bias of today’s forecasts. Because the economy thrived in the late 1990s, economists expected it to continue thriving. Perhaps the trophy for optimism goes to Merrill Lynch’s economists. “We’re sticking with our view that the U.S. economy will avoid an outright recession,” they wrote recently. “Payrolls [payroll jobs] fell for the second consecutive month in April, plunging by 223,000. There have almost never been two consecutive payroll declines that didn’t mark a recession. But the Fed has never eased [interest rates] as rapidly as it is currently doing, which is why we still think a contraction can be avoided.” We’ll see. As a rule, forecasters rely on computer models that try to predict how, say, a rise in consumer spending might affect profits, investment and employment—and how these changes might feed back into inflation, stock prices, interest rates and (again) consumer spending. The idea is that present behavior reflects past behavior, as reflected in various economic statistics. This is not as simple as it sounds. Not only are statistics incomplete and imperfect, but behavior often changes in erratic ways. Every business cycle creates new experiences and expectations that alter how people and businesses think and act. What was true a year ago or a decade ago may no longer be true. Optimism and pessimism feed on themselves. “In long business [expansions], there will be a lot of investment—and then overinvestment,” says Victor Zarnowitz, a retired economist from the University of Chicago and a respected student of business cycles. Certainly that’s the case now—computers and telecommunications equipment being obvious examples. One well-known forecasting service recently conceded that its “equation for investment in computer software and hardware” had underpredicted actual investment by between $35 billion and $60 billion. Worse, forecasting models exclude almost everything interesting and disruptive in life: politics, nationalism, technological change, the weather, greed, fear, ambition, ignorance and stupidity—to name a few omissions. Naturally, blunders occur. To explain and excuse these lapses, economists often blame “shocks.” A “shock” is a catchall label that covers almost anything that the model misses and that spoils the forecast. Oil and energy “shocks” (big shifts in prices) are a common variety. So are food-price “shocks.” In 1998 some economists cited a “demand shock”—an unpredicted spurt in consumer spending—that probably prevented a recession. “Shocks” contradict the premise of the models, which is that economic change is gradual and comprehensible. Most of the time it is. This is why economists and models seem right more often than not. Today is usually like yesterday, which was like the day before. You too can make a credible forecast. Just read the newspapers and make a few minor adjustments to existing economic indicators. The chances of being wildly wrong are slight; inflation won’t jump from 2 to 20 percent in a few months. Because this is true, economic forecasts cluster together. The clustering also reflects herd behavior. To stray too far from your peers is to risk looking foolish—alone. There’s less danger in being wrong with everyone else.
We call this exercise forecasting, but of course it isn’t. It’s telling people what they already know or might know by examining the available information. It creates an illusion of understanding. The trouble is that there are times when radical and dramatic changes do happen, and at these moments economists are almost as clueless as everyone else. They overlooked the onsets of the 1981-82 and 1990-91 recessions and, conceivably, may be repeating the mistake today. The larger paradox is that economic forecasts are least reliable when they are most needed. |
