October 2002
Federal Reserve Bank of Dallas

The National Economy: Where Is It Headed?

By all appearances, the 2001 recession, which began in March of last year, ended last November, at which point the U.S. economy entered into an expansion. In this presentation, I hope to provide you with a snapshot of where that expansion currently stands and where the economy may be headed.

Overview
I’ll begin by discussing the weakness of the expansion. Is this weakness surprising? I’ll try to show that, in terms of output growth, the weakness is largely in line with the mildness of the 2001 downturn, but that in terms of employment growth, the recovery has been sluggish, suggesting a repeat of 1991’s "jobless recovery."

I’ll then examine what looks to be a cooling in the pace of recovery on the economy’s production side, since July. This slowing has raised fears of a possible "double dip" recession. Is the economy tipping back into recession? While the late-summer developments are clearly troubling, we’ll see that the data don’t yet support an imminent second dip.

With production cooling, much will depend on the continued health of demand, and it is here that most of the positive news has been of late. Investment is showing some flickering signs of life, and consumers continue to spend, albeit at a moderate pace. Looking forward, we’ll want to ask whether these sources of demand have sufficient momentum to pull the economy along.

I’ll conclude with inflation, where the marked acceleration in consumer price inflation which began in 1999 and continued through last spring, has been replaced, suddenly, with a marked deceleration, to the point where concerns about deflation have begun to be voiced from some quarters. We’ll want to ask whether those concerns are warranted.

The Weakness of the Expansion
While the National Bureau of Economic Research is the final arbiter when it comes to dating recessions and expansions, every indication is that an expansion began in November of 2001.

Figure 1 shows three coincident indices of economic activity—from the Conference Board, the Economic Cycle Research Institute and academic economists James Stock and Mark Watson of Harvard and Princeton Universities. Coincident indices attempt to measure the overall level of economic activity, and we can see in the figure that all three hit bottom in November—shown by the vertical line—after which point they begin to grow.

As we all know, the pace of output growth in this recovery has been slow. The economy is currently on track for 2.5–3 percent growth over the first year of the recovery. By comparison, the average for the post–World War II period is about 6 percent. As we’ll see, however, the slow rate of output growth is consistent with the mildness of the 2001 downturn.

It is an empirical regularity that mild recessions make for weak recoveries (and, conversely, that deep recessions make for strong recoveries). Milton Friedman dubbed this fact the "guitar string" theory of business cycles: the smaller the pluck downward, the weaker the snap back. By most measures, 2001 was a very small pluck; hence we shouldn’t expect a sharp snap back. Measuring the severity of the downturn by the percent decline in GDP from its peak to its trough, the 2001 recession ranks as the second mildest of the post-War period, and is nearly tied for mildest.

The scatter plot in Figure 2 illustrates the "guitar string" theory. Each point corresponds to a recession/expansion episode. The horizontal axis measures severity of decline, as the percent fall in GDP from its peak to its trough, so that points further to the right correspond to deeper recessions. The vertical axis measures strength of rebound as the annualized rate of GDP growth over the first nine months of output expansion; points which are higher up on the graph correspond to stronger recoveries. The upward-sloping pattern of the data is the "guitar string" idea: the more severe the recession, the stronger the recovery.

The yellow star gives the position of the most recent recession/expansion episode, and while the guitar string relationship is looser for milder recessions, output growth so far does not seem greatly out of line with the historical pattern.

What has been surprising has been the sluggishness of employment growth. Private payrolls continued to fall for several months after the overall economy turned a corner, and are still below their November 2001 level. Since April—the point where employment looks to have turned its corner—they are up a mere 75,000 jobs.

Could this simply be the "guitar string" theory again? The answer is no. Employment growth has been low, even given the mildness of the downturn.

Based on the historical relationship, and given the size of the fall in employment, we would expect payroll growth of about 1.25 percent over the 12 months from last November to this November. In reality, the economy will need to create an average of about 100,000 jobs per month over the next two months just to end up with zero growth for that 12-month period.

The scatter plot in Figure 3 is similar to the previous one, except that severity of recession and strength of rebound are here measured in terms of employment’s decline over the recession and growth over the first 10 months of expansion. It’s pretty clear from the figure that our current experience is close to being a repeat of our nearest neighbor on the graph—the 1990–91 recession and the "jobless recovery" that followed it.

