- The Washington Times - Friday, July 28, 2006

The Federal Reserve’s monetary policy decisions are based on the economic forecasts of its Federal Open Market Committee (FOMC), and forecasts are subject to error.

Twice a year, in February and July, the Federal Reserve submits its semi-annual Report to the Congress, which includes the FOMC’s forecasts of economic growth, core inflation, and unemployment for the current and following year. Since forecast errors can translate into policy errors, it’s instructive to compare the committee’s forecasts in recent years with actual outcomes.

The FOMC predictions for economic growth, as measured by real gross domestic product (GDP), and for core inflation, as measured by the personal consumption expenditure (PCE) price index excluding food and energy, are expressed as percent changes from fourth quarter to fourth quarter. The unemployment rate forecast is a fourth-quarter average. The forecasts represent the central tendency of the individual forecasts of the Fed governors and Fed bank presidents who participate in the deliberations of the FOMC.

Monetary policy operates with a lag, with the economic effect of a change in the federal funds interest rate biting harder over time, then weakening. Thus, in assessing the accuracy of the FOMC forecasts, it’s relevant to focus on the committee’s February prediction for the fourth quarter of the same year. In about a year a change in the federal funds rate should have drawn blood. (Fairness decrees the forecasts be compared to the actual outcomes from Fed reports after the forecast period rather than to subsequently revised data.)

The comparisons that follow are for the years 2003 to 2005, for which both forecasts and data on actual outcomes are available. During this period, the federal funds rate changed direction. Up to June 2003 the rate was falling; then it leveled off at 1 percent. In June 2004, the FOMC started increasing the rate by a quarter point at each committee meeting to its current level of 51/4 percent.

In February 2003, the FOMC forecast real gross domestic product to expand by 31/4 to 31/2 percent from fourth-quarter 2002 to fourth-quarter 2003. Actual growth turned out to be nearly a percentage point higher. Core inflation for the same period was forecast at 11/4 to 11/2 percent, and turned out to be 1.4 percent, within the forecast range. The 2003 fourth-quarter unemployment rate was anticipated at 53/4 to 6 percent and came in at 5.9 percent, also within expectations.

In February 2004, the FOMC expected economic growth later that year to pick up its pace over 2003, but it slowed instead. At the same time the committee expected core inflation to slow, but it sped up. Real GDP was forecast to expand by 41/2 to 5 percent between fourth-quarters 2003 and 2004, but the actual increase was nearly a full percentage point less. Core inflation was predicted to rise by 1 to 11/4 percent but came in about a half-point higher. That year’s forecast range of unemployment was again accurate, encompassing the actual 5.4 percent average for fourth-quarter 2004.

Early in 2005, the FOMC correctly predicted economic growth would slow that year and core inflation accelerate. Growth of between 33/4 and 4 percent was expected, though it came in lower, at just over 3 percent. Core inflation was forecast at 11/2 to 13/4 percent and came in at just under 2 percent. Fourth-quarter unemployment was forecast at 51/4 percent, slightly above the 5 percent actual.

In its report in February, the committee projected fourth quarter year-over-year economic growth at “about 31/2” percent, which it revised to 31/4 to 31/2 percent in its July 19 report. The February forecast of core inflation was “about 2” percent, which this month was revised upward to 21/4 to 21/2 percent. The committee expects a slight slowdown in economic growth and core inflation next year.

Economists sometimes put sophisticated forecasts to simple tests. One such test is to see if using last year’s actual outcome as this year’s forecast does a better job of prediction than the sophisticated forecast. Applying that test to the FOMC forecasts for 2003-2005 shows, for economic growth, the “naive” projection was closer to actual growth than the Fed’s forecast two out of three times. For core inflation, the FOMC forecast was better once, worse once and tied once. On unemployment, the Fed forecast was better twice and tied once.

Forecasting the economy is hazardous and errors are inevitable. All in all, the committee didn’t do a bad job, especially considering the forecast period included war, hurricanes, fiscal uncertainty and energy shocks.

It’s fair to say the FOMC’s prediction errors have not led to serious mistakes in monetary policy, at least judging by the economy’s performance the last few years. To its credit, the Fed’s policies of gradualism and transparency have led to what looks like a soft landing for the economy as it approaches its growth potential.

But the real test for inflation will be played out in the second half of this year and in 2007. Since 2004, FOMC forecasts have underestimated core inflation. A continuation of this pattern could be dangerous.

Alfred Tella is former Georgetown University research professor of economics.



Click to Read More

Click to Hide