• Archive

Redefining Economic Downturns

Not all economic events are unpredictable, says former Goldman Sachs partner Joseph H. Ellis in a new book, Ahead of the Curve: A Commonsense Guide to Forecasting Business and Market Cycles. Read an excerpt.

Early in my career as an investment analyst on Wall Street, I discovered an important truth in providing sound advice to my clients: Although my first job was to determine which companies and stocks (my specialty was the retailing industry) would provide the best investment returns over an intermediate-term (say, one to three years) or longer-term time frame, my second one was to determine when stocks of retailing companies, or the stock market as a whole, might be most favorably owned. Without question, there were long periods during which retail stocks and stocks in general were unrewarding. During such times, a buy-and-hold posture would prove to be destructive. These periods usually coincided with slowdowns in the economy lasting for two or more years. Economic cycles—periods of advancing and then slowing economic growth—were the rule and not the exception. These cycles appeared to occur every three to six years and carried with them corresponding bull and bear markets.

Furthermore, on the downside, there seemed to be a repeating pattern in which businesses and investors were invariably caught, without warning, in economic downturns and the accompanying bear markets. In cycle after cycle, the abilities of the business and investment communities to perceive the downturn as it occurred were typically so belated that there was little capacity for avoiding its damaging effects.

Much of the problem seemed to revolve around businesses' and investors' focus on recession—typically defined as two successive quarters of absolute decline in total economic output (real GDP) on a quarter-versus-prior-quarter basis—as the key economic event to be feared, the big bad wolf of the economy, so to speak. Businesses and investors seemed to feel that, if there were no recession, then everything was basically OK. But, repeatedly, business conditions, corporate profits, and the stock market appeared to suffer badly before recessions ever came into view.

The four stages of economic downturns
I observed four stages of perception associated with these downturns:

  • Stage 1: The peak. The economic environment would be uniformly favorable. Real GDP and its key component, consumer spending, were increasing at a strong pace, corporate profits were showing superb gains, and the employment picture was clearly favorable. The stock market would be reaching new peaks, with investors enthusiastically embracing a bright business outlook.
  • Stage 2: A modest slowing. The uniformly bright outlook of stage 1 would give way to a period of moderate slowdown in the rate of growth of the economy, particularly in consumer spending at the front end of the cycle. Retail sales growth would slow a bit. Interest rates would be rising gradually in response to the strong economy, and some vague concerns about interest rate and inflation would be raised, but not sufficiently to quell business and investor optimism. Capital spending and employment would still be growing at a strong pace (accompanied by much happy talk about a "full-employment economy"). And corporate profits would still be on the rise, but with percentage increases at a somewhat slower pace than the robust pace of the prior period. Few forecasters would yet fear that a recession—an actual decline in total economic output—was in the cards. And the stock market might well be sideways or, more likely, even down 5 to 10 percent, but believed to be only taking a "breather" after the strong market gains of the prior period.
  • Stage 3: Intensifying worrying. The economy would now enter a period of more intense worry, in which interest rates and inflation were higher, the rate of growth in real consumer spending and real GDP had slowed from a peak of 5 to 6 percent to "moderate growth" of perhaps 2 to 3 percent, and economists and others would begin to contemplate the possibility that the economy could enter a recession. Conjecture would now begin to center on how long and deep such a recession (if it occurred) might be. The stock market by now might have declined an additional 10 percent or more from stage 2; a bear market was now under way. However, some comfort would still be taken from the fact that capital spending was still relatively strong and the unemployment rate remained low.
  • Stage 4: The advent of recession. Now the economy would actually enter the recession, a period of absolute decline in real GDP, with corporate profits falling significantly, capital spending beginning to weaken, and—perhaps most alarmingly—the first major increases in the unemployment rate. And if an accelerating number of workers were losing their jobs, where would the impetus for an upturn come from? Fears of a protracted decline would now become more widespread, with a great deal of economic discourse devoted to determining when the recession began: when quarter-to-quarter real GDP comparisons fell below zero—that is, began to decline (as if the zero demarcation level were any more important than any other number). During the early phases of stage 4, stocks would continue to decline as investors confronted these fears. As a period of unusually wide-spread pessimism, stage 4 appeared very much like a mirror image of the unquestioned optimism of stage 1. And, most perplexingly, at some point during the recession, when economic conditions were at their worst, stock prices would stop declining and—mysteriously—would begin to advance.

As an analyst responsible to my clients for providing guidance on the performance of my (retail-industry) stocks, I began to recognize patterns of economic and investment damage. Not only my retail stocks, but also the stock market as a whole, had peaked during stage 1 when economic growth was still at its best and optimism was rampant. Stock prices were already in decline as sales-growth rates for business throughout the economy began to slow in stage 2, and this decline intensified in stage 3 as sales increases slowed even further, but before an actual recession was even considered likely. By stage 3, it was already too late for most companies to react to slowing business conditions. Investors, too, were faced with a dilemma by the time stage 3 ran its course: the difficult choice of either waiting things out in the hopes that a recession could be avoided, or selling—to avoid further losses—at what could be the bottom.