Cooling Production Since July
While the slow pace of output growth observed earlier in the year is probably not a cause for concern, signs of a recent cooling in the pace of production certainly are. Recent data seem to indicate a slowdown in the pace of the expansion beginning in late summer, and this evidence has come primarily from the economy’s production side.

The July employment report from the Bureau of Labor Statistics, for example, showed a marked drop in hours worked in July. Figure 4 shows the Bureau’s index of aggregate weekly hours, which had been basically flat since last November. While the July drop in hours was partly offset in August and fully offset by September, the hours data seem to indicate at least a pause in the pace of activity on the economy’s production side.

The July hours data is corroborated by other barometers of production-side activity. For example, the purchasing managers’ surveys conducted by the Institute for Supply Management show a significant deceleration in both the manufacturing and nonmanufacturing sectors since July.

Figure 5 plots the ISM indices for manufacturing, in blue, and nonmanufacturing, in red. These are what are known as "diffusion indices," where readings over 50 indicate a sector is growing, and readings below 50 indicate decline. After having climbed into solid +50 territory early in the year, indicating growth, both indices slipped back towards "neutral" in July and August. The index for the non-manufacturing sector rebounded somewhat in September, but that for the manufacturing sector fell further, to a shade under 50.

As further evidence of cooling, industrial production, measured by the Federal Reserve Board, also fell slightly in August, breaking a string of 7 consecutive monthly increases. Is the economy tipping back into recession? While concern is definitely warranted, the data—which we’ll examine—don’t yet point to a double-dip scenario. First, while leading indicator indices have dipped in recent months, the movement has been quite small. Among financial indicators that have predictive power for future growth, signals are mixed. Perhaps most importantly, inventories have been pared down at all supply stages; consequently, if demand growth holds up, production will have to follow.

Let’s look first at leading indicators. Most of the various indices of leading indicators show declines over the past few months, but those declines have been small. By and large these measures have been flat since spring, consistent with the good-news-bad-news feel to the recent data. The probability of renewed recession remains small, but I should note that economists’ forecasting models are notoriously poor at predicting turning points.

Figure 6 offers one summary of leading index behavior. This figure shows recession probabilities calculated by the academic economists Stock and Watson. The red line is their coincident recession probability; each point on the line gives the probability that the economy is in recession at that date. The blue line is their leading recession probability; each point on the blue line gives the probability the economy will be in recession six months in the future. The gray shading indicates the 2001 recession, with November the assumed ending date.

Both probabilities are currently low, with the last value for the leading probability being 4 percent.

The figure also illustrates the sense in which economists are bad at predicting turning points. The work of Stock and Watson is about as sophisticated as economic forecasting gets; nonetheless, their leading recession probability gave only faint signals of the oncoming 2001 downturn.

What are financial market indicators telling us about the future pace of economic growth? The signals here are mixed. On the one hand, the yield spread—the difference in interest rates between bonds of long maturity and bonds of short maturity—remains high, though it has declined somewhat of late. Economic theory tells us that when interest rates on long-maturity bonds exceed interest rates on short-maturity bonds, markets are expecting short rates to rise, something generally associated with more rapid economic growth. Thus, a high yield spread generally signals a faster pace of economic activity down the road.

On the other hand, the junk bond spread—the difference between interest rates paid by issuers of junk bonds and issuers of high-quality corporate debt—has widened since the spring. A bigger junk bond spread indicates tighter credit conditions for below-investment-grade firms, and while this indicator has a short track record, increases in it have generally been a portent of slower economic growth.

Figure 7 shows the behavior of both spreads, measured in basis points, over the past two and a half years. The yield spread, the blue line, is measured as the difference between yields on 10- year Treasury bonds and yields on 3-month Treasuries. The junk bond spread is measured by the difference between an average interest rate paid by issuers of junk bonds and an average rate paid by Aaa-rated industrial firms.

Note that just prior to the 2001 downturn, the junk bond spread jumped up, while the yield spread fell into negative territory.

A final factor to consider when weighing the possibility of a second dip is the position of inventories. The great inventory correction that began in early 2001 seems to have run its course, with inventories bottoming out at all stages of the economy’s supply chain.

