David E. Runkle - Senior Economist
Published December 1, 1992 | December 1992 issue
This article is an excerpt from the Fall 1992 Quarterly Review, published by the bank's Research Department.
Last year, we predicted that the first year of the current recovery would be unusually weak because consumers were pessimistic about their long-run economic prospects. We were right. Consumer spending was exceptionally low in 1992, and following the weakest start to any recovery during the past 45 years, growth has remained below its postwar average.
Now, we predict that slow economic growth will continue for at least the next two years because consumers are still pessimistic and nearly all other areas of the economy are weak. This slow growth prediction seems quite reasonable, based on the recent performance of the Fed's forecasting model and the evidence suggesting that the economic problems we stressed a year ago appear unlikely to be solved soon.
This recovery has proven to be weaker than the previous seven recoveries by almost every measure.
Table 1 [tables and charts are printed in the publication] compares the performance of some key economic indicators during the first year of the current recovery with their average performance during the first year of other recoveries since 1945. From the beginning of the third quarter of 1991 to the end of the second quarter of 1992, real gross domestic product (GDP) grew only 1.6 percentabout one-fourth of its first-year average growth rate during other recoveries.
But the true weakness of this recovery can be seen by noting that every indicator in Table 1 showed lower growth in the first year of this recovery than in the typical recovery. Investment grew by only half its average rate. Consumption grew at less than one-third of its average rate. Employment grew by less than one-twentieth of its average rate. And instead of growing 4.4 percent, government spending actually declined.
Comparing this recovery to the average recovery doesn't fully reveal the current poor performance of the economy, however. Not only has this recovery been far below average, by most measures this has been the weakest recovery in 45 years. While real GDP grew only 1.6 percent in the first year of this recovery, it never grew less than 3.5 percent in the first year of any other postwar recovery. Consumption, investment and employment growth were also well below their lowest recorded growth levels in previous recoveries.
When compared to other recoveries, the current recovery certainly appears to be weak. But this recovery seems weaker still when we compare it to average economic growth. Economic growth is typically fastest at the beginning of a recovery. But this time growth in the first year of the recovery was below its average during the postwar era. Over the last 45 years, real GDP has grown at an average rate of about 3.3 percent. However, there has not been a single quarter since the end of 1988 when real GDP grew above its average rate. This means that the current recovery is just a continuation of a long period of below-average growth.
How long will this below-average growth last? The Fed model predicts this period of below-average growth will continue until at least the end of 1994. Another recession is not predicted, but the Fed model predicts that real GDP growth will remain well below the postwar average.
Table 2 shows the 1993 and 1994 forecasts for several important economic variables made by the Bayesian vector autoregression (BVAR) model used at the Minneapolis Fed. The table also shows, for comparison, the average values for those variables since 1948.
The model predicts that real GDP will grow at an annual rate of 2.3 percent in both 1993 and 1994, substantially below its average growth since 1948. It also predicts below-average growth over the next two years for three major components of GDP: consumption, business fixed investment (that is, investment in equipment and buildings), and government purchases of goods and services. Only residential investmentthat is, spending on the construction of new housesis predicted to grow faster than its average growth rate.
One piece of good news this forecast offers is that inflation is predicted to remain low over the next two years. Both the consumer price index and the GDP deflator are predicted to grow by around 3 percent in 1993 and 1994, well below their average annual growth of around 4 percent to 5 percent.
Two kinds of evidence give credibility to the model's prediction. First, the model was fairly accurate in predicting how weak the first year of this recovery would be. Second, the long-term economic problems that provide support for the model's predictions are not likely to be resolved soon.
Last year at this time, the model predicted that the first year of the current recovery would be much weaker than normal. That forecast turned out to be fairly close to the mark and gives credence to the model's current forecast that below-average growth will continue for another two years.
