Overview of the theory
In previous articles I have introduced the concept of secular market trends and relative P/R, a valuation concept closely coupled with secular trends. A thorough development of this theory is found in my book Stock Cycles, but for the purposes of this article the summary given in the two referenced articles should give the reader sufficient background to follow the arguments presented here. The secular trends concept holds that the long-term rise in stock prices displays a "stair-step" pattern. During the secular bull markets (corresponding to the riser) stocks move strongly upward, averaging a 13-14% return in real terms. During the secular bear markets (corresponding to the step) stocks move basically sideways, averaging close to no return in real terms. The market spends roughly equal amounts of time, in 10-20 year blocks, in each kind of secular market trend. I propose that the 1982-2000 period constitutes a secular bull market and that since last year we have entered a secular bear market which could last as long as 20 years.
The P/R and relative P/R valuation concept starts with the idea that business resources (R) constitute the true measure of the intrinsic value of broad-based representative index of stocks such as the S&P500. By definition, R has no meaning for an individual stock. The value for R can be calculated for such an index using earnings, dividends and a consumer price index as is described in Stock Cycles.
Application of the model during a secular bull market
The identification of secular bull and bear markets in real time makes use of P/R, the ratio of the S&P500 index value to R. For example, in fall 1995, when I first noted the secular market trends, it was apparent we were in a secular bull market that had begun in 1982 and P/R, at around 0.75, was still well short of the >1.0 values it had typically shown at the end of previous secular bull markets. Thus, it was clear that the secular bull market had years left to run. During a secular bull market, the trend is strongly upward and long-term asset allocation should be 100% in stocks. For traders, methods should be backtested using previous secular bull market periods. The only question to be answered is when will the secular bull market end? Starting in May 1997, I produced sporadic updates exploring this question (see Dec 1999, Oct 1998) before calling an "end to the secular bull market", and moving my 401K from stocks to cash in summer 1999.
Application of the model during a secular bear market
I must stress again that secular trend theory is suitable only as a guide for long-term asset allocation strategy. It is not a trading tool, although a trader can use it to help formulate a market view, the details of that view will necessarily depend on a number of short-term technical analysis tools. Once there is reason to believe that a secular bear market has begun, asset allocation strategy changes from fully invested in stocks to a mixture of stocks and non-stock investments. The stock index can still outperform money markets if they are purchased at "low enough" prices, or if purchased at short-term low prices and sold after a relatively short period of time. Thus, when stocks are expensive relative to their long-term future performance one should over weighted in non-stock investments. Conversely, when stocks are cheap relative to their long-term future, one should be overweighted in stocks. Simplistically, at the beginning of a secular bear market (e.g. last year) one should have mostly cash and at the end of the secular bear market (e.g. the early 1980's) one should be mostly invested in stocks.
How to change asset allocation as the secular bear market progresses will depend on two things, market valuation (as measured by relative P/R) and investor risk tolerance. For investors with low risk tolerance and a 5+ year time horizon, the "safe" level is given by the following expression:
1) safe level = 500 (1.02 + i / 100)N
Here i is the average inflation rate in the future and N is the number of years after 2001. The safe level refers to the level of the S&P500, below which long-term investments in this index will almost certainly outperform money markets if held for 5 years or more. The levels given by equation 1 are so low they are not likely to be reached anytime in the next few years (although they could be). The purpose of equation one is to keep risk-adverse investors entirely out of the market during the secular bear market, only reentering when it is nearing its end (or a stupendous buying opportunity like summer 1932 emerges). More enterprising investors will want to avail themselves of the potential gains to be had during secular bear markets, keeping the risk in mind, of course.
The key parameter we will focus on is relative P/R, a tool that allows one to estimate the value of the S&P500 index relative to its long-term (30 years) future performance. The lower relative P/R becomes, the higher the probability that long-term future returns will be better than money markets, meaning one should own the stock index rather than a money market. Relative P/R reached an all-time high of 2.29 in July 1999, based on the monthly average of the index. In March and August of the next year, relative P/R reached the same peak level, before beginning a dramatic drop to a low (so far) of 1.77 in March 2001. Figure 1 shows a graph of the behavior of relative P/R (rPR) around the July 1999 peak. Also shown is the same behavior for four previous peaks in rPR.