Figure 8 plots the ratio of inventories to sales for the manufacturing sector, since 1992. This ratio has been trending downward for some time, owing to technology-driven improvements in inventory management. Its current value—at about $1.30 in inventories for every dollar in sales—seems close to where one would’ve expected it to be, had the 2001 downturn—and consequent inventory blowup —not occurred.

With inventories stripped down, production will have to increase, so long as demand growth continues.

The Health of Demand
Hence, we turn to demand. It is from here, as I noted, that most of the good news has been coming lately. We’ll look first at firms’ investment in capital, where we have recently seen some spark of life. From there, we’ll turn to consumers, whose spending continues to grow at a moderate pace. In each case, we’ll want to see whether these sources of demand have sufficient stamina to maintain their forward progress.

First, investment. The 2001 downturn was, if not an investment-led recession, certainly an investment-fed recession. The declines in fixed investment alone—let alone inventories—more than accounted for all of the decline in GDP in the three quarters in which output fell.

Has investment begun to rebound? As Figure 9 shows, the answer is "somewhat." This figure displays the annualized quarterly growth rate of business fixed investment and some of its components. The data run through the 2nd quarter of this year. Overall fixed investment (shown by the green bars) fell in the second quarter, though by a much smaller amount than in prior quarters. If we look at the components, we see that investment in equipment and software (the blue bars) grew in the second quarter, for the first time since mid-2000. Within equipment and software, the information-processing, or high tech, portion (shown by the red bars) registered growth in both the first and second quarters. As one can see from the graph, the components that are growing are not growing at nearly their pre-recession rates, but they are growing nonetheless.

Somewhat more timely data is available from the Census Bureau’s reports on manufacturers’ shipments and new orders for capital. They are both shown in Figure 10; the data, which run through August, are shipments and orders for non-defense, non-aircraft capital goods, measured in millions of dollars. In spite of the wiggles, the basic pattern is one of slow growth since the end of last year.

Can investment growth be maintained? The outlook here is unclear. On the plus side, corporate net cash flow over the past three quarters has been up a bit, relative to where it had been for the past few years. Looking carefully at the data on orders and shipments, though, there is little evidence of forward momentum. (See Figure 10) In particular, orders (the blue line) are still slightly below shipments (the red line). That is, manufacturers of capital goods are getting slightly less than a dollar in new machinery orders for every dollar’s worth of machinery they ship, which would seem to portend slower future growth in shipments.

We now turn to consumers. Consumers, as we know, were the force that kept the 2001 downturn as mild as it was. How have they fared in the expansion? Thus far, as we’ll see, they are a bit bent, but unbroken. Growth in consumption spending and income have both rebounded, after slowing late last year.

Figure 11 presents one measure of the pace of consumption, the behavior of retail sales, since the start of 2000. The figure should make clear that consumers have been spending on more than just cars. Certainly, the on-again-off-again-on-again incentive programs of automakers have led to some large bursts of growth in overall retail sales, the red line. The more important story here, though, is the behavior of retail sales stripping out motor vehicles, the blue line. The picture that emerges from that data—which also run through August—is one of steady, moderate growth.

Can consumers keep up the pace? The picture here is, I think, brighter than was the case with firms. First of all disposable income has been growing faster than consumer spending for most of 2002—as a result, households savings rates have risen, considerably. Consumers, like firms, have been engaging in some "balance-sheet repair." And, while consumer indebtedness remains high, so too does household net worth—historically so—even given the stock market’s recent woes.

Figure 12 shows one gauge of household sector wealth—the ratio of household sector net worth relative to the size of GDP. These data run only through the 2nd quarter of this year, but even taking into account the further erosion in equity markets since then, the ratio is—at worst—at or above what had been its historic upper bound prior to the late 1990s.

The Inflation Outlook
Finally, we turn to inflation. Since the middle of last year, what had been a disturbing acceleration in the rate of consumer price inflation has turned into a very sharp deceleration. The consumer price index in August registered a 12-month inflation rate of 1.75 percent; most core or trend measures of CPI inflation are hovering near 2 percent.

The next figure (Figure 13) plots the 12- month growth rate of the CPI, along with a measure of core inflation, the CPI excluding energy goods and services, since 1993. The rapid reversal of the inflation outlook over the past year is striking.

The very low rates of "headline" CPI inflation, together with falling prices in some components of the CPI, such as goods prices or durable goods prices, has led to concerns that overall deflation may now be a real danger.