The greatest economic damage was done when rates of economic growth began to slow.

I now began to understand that the central role given to recession—the advent of stage 4—as the accepted definition of economic harm was itself most damaging to businesses and investors. The emphasis that economists, businesspeople, and investors placed on recession as the key negative economic event actually led them to miss the fundamental fact that the greatest economic damage was done when rates of economic growth began to slow, and this period began at the end of stage 1 and continued through the beginning of stage 4, long before stage 4's absolute decline in economic activity. It was in stages 2 and 3 that an unexpected slowdown in sales growth would begin to cause rising inventories, pricing weakness for businesses, falling profits, and declining stock prices. In fact, in a number of downturns (for example, 1966-1967 and 1984–1985) the economy never even reached stage 4; that is, a recession, or actual decline in economic activity, never even occurred. Yet there was clear damage to corporate profits and the stock market.

As I write, I have been tempted to draw a weather analogy with this economic pattern. Stage 1 is the bright, unclouded sunny day, stage 2 is sunny but with a growing haze, stage 3 is cloudy with the threat of thunderstorms developing, and stage 4 is the violent thunderstorm itself. But I then realized how inaccurate this analogy is, and yet helpfully descriptive in its very inaccuracy. With weather, we can wait through hazy and then cloudy and worsening skies (stages 2 and 3) and board up the windows shortly before the onset of the storm (stage 4). But in the economy and the stock market, much of the damage is already done between the end of stage 1 and the end of stage 3, before stage 4. In the economy and the stock market, stages 2 and 3 represent, in reality, the first half to two-thirds of the damage, and stage 4—the recession—is the beginning of the end of the harm.

Two great flaws in conventional economic analysis
It now became clear to me that to get ahead of the curve in predicting the economy, it would be necessary to alter the most widely used methods of tracking economic data. The goal was to correct the two major dysfunctions in the conventional means by which most businesses, investors—and yes, even economists—measure economic cycles.

  • Economic flaw 1: recession as the primary measure of economic slowdown. To use the concept and definition of recession—a period of absolute economic decline in real GDP—as the most widespread measure of economic harm is fundamentally tardy and, therefore, useless as a pragmatic economic measure for businesses, investors, and policymakers. In every economic cycle—those I experienced in my first decade as an analyst, and those since—by the time a decline in real GDP occurs, or is even close at hand, the general economy, businesses, and stock market have already suffered considerable damage. Most damage to the economy, businesses, and the stock market happens much earlier, when the rate of growth in the economy begins to slow from peak levels. By the time the economy is actually declining in absolute terms, most of the damage is already done. Recession needs to be replaced by a more sensitive and agile concept of economic slowdown.
  • Economic flaw 2: tracking economic data on a quarter-to-quarter and month-to-month basis. The method that is used to track rates of change in most economic series and indicators creates unneeded confusion and often obscures viable economic cause-and-effect relationships that might have been used more effectively to forecast future events. Government data bureaus and other sources of economic data, and most economists, track the rate of change in most major economic indicators on a quarter-versus-prior-quarter basis (headline: "Real GDP rose 3.5 percent in the second quarter"). This method requires (1) seasonal adjustment of the data for the quarter and then the one that preceded it, (2) multiplication by 4 (or 12, if the data is monthly) to determine a seasonally adjusted rate of change, and (3) frequent revision. There are some pragmatic short-term reasons for reporting economic statistics in this manner. But looking at economic data on a quarter-to-quarter basis over longer periods results in so much volatility, or noise, that causality between economic series is often obfuscated, when a cause-and-effect relationship in fact exists. I found that, as we seek to find longer-term predictive relationships between economic indicators, the way data is charted is fundamental to achieving clarity.

When we combine recession, as the wrong definition of economic harm, with quarter-to-quarter rates of change as the wrong method for measuring it, the result is confusion that obscures the real inflection points in the business cycle—when rates of change begin to turn.

By correcting these two overarching problems associated with most economic analysis, we can learn to look at the same data everyone else has, but from a different perspective that enables us to see patterns others may miss. Like the pilot of an airplane at thirty thousand feet, we can look down and see patterns that the driver of a car—lost in the maze of "recession," "growth," and confusing quarter-to-quarter data comparisons—will not recognize. Eliminating these two errors will make us better able to evaluate the public conversation from a detached and illuminated point of view and help us wean ourselves from the confusion of much of current economic discourse.

Excerpted by permission of Harvard Business School Press from Ahead of the Curve: A Commonsense Guide to Forecasting Business and Market Cycles. Copyright 2005 Joseph H. Ellis; all rights reserved.

[ Buy this book ]

Joseph H. Ellis was a partner at Goldman Sachs and was ranked for eighteen consecutive years by Institutional Investor magazine as Wall Street's No. 1 retail industry analyst. For more information about Ahead of the Curve, see the Web site.