November 2001
Federal Reserve Bank of Dallas

Down But Not Out: The U.S. Economy After September 11

The terrorist attacks of September 11 have profoundly affected the well-being of U.S. citizens. Our sense of invulnerability is gone. Comparable events are the Arab oil embargo of October 1973—which challenged our assumptions about the continued availability of abundant, cheap energy—and the October 1957 Sputnik launch—which raised fears of intercontinental missile attack. Both of those shocks triggered important changes in spending priorities. Both hit a U.S. economy that had already been slowing. Both were accompanied or promptly followed by recessions.

We can never know, with certainty, how the economy would have evolved had the Sputnik launch, the oil embargo or the September 11 attacks never occurred. Such events are rare, and each has unique aspects. Moreover, our understanding of the terrorist threat and the measures necessary to combat it is still developing. With this caveat in mind, it appears that the September 11 terrorist attacks subtracted perhaps 1 percentage point from annualized third-quarter GDP growth, making what would have been a small, positive number small and negative. Spillover from the attacks makes a much more significant GDP decline likely in the current quarter. In contrast, the outlook for the first half of 2002 has been little-affected. Unfortunately, that outlook calls for output growth so sluggish that jobs will shrink and the unemployment rate will continue to rise.

Pre-Attack Trends
Figure 1 nicely summarizes the economic situation we were facing leading up to the attacks. Consumer spending decelerated early last year, but continued to increase right through August 2001. Industrial production kept rising, unabated, until September 2000, and fell more or- less steadily since. Obviously, output cannot contract indefinitely in the face of rising consumer demand. Consumer demand cannot expand indefinitely if firms continue to cut production and jobs. One or the other of these trends was going to have to give way.

There were hints, at least, that industrial production might soon stop falling. In early August and early September surveys by the National Association of Purchasing Management (NAPM), more manufacturers reported increases in orders than decreases (Figure 2). The Conference Board’s composite leading index was also signaling improvement. Most analysts were calling for a modest pick up in GDP growth during the third quarter and a further increase in the fourth quarter.

Monetary policy played an important role both in slowing demand growth back in the second quarter of last year and in maintaining positive demand growth in the face of a rising unemployment rate in 2001. Judging by the inflation-adjusted, or "real" federal funds rate, monetary policy tightened from early 1999 through the middle of 2000, and has eased almost continually during 2001.

For evidence that monetary policy still packs a punch, one need only look at the construction-materials and consumer-durables manufacturing industries, two important interest-sensitive sectors. In both, new orders topped out in early 2000–less than a year after the Federal Reserve began to raise short-term interest rates, and coincident with the peak in real rates. In both sectors, demand growth resumed quickly once the Fed began easing in 2001 (Figure 3).

Was the policy tightening in 1999 and early 2000 a mistake? Inflation statistics released over the past 2½ years suggest not. Inflation as measured by the GDP price index, for example, increased by a full percentage point during 1999 and 2000 before leveling off (Figure 4). While one can quibble over the exact timing of the Fed’s moves, tightening appears to have been necessary to hold inflation in check.

Supply-Side Impact of the Attacks
What was the likely impact of the September 11 attacks on the economy’s capacity to produce goods and services? A good place to start is with the effects of a natural disaster like the 1994 Northridge quake in Southern California.

Figure 5illustrates how the level of output is typically affected by a Northridge-style event. The figure assumes that output has been rising at a more-or-less steady pace and is expected to continue to do so in the future (as indicated by the dotted red line). Instead, disaster strikes, causing output to drop sharply. The level of output remains depressed for a time, but as damaged homes and factories are rebuilt and damaged furnishings and equipment are replaced, output growth is elevated. (See the dashed blue line.) The economy is soon back on its pre-disaster path.

Although the events of September 11 fit the natural disaster mold in many ways, they are also reminiscent of the 1973 Arab oil embargo and the 1957 Sputnik launch. Like these earlier events, the attacks brought previously unappreciated, continuing risks to the public’s attention. It’s as if we not only experienced a damaging earthquake on September 11, but also discovered a whole network of fault lines beneath our major cities.

Consequently, we are likely to see a larger and more sustained shift of resources than would typically follow a natural disaster—instead of simply rebuilding we must build anew, differently from before. Unfortunately, it takes time to plan new factories and train workers in new skills, so layoffs, closings and bankruptcies will initially dominate the headlines and the statistics. Instead of the immediate, strong boost to growth that occurs during the recovery from an earthquake or hurricane, we end up with an output path that looks like the solid blue line in Figure 5.

Note that output never quite makes it all the way back to its original path. That’s because going forward we will have to sacrifice efficiency gains for the sake of enhanced security. For example, firms may hesitate to consolidate their operations or rely on foreign parts suppliers. A larger military budget will take resources away from the private sector. Despite our efforts, some future terrorist attacks may succeed.

The actual and prospective destruction of capital, the disruption associated with resource reallocation, and the prospect of higher military and security spending all make households financially worse off by lowering asset values and by reducing their future after-tax earnings.

The evidence suggests that, given a constant real interest rate, consumption shifts sharply downward in response to a decline in wealth or earnings prospects. There’s the rub. For in the wake of September 11, new investment projects will not get under way immediately, and military and security spending will take time to ramp up to their new, higher levels. Any sudden decline in consumer spending may consequently cause a shortfall in aggregate demand.

To mitigate this potential problem the Fed can lower real, short-term interest rates by enough to induce households to scale back their spending plans gradually, rather than all at once. As military, security, and investment spending pick up, monetary policy will need to reverse course and raise short-term real interest rates to normal or even above-normal levels. Getting the timing of this switch right will be the major monetary-policy challenge in the year ahead.

