Bad News Ends Long After Bull Markets Start

William Hester, CFA
November 2008
All rights reserved and actively enforced.

"Let me be clear on one point: I can't predict the short-term movements of the stock market. I haven't the faintest idea as to whether stocks will be higher or lower a month — or a year — from now. What is likely, however, is that the market will move higher, perhaps substantially so, well before either sentiment or the economy turns up. So if you wait for the robins, spring will be over."

- Warren Buffett

The data is worsening and sentiment is bad. Neither is likely to improve over the near term. This set of conditions certainly doesn't provide comfort for investors. If the bad news continues for several months, well into next year, should investors rule out the potential for stock prices to recover?

The case for stock prices moving lower into 2009 usually rests on two arguments. The first is that corporate earnings will fall sharply as the economy continues to slow. The second is that the labor market will continue to deteriorate, eventually sinking consumption further. Both may happen. Analysts have recently been quickly lowering their earnings growth estimates for next year. Earnings are expected to grow about 1 percent in the first quarter and about 5 percent in the second quarter. But analysts tend to be overly optimistic. If past patterns hold, estimated earnings growth will likely turn negative as these quarters approach.

It's also difficult to find anything positive to say about the labor market. Jobless claims hit 542,000 this week, the highest level since 1992. The four-week moving average of that series rose above 500,000, something that has occurred only four times in the last forty years. The other occurrences include some of the worst recessions, including 1974 and 1982.

With the market now down by half from its peak, stock prices certainly reflect a lot of negative expectations. Investors have moved from questioning whether the economy is in recession to what will be its depth and duration. Comparing this recession to previous economic downturns may be helpful.

Based on the data that the NBER uses to mark the peaks in the economy, the recession likely began sometime around January, or possibly late last year. The average post-war recession has lasted 10 months, which means the current recession may already be longer than average.

The contraction in the economy during the Great Depression lasted 3.5 years. Short of that, the 1974 and 1982 recessions were the steepest based on the contraction in GDP. They both lasted 16 months, ranking them as the two longest recessions since the Great Depression. If the length of the current recession matches those two recessions in duration, the economy could be expected to contract through the middle of next year.

We know that we can't fall back on sayings like "the stock market will bottom 6 months prior to the end of a recession." As John Hussman noted in How Low, How Bad, How Long , there's a tremendous amount of variation around where stocks bottom in relation to the end of a recession. But in post-war data, stocks have always eventually rallied prior to the conclusion of each recession.

It is important to keep in mind that even if the recession were to end sometime in the middle of next year, that doesn't mean the data will become uniformly optimistic. Just as this beginning of this recession has been difficult for NBER officials to call because of the mixed signs between housing, employment, and industrial production, the signs of recovery are likely to be just as mixed. The housing sector went into recession long before its typical lead time prior to a slowdown. The slump could last longer based on the unique characteristics of this cycle. And the economies of Europe and Japan look to be entering into their own recessions later than the US, which may weigh on the US exports well into next year. So expect the early recovery data to be as confusing as the early contraction data.

Earnings Will Decline Next Year

It's best to tune out any forecast for the performance of next year's stock market that is based on expectations for near-term earnings growth. There's almost no correlation between year-over-year earnings growth and stock market performance. But the habits of investors are hard to break, and the focus on quarterly earnings is unlikely to dim soon. If we assume that earnings will decline during most or all of next year, should we also assume that stocks will decline?

The graph below plots the percentage change in earnings per share of the S&P during the first two years of each bull market that followed a recession. I included the nine recessions since 1950 and used the more-familiar operating earnings series, which was provided by Ned Davis Research. (The data prior to the early 1980's - during which period operating earnings did not exist - is based on the typical relationship between operating earnings and reported earnings.) The chart is scaled to the current level of operating earnings. This gives a range of where earnings could evolve over the next two years, even if the market was to rally.

The top line is the bull market following the 2001 recession. This instance is unique because that bull market began almost 16 months after the recession ended (partly because valuations were still unattractive at the end of the recession). The green line shows the growth in earnings during the first two years of the bull market beginning in September of 1953. These first two instances may be what investors and analysts have in their heads when they think about the stock market. They often argue that stocks can rally only when earnings are expected to improve quickly.

