To say that the market is closer to a bottom than a top is true by definition, but essentially irrelevant. The S&P 500 topped at 1553 in March 2000, at 1576 in October 2007, and is now at 845. The key question, though, is the level and timing of the eventual bottom. Although we don't know the answer, history provides us with some rough guidelines that indicate some of the possibilities.
In the table below the first column shows the date of all post-World War II market lows that were accompanied by recessions. The second column marks the date of the associated economic cycle low as determined by the National Bureau of Economic Research. The third column indicates the number of months by which the market low preceded the economic low. The fourth column shows the price-earnings ratio on smoothed "reported" (GAAP) S&P 500 earnings at the market low.
|
|
|
Lead Time |
|
|
Market Low |
Cycle Low |
Months |
P/E Ratio |
|
6/1949 |
10/1949 |
-4 |
6.3 |
|
9/1953 |
5/1954 |
-8 |
8.0 |
|
10/1957 |
5/1958 |
-7 |
12.3 |
|
10/1960 |
2/1961 |
-4 |
15.4 |
|
5/1970 |
11/1970 |
-6 |
11.0 |
|
12/1974 |
3/1975 |
-3 |
7.4 |
|
3/1980 |
7/1980 |
-4 |
7.0 |
|
8/1982 |
11/1982 |
-3 |
7.1 |
|
10/1990 |
3/1991 |
-5 |
14.1 |
|
10/2002 |
11/2001 |
+11 |
15.5 |
| |
|
|
|
|
Average |
|
4.9 |
10.4 |
|
Median |
|
4.0 |
9.5 |
The above table enables us to make some useful observations. First, in 9 out of 10 instances the market bottomed before the economy, and by an average of 4.9 months (say 5 months with rounding). On 6 of these occasions the market bottomed from 3-to-5 months before the economy. Our conclusion here is that the market usually bottoms during a recession, but not too long before the end.
Second, the average P/E ratio at market bottoms was 10.4 with the median at 9.5 (we'll round this off at 10). We note, too, that in 5 of the 10 observations, the P/E bottomed in single digits ranging from 6.3-to-8.0. In the other 5 instances the P/E troughed in double digits between 11.0 and 15.5.
Third, while the trough P/Es cover a wide range of 6.3-to-15.5, a closer inspection of the data provides some hints as to whether an eventual P/E will bottom in single or double digits. For this purpose we note that a secular bear market ended in 1949 with the subsequent secular bull market lasting until 1966. The market then remained in a secular bear market through 1982, followed by a bull market from there until 2000, when the current secular bear market started. Since secular bear markets comprise periods where P/E ratios are declining while secular bull markets are periods where P/E ratios are expanding, it is logical to assume that cyclical P/E ratios are low at the end of secular bear markets and the first few years of bull markets. In contrast, cyclical P/E ratios are high in the last few years of bull markets and at the beginning of bear markets. It is not a stretch to say that this is exactly what our data shows.
Given the above observations, what can we guess about the timing and level of the of the market bottom? In terms of timing, we believe that the current recession will be long and deep as result of the massive credit problems, the deleveraging of the economy, the rapid drop in the numbers and the global nature of the decline. Therefore, unless one believes that this is merely a garden-variety recession that bottoms as early as the 2nd quarter of 2009, it is unlikely that a market trough has been reached currently. While no one knows when the economy will bottom, any trough that goes beyond the 2nd quarter means that an eventual market bottom will be pushed off until well into 2009.
As for levels, we believe that the secular bear market is now in its 8th year, meaning that the P/E ratio is more likely to end in the bottom of its historical range than the top. Therefore a final low P/E ratio of about 8 on smoothed earnings is not far fetched.
Taking all of the above into account, where are we now? At today's S&P 500 close of 845, the market is at 12.8 times our trailing smoothed reported earnings of $66. That is nothing to sneeze at since it's the lowest P/E seen since the 1987 crash, and is also under the long-term average of about 15. It is also a vast improvement over the P/E ratio of 37 at the 2000 high and a P/E of 24 at the October 2007 peak. Therefore, predicting a lower P/E from this point is no longer the layup it was at far higher levels. However, we note that in 5 of the 10 data points in the table the P/E bottomed at 8 or less, and that would entail a big decline, even from this point. In addition those occurred either during secular bear markets such as the one we are in now or early in secular bull markets when investors did not yet recognize the new uptrend. It also noteworthy that in 24 of the 38 years preceding 1987, the market sold at lower than the current P/E at some point during the year
So far the low for the current bear market was 741 on November 21 when the P/E ratio would have been 11.2. That is close enough to the long-term average of 10.4 at bottoms to raise the possibility that it could have been the final low. However there is still a real risk that the eventual P/E ratio could drop as low as 8 times smoothed 2009 earnings of about $70, resulting in an S&P 500 index of about 560. Since that constitutes another 33% decline from current levels, it is not a risk that can be easily dismissed. We also note that our assumption of $70 in trendline earnings is a generous number since S&P now estimates 2009 reported earnings at only $49. All in all, although we think there is some possibility that the market has already bottomed, it is more likely that the market decline still has quite a bit further to go.