Support for the model's recent successful prediction of continued slow growth also comes from understanding several long-term problems facing the economy. Constrained consumption growth is probably the most crucial one, since consumption makes up the largest percentage of real GDP. Consumers' pessimism about prospects for income and employment growth, as well as demographic trends, will result in continued slow consumption growth.
Other long-term problems include the oversupply of commercial real estate, which is likely to hold down investment spending for years to come, and budget crunches at all levels of government that will likely slow the growth of government spending. Hobbled by all these long-term problems, growth in the U.S. economy seems unlikely to break free any time soon.
Consumer spending usually provides a big percentage of the boost to the economy in any recovery. The reason is that consumption comprises about two-thirds of real GDP. Total consumption growth can be split into growth in consumption per employee and growth in employment. Both are likely to remain low for the next two years.
Growth in consumption per employee is likely to be low because people are pessimistic about their long-run economic prospects, and growth in employment is likely to remain low because of demographic trends. Of course, at around 2 percent per year, slow consumption growth is exactly what our model predicts.
We first suggested last year that consumer pessimism about future economic conditions would restrain economic growth. A year later, consumers' behavior has not changed much. Their responses to surveys, their unwillingness to buy and their desire to reduce their debt burden all indicate continued pessimism.
One indicator of how consumers view the future comes from the University of Michigan's Index of Consumer Sentimenta monthly poll taken to find out consumers' attitudes toward making different kinds of purchases. In October 1992, that index was 73.2, which is lower than it was during most of the recent recession.
But the true indicator of how pessimistic consumers are, as we argued last year, can be found in how willing they are to spend. The permanent income hypothesis, a theory which suggests that people base their current consumption decisions on their expectation about their long-run income, rather than their current income, offers one explanation. If consumers are optimistic about the future, the theory suggests, they will be willing to spend based on their expectations about the future, even if their incomes are growing slowly today. If consumers are pessimistic about the future, however, not only will they be unwilling to spend much now, they will also want to reduce the amount of debt they hold.
Judging their behavior against this theory, consumers appear persistently pessimistic. Real consumption has increased at an annual rate of only 0.7 percent since the beginning of 1989by far the longest period of slow consumption growth in the last 30 years. And real consumption spending per employee has grown at a rate of only 0.5 percent since the beginning of 1989.
Consumers' willingness to take on debt provides an additional indicator of their pessimism. If consumers become pessimistic about the future, they will want to reduce their debt burden because they will no longer believe that future increases in income will allow them to pay their debts without reducing their consumption. Chart 1 shows that consumers rapidly built up their installment debt in the 1980s, and they have cut back that debt since 1989. That cutback in debt is an important indicator of consumers' pessimism. If consumers continue to be pessimistic, they will probably continue to reduce their debt for some time to come.
Clearly, consumers are pessimistic. But is their pessimism reasonable? Two factors suggest that it is.
First, employment growth has been very low during the first year of this recovery. Employment increased by only 0.2 percent in that first year. Income growth has also been very low. Real disposable income grew by only 2 percent in the first year of the recovery, far below its lowest growth in the first year of other recent recoveries.
And real disposable income has grown at an annual rate of only 0.9 percent per year since the first quarter of 1989. So real disposable income per employee has grown at an annual rate of only 0.7 percent since the first quarter of 1989.
But slow growth in income and employment are not the only grounds for consumers' pessimism. A second factor is the increase in permanent layoffs. The increasing trend toward permanent layoffs has caused consumers to be more uncertain about their future employment prospects and thus more pessimistic. Even as the recovery started, permanent layoffs increased. The percentage of layoffs that were permanent hit a new high of 78 percent in the second quarter of 1992.
Even if consumers weren't pessimistic, total consumption growth would remain low over the next few years because fewer people will be entering the work force and forming households. We suggested last year that much of the consumption growth of the last 20 years was caused by an increase in employment as women and baby boomers entered the work force. As employment grew, so did total consumer income and spending. But now there will be relatively few potential new employees, which means that growth in total consumption will remain low, even if growth in consumption per employee is close to its historical average.