Figure 1. Profiles of relative P/R (rPR) shortly before and after long-term peaks
Peaks in rPR tend to occur near the end of the secular bull markets, sometimes right at the top such as in June 1881 and September 1929. Other times they precede the top, sometimes by a considerable amount of time. For example, rPR peaked in December 1961, four years before the secular bull top in January 1966. The June 1901 rPR peak preceded the September 1906 top by more than five years. Finally the first rPR peak in July 1999 preceded the (so-far) August 2000 top of the recent secular bull market by more than a year. We still cannot rule out yet another bull market which will carry the index to a marginal new constant-dollar high in the next few years, which will make the official end of the secular bull market later than August 2000.
To get an idea of the valuation territory we have entered since the year 2000, we will look at capital gains returns in constant dollars following the months with values of rPR greater than 1.65, which corresponds to a present value of 1100 on the S&P500. We will look at future capital gains over one, three, five and ten year periods. For one-year returns we have 50 data points, including 9 from the July 1999 to March 2000 period. For the longer periods we have 41 data points, mostly from the 1960's. We focus on capital gains since the level of dividends today are less than the real return available from money market funds. Thus we can interpret the real capital gains returns shown in Table 1 as roughly equivalent to the expected total return from the stock index relative to that from a money market fund. A negative value in Table 1 indicates a situation in which the stock index investment would have failed to outperform the safe money market given today's low dividend.
Table 1: Projected capital gains in real terms from historical markets with rPR > 1.65 (S&P500 = 1100)
| X percent of returns better than ® | 1-year | 3-year | 5-year | 10-year |
| X = 10% | +8% | +22% | +22% | +1% |
| X = 25% | +5% | +5% | +4% | -19% |
| X = 50% | +0% | -3% | -5% | -36% |
| X = 75% | -11% | -5% | -18% | -41% |
| X = 90% | -18% | -27% | -20% | -44% |
The interpretation of Table 1 is straightforward. Looking at the one-year gains we see that 10% of historical returns from markets showing rPR > 1.65 were 8% or greater. The median return was 0%. Ten percent of returns involved a loss of 18% or more. Over a three year period, 10% of the time the S&P500 index advanced 22% or more. Another 10% of the time the market lost 27% or more. The median return was a 3% loss over three years. Examination of the median returns show that the longer one holds the index purchased at these levels, the more likely one will lose money relative to money markets. This is the reason for my book's title: Stock Cycles: why stocks won't beat money markets over the next twenty years.
We would also like to know how further declines in the market will affect the probabilities shown in Table 1. To address this question we can look at the next 50 highest rPR values, which fall into the range 1.45 to 1.65, and correspond to a present S&P500 level of 970 to 1100. These values come from the 1960's too, but also from 1930 and the 1880's. Should the S&P500 fall below the April 4th low we will enter into this category.
Table 2: Historical real capital gains for rPR between 1.45 and 1.65 (S&P500 = 970-1100)
| X percent of returns better than ® | 1-year | 3-year | 5-year | 10-year |
| X = 10% | +16% | +31% | +16% | +4% |
| X = 25% | +14% | +14% | +9% | -7% |
| X = 50% | +6% | -8% | -8% | -42% |
| X = 75% | -14% | -23% | -18% | -48% |
| X = 90% | -20% | -32% | -22% | -48% |
Table 2 shows that long-term positions taken even at these lower levels are also unlikely to beat money market funds. In the shorter term some gain is possible, as the median return is positive. Going still lower in rPR, Table 3 shows the 200 months with rPR falling between 1.17 and 1.45, corresponding to an S&P500 of 780 to 970 for today. These data come from all previous secular bear markets. Note that returns over all time scales are quite poor.