But look at the direction of earnings during the first two years of the other 7 bull markets. Earnings sometimes fell another 15 to 20 percent over a year or longer after the bull market started, and were typically flat even two years later. In 1990, earnings fell almost 20 percent over almost a 2-year period. During the same period the S&P rose more than 40 percent.

Based on the average decline in earnings during the early stages of a bull market, operating earnings for the S&P 500 could fall to $57 next year even if a bull market was to start today.

It's also important to keep in mind how much earnings have already declined. From their peak last year, operating profits have fallen 24%. That's about as much as profits fell from their peak during the entire 1991 recession, and three quarters of their total drop in 2002, which was one of the worst earnings downturns on record. During the 2002 downturn, operating earnings fell by 32 percent from peak to trough (reported earnings fell by 50 percent). If operating earnings fall to $57 next year, that would be a plunge of almost 40 percent from last year's peak.

The Labor Market

Economist's forecasts for the labor market have become much more pessimistic, relative to just a few months ago. As I noted in The Beginning of the Middle , economists were expecting the unemployment rate to rise to 5.9 percent by the second quarter of next year, which was far below what one could estimate based on the typical pattern during recessions. In the most recent survey, economists now think the unemployment rate will rise to 7.25 percent in the second quarter, and to continue climbing throughout the year, ending at 7.7 percent.

If the unemployment rate continues to rise throughout all of next year, will that prohibit a stock market rally? The graph below attempts to put this into perspective. It takes the absolute change in the unemployment rate during the first two years of each bull market following a recession and scales it to the current rate. The unemployment rate rose to 6.5 percent in October, up from its trough measure of 4.4 percent two years ago.

The two lines that reach the highest points (the green and blue lines) reflect increases in unemployment from the very low levels of the 1950's. The purple line represents the rise in unemployment that resulted from the 1974 recession. Its starting level matches today's well. In 1973 the unemployment rate bottomed at 4.6 percent (versus the current trough rate of 4.4 percent) and would eventually rise to 9 percent a year and a half later. Along the way, it reached 6.5 percent (the present rate of unemployment) in November 1974. From November 1974 when the unemployment rate reached 6.5 percent to when it reached 9 percent six months later, the S&P rallied by more than 30 percent.

More than half of the estimates in the graph sit between a 7.5 percent and an 8 percent unemployment rate, which is in-line with today's average estimate for next year's jobless rate. So even if the unemployment rate is fully expected to increase significantly, investors should not conclude that stocks will persistently decline. Even if the labor market contracts to a greater extent than is expected, forcing the unemployment rate to 9 percent, it would still be consistent with the first two years of the bull market that started in 1974. (On Friday Goldman Sachs economists said that they now expect the unemployment rate to reach 9 percent by the end of 2009.)

One can apply the same analysis to a number of different data series. Though the lag times vary somewhat, they reinforce the same message: the early stages of a bull market almost always coincide with poor and deteriorating earnings and economic data - sometimes extremely poor. That bad news can easily continue for a year or more after the market reaches its low.

None of this data argues that stocks will rebound in the near-term, or that a new bull market is within reach. But it does suggest that stocks will begin to recover long before the earnings and jobs data does.

The combination of the current valuation of stocks and the amount of bad news that has already been announced is one that has historically favored longer-term investors. As I mentioned above, the four-week moving average of the jobless claims data breached 500,000 this week, which has happened only 4 other times. It occurred in December of 1974, in April of 1980, in November of 1981, and in March of 1991. During the 12-month period following these periods, the S&P rose 32 percent, 30 percent, 20 percent, and 9 percent, respectively. These periods also shared attractive valuation. Over the four periods the price-to-peak earnings ratio averaged 8.75, which is about right where the market's current valuation is. Although it's a small sample, low valuation, coupled with economic data confirming a substantial contraction in the labor market, has offered longer-term investors very strong average returns.

Those returns aren't restricted to bull markets that follow the worst recessions. Returns following all of the recession-induced bear markets have been quite strong. The graph below shows the first-year returns following the 9 recessions we've examined. First-year returns following a recession have averaged 37 percent with surprisingly little variation. Not including the out-sized gains following the 1982 bottom, all of these first-year bull markets gained between 29 and 44 percent.

Warren Buffett recently suggested that if you wait for the robins, spring will be over. Not only should investors not wait for good news on profits and the economy, they should actually brace themselves for the news to worsen predictably, well after the stock market reaches its lows.