Over the past 20 years, employment grew much more rapidly than did population in the United States. From 1969 to 1989, employment grew at an annual rate of 2 percent, while population grew at an annual rate of only 1 percent. Of course, it is impossible for employment to grow much faster than population forever, and there are reasons to believe that employment growth will be slower over the next two years than it was over the last 20 years.
One reason employment grew so quickly relative to population in the United States over the past 20 years is that the working-age population grew more quickly than the population as a whole. As baby boomers reached working age, employment naturally grew faster than population because of the increase in the fraction of the total population that was of working age.
But labor force participation is unlikely to rise much in the near future, because it is already at quite high levels. Over the past 20 years, the movement of women into the labor force increased labor force participation and caused much of the total employment. Labor force participation among women will probably not continue to rise as rapidly as it has during the past 20 years, because women's labor force participation is already high. In fact, the Bureau of Labor Statistics estimates that the annual growth in women's employment will drop from 2.8 percent in the past 15 years to 1.6 percent over the next 15 years.
If both the growth in the working-age population and labor force participation stay well below their average rates during the past 20 years, then employment growth will certainly be below average.
Consumption is not the only component of GDP with long-term problems. The commercial real estate component of GDP has a serious oversupply problem; however, its influence on real GDP growth will be smaller, since its fraction of total GDP is smaller.
A huge commercial real estate spending boom occurred in the
mid-1980s, followed by a bust that continues today. Starting in 1982, there was an explosion in commercial mortgage lending, but the real value of commercial mortgages outstanding has fallen steadily in the last few years.
One consequence of the commercial real estate bust is that there will probably be little investment in business structures for at least several more years. Real investment in business structures is now about 25 percent below its peak. Such investment dropped by 8 percent in the first year of the recovery alone. And the number of square feet of business structures that were completed fell by 11 percent during the first year of the recoveryits lowest level in 30 years.
Recent surveys indicate that the vacancy rate for both industrial and downtown office properties are at historic highs and that net absorption of office spacethe change in the total number of occupied square feethas fallen to 10 percent of its rate during the second quarter of 1987. The dismal prospects for commercial real estate are reflected in the model's low forecast of growth in business fixed investment over the next two years.
Along with consumer spending, government spending typically provides a boost to the economy at the start of a recovery. But with well-publicized fiscal problems at all levels of government, it is unlikely that government purchases of goods and services will provide any economic boost to the economy. Indeed, the model predicts that real government purchases will increase by only 0.6 percent in each of the next two years, well below its postwar average of nearly 4 percent per year.
The big constraint on government spending will come from lower growth in state and local spending. Over the past eight years, real state and local spending has grown at an annual rate of 3.5 percent - far faster than either real GDP or real disposable income. Since most state budget plans were based on the assumption that tax revenues would continue to grow at a high rate, most states now have serious financial crises.
Most state budget surpluses are now at their lowest level in 15 years, and the National Governors Association predicts that state spending will grow below the rate of inflation in 1993only about one-third as fast as it did in the 1980s. The Governors Association also predicts that states will trim employment nearly 2 percent by the end of 1993.
Along with the three major long-term problems afflicting the U.S. economy, other components of real GDP may worsen the already dismal outlook for growthat least in the near term. Although the Fed model predicts export growth will be well below its level of the late 1980s, export growth could be even lower because of world economic conditions. Many of our major trading partners have had slower real GDP growth in recent quarters than the United States has. Since their growth is the major determinant of our export growth, U.S. export growth could be quite low in the near future.
One of the only bright spots in the model's forecast is a predicted increase in residential construction of around 3 percent over historical averages for the next two years. Demographic trends may well dilute this predicted increase, however, since the potential first-time home buying populationmostly those 25 to 35 years oldhas greatly declined.
Both the model's predictions and problem areas in the U.S. economy clearly signal real GDP growth will be hobbled for at least the next two years, and there appears to be no relief in sight.