While falling goods prices are not, in themselves, evidence of deflation—and may, in fact, have reasonable explanations in terms of productivity growth—nonetheless, very low overall rates of inflation may still warrant concern.

Why should we worry at all about the possibility of deflation? While I don’t wish to suggest any similarity between conditions in the U.S. and in Japan, nevertheless, the Japanese experience since the early 1990s provides a vivid demonstration of the difficulties that can arise in a deflationary environment. As interest rates reach zero, the traditional tools used by central banks become unavailable, and policy can only be conducted through extraordinary measures. An economy may find itself mired in a "deflationary trap," which policy is powerless to break. Some caution would thus appear to be warranted.

But how close are we? With core CPI inflation around 2 percent it would seem that we’re not really that close. However, there may be good reason to view that 2 percent figure as an upper bound on the economy’s actual rate of inflation. First of all, in spite of the many technical improvements to the CPI made over the past several years, it’s likely that the CPI is still biased upward.

Figure 14 shows 12-month CPI inflation together with inflation in a new, "chained" CPI compiled by the Bureau of Labor Statistics since the start of 2000. The chained CPI is designed to reflect in a more timely way the impact of changes in consumers’ spending patterns on measured inflation. One should bear in mind that the chained index is still very much "experimental," and the numbers, unlike the usual CPI, are subject to revision. As the data currently stand, though, inflation as measured by the chained CPI is running at about 4/10 percent below the more usual CPI inflation rate, and has at times trailed the usual CPI by as much as a percentage point. Still, one might say, 1.34 percent—the index’s most recent value—seems quite a ways from zero.

If one looks at broader inflation gauges than the CPI, though—for example, price indices for GDP or some of its major components—one finds rates of inflation in those measures that are near 50-year lows. Those rates of inflation are also quite a bit closer to zero, though, I must add, still positive.

Figure 15 shows four-quarter inflation rates computed from the price index for GDP (the red line) and the price index for GDP, less government, farms and housing (the blue line). The current annual inflation rate for the latter—at a little over 4/10 of a percent—is below all but one observation in the past 50 years. We are, by these measures, as close to price stability as we have been at any time in the past 40 years.

What is the bottom line on deflation? By any reasonable measure, the economy’s overall inflation rate is quite low, but still positive. Given the possibly dire consequences of deflation, though, we may not want to allow further declines in inflation. For the past forty years, "price stability"—understood as a low, stable rate of inflation—has been a goal fought for by restraining inflation from above. In the current environment, maintaining price stability may now entail supporting inflation from below.

Conclusions
Let me offer some conclusions, and then some risks to those conclusions. First, we have seen that the current weak expansion is, at least in terms of output growth, not surprising, while the weak growth in employment that we have seen is. The recent cooling on the production side of the economy is a definite cause for concern, though it is premature to conclude that we are facing a double dip recession. Demand growth has thus far held up and—given the stripped down state of inventories—may yet carry the day. Finally, the reversal of fortune that has taken place on the inflation front has put us in a position where maintaining price stability may—for the first time in decades—mean boosting inflation rather than containing it.

Now the risks. There are any number of wild cards, which my analysis has (purposely) ignored. Foremost among them is the prospect of war with Iraq. Conflict could deal a shock to the economy in the form of higher oil prices, while a non-military resolution to the situation could dispel a great deal of the uncertainty currently hanging over the economic environment. We also have to consider the possibility of financial shocks, whether banking collapse in Japan, a debt crisis in Brazil, or problems emanating from elsewhere. Also, despite the stock market’s two-year slide, stock prices are—by some measures, such as aggregate price-earnings ratios—still at historic highs. While there may be very good fundamental reasons for those valuations, we can’t rule out the possibility that stocks could tumble a great deal further—that is, more equity market meltdown. And, while the past year has, thankfully, proceeded without incident, we have to acknowledge the chance for further September 11-type acts of terrorism.

On the upside, we should remember the long, often variables lags with which monetary policy takes effect, and the considerable amount of stimulus which may still be "in the pipeline."

Certainly, things may resolve positively in the end, along many of the dimensions I’ve just mentioned, but, for now, the downside risks appear to dominate.

—Jim Dolmas

About In Depth This article is based on a presentation by Jim Dolmas, 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.