How Bad Is It Likely to Get? When Will Growth Resume?
Just how big a hit is the U.S. economy likely to take from the September attacks? When will their impact begin to fade and growth resume? Two forecasting tools developed here at the Dallas Fed can help answer these questions.

The first tool is a forecasting equation for current-quarter GDP growth. Official GDP growth estimates don’t come out until a full month after the end of each quarter. Our forecasting equation provides us with a GDP prediction a month-and-a-half earlier than these estimates. The forecast is based on monthly employment, industrial production, and retail sales figures for the first two months of the quarter. The forecasting equation’s unique feature is that it is estimated using only data that were actually available at the time, instead of data that have gone through many rounds of revisions. The resulting performance is superior to that of the average professional in the Blue Chip survey of forecasters.

Based on monthly data through August, our model forecasted 0.7 percent GDP growth in the third quarter. Actual third-quarter GDP growth came in at -0.4 percent according to the Commerce Department’s "advance" release. So, our best estimate is that the September 11th attacks subtracted about 1 percentage point from third-quarter growth, turning a small positive number into a small negative number.

The impact of the terrorist attacks on third-quarter GDP growth would have been even larger had the attacks taken place in July or August instead of September. An extreme example illustrates the point. Suppose that the attacks had occurred on the very last day of September. Then the average level of output in the third quarter would have been hardly affected, and third-quarter GDP growth–which compares the average third-quarter level of output to the average second-quarter level–would also have been hardly affected. Instead, we would have seen weak fourth-quarter GDP growth.

Well, the 11th of September isn’t quite the very end of the quarter, but it’s pretty close to the end. So if the direct impact of the attacks subtracts 1 percentage point from third quarter growth, they are likely to subtract roughly 3.5 percentage points from fourth quarter growth. This timing story helps explain why most private forecasters are calling for a moderate decline in fourth-quarter GDP instead of a moderate increase.

Our second tool is an equation that forecasts future employment growth using financial-asset and oil prices. Financial-asset prices are available daily and are not subject to revision. Because they reflect investors’ expectations, they often provide the earliest warnings of changes in the economy’s direction. Although they are often individually unreliable, false signals often cancel one another out when several indicators are considered as a group.

The first indicator we use to forecast employment growth is the junk-bond spread, equal to the difference between the returns on high-yield and AAA-rated corporate bonds. It measures the risk that marginal borrowers will default on their loans. The spread widened markedly in September and rose further in October, to its highest level since the end of last year (Figure 6). Bond investors are clearly concerned that the economy will be weak in the months ahead.

Stock prices are another important (but not very reliable) indicator of future employment growth. As of September 10, the Standard and Poors 500 index was down 28 percent from its all-time high in March 2000. At its postattack low, it was down almost 37 percent (Figure 7). However, as of this writing the index has recovered its September 10 level. So the stock market’s signals, although not encouraging, are no worse than before the attacks.

Oil-price increases both disrupt the economy and act much like a tax hike imposed by oil exporters. Oil prices initially rose following September’s attacks, but have since fallen substantially (Figure 8). Unfortunately, since the economy responds to oil prices with a long lag, the residual effects of the relatively high prices of 2000 and early 2001 will remain a drag on growth in 2002.

The only indicator that is giving us a positive signal about future employment growth is the real short term interest rate. It fell sharply in the first half of the year, as the Fed aggressively eased monetary policy, and fell sharply again following September 11 (Figure 9).

Figure 10 shows actual employment growth (in red) along with a forecast made 9 months earlier (in blue). Forecasts are calculated using the four indicators discussed above. You can see that earlier this year, the forecasting model was predicting essentially zero job growth in late 2001 and early 2002. However, in August, just before the terrorist attacks, forecasted employment growth turned sharply negative. Employment growth forecasts calculated in September and October were also negative. The most recent (October) forecast indicates that jobs are likely to decline at a 0.7 percent annual rate over the first six months of 2002. So, the terrorist attacks didn’t make the early 2002 outlook any worse than before, but that outlook wasn’t bright to begin with. Although job cuts will not be so great as to keep GDP growth negative, they will drive the unemployment rate up to about 6.0 percent by June.

Summary and Concluding Remarks
There were conflicting trends in production and sales prior to September 11, with production falling despite rising consumer demand. Sooner or later, one of these trends had to give way, and there were encouraging signs that production might soon bottom out. The attacks had a mild, negative effect on third-quarter GDP, turning a weak increase into a small decline. We’ll see a bigger negative impact in the fourth quarter statistics. The already bleak growth outlook for the first half of 2002 hasn’t really changed very much, however. We’re likely to see output rising, but too slowly to prevent further increases in the unemployment rate.

There are several risks to these forecasts. For example, we may see major new terrorist acts attempted or political upheaval abroad. A less obvious risk is that the Fed will "get behind the curve" much as the Japanese central bank did in the 1990s, and lower interest rates too slowly to keep up with declining inflation expectations. The October University of Michigan survey of households shows a sharp fall in expected inflation that bears watching.

On the plus side, the Fed has demonstrated a willingness to act quickly and boldly when economic developments warrant it. Policy has proven itself to be effective, first by slowing consumer spending growth in 2000, and then by sustaining it in the face of rising unemployment during the first eight months of 2001. By the spring of 2002, the economy will benefit from the additional stimulus that the Fed has added to the pipeline since September 11. Tax incentives designed to kick start investment spending seem all but certain. Finally, no other economy can so quickly shift resources from shrinking to expanding industries.

We’re down, but not out. Brighter days lie ahead.

—Evan F. Koenig

About In Depth This article is based on a presentation by Evan F. Koenig, vice president and senior economist, Research Department, Federal Reserve Bank of Dallas. The views expressed are those of the authors and do not necessarily reflect the positions of the Federal Reserve Bank of Dallas or the Federal Reserve System.