PLAYING WHAT-IF AT THE FED
A revised model of the economy can factor in expectations
It's September, 1997. The economy is slowing, but President Clinton and GOP congressional leaders have finally hammered out a seven-year plan to balance the federal budget. While politicians are high-fiving at the White House, Federal Reserve Chairman Alan Greenspan faces a tough question: Does he need to cut interest rates to revive the economy? If the financial markets expect Washington will have the stomach to follow through on spending cuts, long-term interest rates will fall--thus making any Fed cuts unnecessary. But if traders discount the budget deal, the Fed may have to act fast to offset a slump.
It's too soon to tell how Greenspan will react when--if ever--such a budget deal is struck. But the Fed now has a better way to analyze such sticky problems. The central bank is using a new computerized model of the U.S. economy that lets policymakers look directly at expectations--the effects that financial markets, businesses, and consumers expect from changes in taxes, spending, and interest rates. With the new system, Federal Reserve officials can play what-if games with potential moves and responses. "The range of policies that we can try out in the model is vastly greater," says Fed Governor Laurence H. Meyer, a leading economic modeler and forecaster before he joined the central bank.
While the Fed's exact predictions are a closely held secret, officials indicate that the new model shows a slowing economy with little need for Fed action at the moment. And some economists argue that the new model's ability to fine-tune expectations will take it further from the real world, not closer.
But the shift from the Fed's old model, rooted in the Keynesian views of the 1970s, to the expectations-based system reflects a powerful current sweeping through the economics profession--and through the world's central banks. The emphasis on market reaction is pushing monetary policymakers from London to Tel Aviv to Canberra to adopt explicit targets for inflation and to spell out how they'll reach those goals. The idea: If the public is convinced that a central bank is determined to bring down inflation, officials can harness those expectations to magnify the impact of their policy moves.
NO PANIC. The Fed hasn't adopted such explicit targets. But many officials, including Greenspan, argue that the Fed's growing credibility played a crucial role during this election year. As the economy roared to 4.7% growth in the spring, Greenspan resisted urgings to hike interest rates to head off an increase in inflation. Instead, he maintained that bond traders--confident that the Fed wouldn't let price hikes get out of hand--would raise rates just enough to slow the economy, without setting off a panicky spike that would halt growth altogether. That forecast was borne out as the economy slowed to 2.2% in the third quarter without any Fed action.
Credibility has always been crucial in matters of money. But in the decades after World War II, policymakers assumed that the economy could be tweaked and fine-tuned through constant adjustments in interest rates, taxes, and spending. That worked--until the late '70s, when "stagflation," a lethal combination of inflation and unemployment, disconnected the fine-tuners' policy levers. At the same time, economist Robert E. Lucas Jr., 1995 Nobel laureate, started demonstrating the importance of "rational expectations." Markets and individuals were able to forecast the economy just as well as the professional seers, Lucas argued, and that would negate the impact of the fine-tuning moves. Consumers would not step up their borrowing and spending in response to a stimulative rate cut, for example, if they were able to see that it was likely to be temporary.
Ever since, economists have been trying to build expectations theories into their forecasting models, with varying degrees of success. The Fed's new model is the most ambitious effort yet in this direction. The model is a set of about 300 complex, interlocking equations that try to describe the economy in mathematical terms. Working with a network of high-powered personal computers, the Fed staff can feed in a variety of conditions--an oil shock, say, or a tax cut--and solve the equations to see how gross domestic product, prices, and wages will respond. A run of the model can take from two seconds to 15 minutes and produces from 4 to 40 pages of charts and tables, depending on how much detail the modelers need.
For simple forecasts, the model closely resembles the Fed's old system. But suppose Congress were debating a proposal by Senator Connie Mack (R-Fla.) to order the Fed to concentrate solely on bringing inflation down to zero. The new model lets the Fed test the mandate's effects on public expectations and the central bank's inflation-fighting credibility. The model shows, according to staffers, that such a mandate would lower inflationary expectations and would mean that the Fed could reach its goal without tightening as much.
Some doubters still believe the whole idea of adjusting expectations is overblown. "It's a myth that central banks can change inflationary expectations by words alone," says Princeton University economist and former Fed Vice-Chairman Alan S. Blinder. The model's new levers for adjusting expectations are "technically innovative--but I would put a stop sign in front of them," Blinder says.
So far, the model's impact on day-to-day policy may be slight. The governors and Federal Reserve Bank presidents who actually vote on policy can be swayed as much by the help-wanted signs they see as by sophisticated econometrics. But the new model, if it works, may help make setting monetary policy more of a science and less of a black art.By Mike McNamee in WashingtonReturn to top