Table 3: Historical real capital gains for rPR between 1.17 and 1.45 (S&P500 = 780-970)
| X percent of returns better than ® | 1-year | 3-year | 5-year | 10-year |
| X = 10% | +12% | +23% | +20% | +17% |
| X = 25% | +3% | +4% | +9% | +9% |
| X = 50% | -7% | -12% | -9% | -8% |
| X = 75% | -18% | -31% | -33% | -21% |
| X = 90% | -28% | -48% | -44% | -47% |
Going still lower (Table 4) we finally reach a level at which 10-year investments in the index will match returns on a money market fund. Shorter term investments are still quite poor. For each table I have given the range of valuation in terms of both rPR and today's S&P500. Unless the market crashes this year, it is unlikely that rPR levels as low those in Tables 3 or 4 will be seen this year. For future years the values for the S&P500 equivalent should be increased by about 2% plus the inflation rate for each year after 2001. That is, Table 4 corresponds to an S&P500 of 650-780 for 2001, but for 2007 it would correspond to 870-1050 (assuming 3% inflation). Unless we get a repeat of the 1929-1932 debacle, the market may take years to work its way down to Tables 3 and 4. Relative P/R will fall as time goes on even if the index does not because R increases with time. Hence, we may spend quite a bit of time in Tables 1 and 2.
Table 4: Historical real capital gains for rPR between 0.98 and 1.17 (S&P500 = 650-780)
| X percent of returns better than® | 1-year | 3-year | 5-year | 10-year |
| X = 10% | +19% | +25% | +20% | +57% |
| X = 25% | +7% | +1% | +6% | +41% |
| X = 50% | -7% | -8% | -10% | 0% |
| X = 75% | -19% | -26% | -21% | -17% |
| X = 90% | -30% | -35% | -39% | -48% |
An important thing to note is during the early stages of the secular bear market, described by Tables 1 and 2, short term returns can be quite good despite the high level of rPR. Later, as we penetrate deeper into the secular bear market and rPR moves lower, both short term and long term investments become poor. Finally, as rPR falls below 1.0, long-term stock returns start to become more promising again. Short-term investments show promise during the early part of the secular bear market precisely because of the uncertainly that a secular bear market has in fact begun. Until a new bull market fails to exceed the previous bull market peak, some investors will believe that the secular bull market is still in progress. Bullish models, such as Harry Dent's, also will provide support for the bullish thesis. As a result, investors will be much more aggressive with long positions today than they will be years from now (despite rPR being higher today). Users of the secular trend model can take advantage of this by selling into rallies in the coming months. Enterprising traders can enter long-side trades at potential technical bottoms and go short as the market approaches previous tops. One must be careful to maintain a large reserve of cash against a possible sharp drop. Secular bear markets can unfold with blinding speed (e.g. 1929-1932), but this is rare (it has only happened once). More likely is a prolonged period in which the market retains a high rPR for a while, and operates in the region covered by Tables 1 and 2.
As time goes on, P/R and rPR will drop, if only because R rises without a commensurate rise in P. For example, the development of high levels of inflation would cause R to climb rapidly, and P/R to fall. As rPR falls, those who employ secular trend theory will move to cash and wait for inflation to push the market right through Tables 3 and 4 to very low levels of rPR. As Figure 1 shows, rPR typically trends downward as the secular bear market rolls on. When rPR levels fall below 0.6 stocks become hands-down superior long-term investments (see Table 5). When one also considers that at these low values of rPR stock dividend yields will likely be larger than real interest rates, the already good situation displayed by Table 5 become even better. But these low levels for rPR could easily be 10-20 years away.
Table 5: Historical real capital gains for the lowest rPR values between 0.35 and 0.57
| X percent of returns better than ® | 1-year | 3-year | 5-year | 10-year |
| X = 10% | +47% | +39% | +88% | +277% |
| X = 25% | +23% | +33% | +75% | +137% |
| X = 50% | +2% | +6% | +24% | +122% |
| X = 75% | -9% | -17% | +10% | +71% |
| X = 90% | -20% | -28% | -3% | +24% |
