June 2015

Progress of the secular bear market: position as of May 31, 2015

The value for R is 2029 as of May 2015.

10000 point decline in the Dow in the cards over the next three years

I have been tracking the progress of the secular bear market since forecasting it in Stock Cycles in 2000. The valuation tool I employ is P/R, which I outlined in my first article at Safehaven in 2001. R stands business Resources. It refers to the non-labor resources businesses employ to earn a profit. R can be thought of as a constant-dollar book value. R in year t is the sum of retained earnings from 1871 to year t plus the value of R in 1871. I assume that R in 1871 is equal to the index value in 1871 and proceed from there. Values for the index price, earnings and dividends were obtained from Professor Shiller's website. The figure at the top of this article shows plots of P/R for previous secular bear markets. Until about a year ago the current secular bear market had been behaving in a very conventional way--even the disruption of the crisis in 2008. But over the last year the behavior of the market has been unprecedented, assuming that we remain in a secular bear market.

I believe the recent rise in the market reflects a bout of irrational exuberance. Consider, in January 1999, P/R reached a level equal to its preceding record value, set in 1901. I had sold half my portfolio the previous month and would sell the rest by the end of the summer. Yet the S&P reached its highest value (on a monthly average basis) in August of 2000, 17 months later. The story told by the NASDAQ is even more interesting. In January 1999, it stood at 2500 and P/R predicted serious declines in the next few years. Indeed, the NASDAQ averaged about 1240 in October 2002, half its value in January 1999. What is interesting is the path taken by the "decline" from 2500 in January 1999 to 1240 in October 2002. On the way "down" it (briefly) was over 5000. Fed Chairman Alan Greenspan had called the late 1990's market "irrational exuberance" and that is what it was.

The figure at the top of this article shows that the stock market reached an all-time high P/R for our current position in the stock cycle in May of 2014 and since then has moved much higher. But this assessment assumes that we are still in a secular bear market, and that stock market cycles are meaningful. How does the market compare to history if we ignore stock market cycles? The current value of P/R of 1.04 is well below the value of 1.47 seen in 2000. It is even below the absolute P/R of 1.06 in 1966, the lowest of all the secular bull market peaks. (Note: the 1966-1982 bear is scaled in the top figure so as to fit in with the other secular bear markets in order to show relative levels). So is the current level really all that high?

Why the market level today is problematic

The stock market is essentially a capital market; it serves the function of putting a price on the basket of capital associated with a specific firm. A broad-based stock index, like the S&P 500, provides a price for the capital collectively possessed by the companies in the index. R represents the cumulative investments responsible for that capital and can be thought of as an independent measure of the capital in the index.

It follows that the index earnings (E) divided by R gives the return on capital (ROC) for the index. The quotient of ROC and P/R gives the return on price (ROP) which should be indicative of the typical return on investment (ROI) from a purchase of a diversified portfolio of stocks in the index. Since capital is a real thing, ROC is a real return and so is ROP. ROC declined after WW I and never recovered. Table 1 shows ROC for the period before and after 1916. ROC was much lower after 1916 than before. The stock market responded by pricing capital lower, the average price for R was 24% lower after 1916 than before. As a result ROP did not drop nearly as much as ROC. Actual observed total return was only modestly lower.

Table 1. Estimation of stock market total real return using ROC and P/R

So far we have seen that around 1916 the way the stock market works changed in ways that reduced the average price of capital. There is an old Wall Street saying that the market is driven by fear and greed. Before 1916, the balance between fear and greed maintained P/R at about 1, and investors reaped a real return of 7.0% on a ROC of 7.3%. After 1916, the average level of fear increased so as to reduce average P/R, enabling an ROC of 4.5% to give a 6.2% real return. To do this required large swings in market levels (i.e. the secular bull and bear markets) to generate the fear necessary to produce low average prices. The figure at the top of the page shows that secular bear markets after WW I ended at much lower levels than those before WW I. Presumably, the current secular bear market should be no different than the other three post-WW I ones.

It is a reasonable assumption that the S&P 500 and Cowles index are sufficiently broad-based to be representative of the economy. In this case, R serves as a proxy for the amount of capital in the economy. As capital and labor are both factors of production, GDP should be correlated with both. Since the emergence of a modern industrial economy after the Civil War, GDP per capita (GDPpc) has been proportional to R.1 This means the ratio of GDPpc to R should be approximately constant if capital is being used efficiently. Indeed, from 1871 to 1907 GDPpc/R (capital productivity) showed an average value of 44±2 and over 1942-2000 it showed an average value of 42±3 (Figure 1). Following the Panic of 1907, GDPpc/R began a steady decline reaching the mid 20's during the heights of the Depression. The ratio GDPpc/R can be thought of as a measure of sales per person (consumer) per unit of capital invested. A decline in the top line eventually falls to the bottom line, and after about a decade lag, pre-tax ROC also began to decline.

Figure 1. Declining capital productivity (GDPpc/R) led to the decline in ROC after 1916

So the fundamental problem was that capital ceased to produce as much output per person after 1907 as it had before. Less output means less pre-tax and post-tax ROC. Less ROC means more stock market and economic volatility to drive down P/R, which leads to increased risk, lower investment returns or both. Since the ethos of capitalism is the maximization of profit, this constraining of profit can be thought of as an operational crisis in capitalism, that culminated, decades later, in the general crisis of the Great Depression.

Capital that has a lower productivity than normal (i.e. between 1907 and 1942 and since 2000) does not have the same value as capital that produces normal output. We can account for changes in capital productivity by using the observation that during "normal" times GDPpc/R is approximately constant allowing us to express R as GDPpc/constant. If we use this in place of R we get:

1. P/R = constant * P / GDPpc

Figure 2 shows a plot of equation with a constant of 44 used before 1942 and 42 afterward. This shows how the market priced capital based on its performance. Three peaks stand out, 1929, 2000 and now.

Figure 2. P/R adjusted for capital productivity

That the stock market was in a bubble in 1929 was confirmed by the tremendous decline afterward. The situation in 2000 is different. Like 1929, 2000 was the end of a secular bull market during which conditions had long been favorable to investment and investors were understandably in a bullish mood. When the market began to decline, the Fed sprang into action and slashed interest rates. The real estate market, which had been in a bull mode for several years, but was not yet seriously overvalued, responded favorably and housing prices and building activity remained fairly strong during the 2001 recession, resulting in a milder downturn both economically and in the stock market than might otherwise had happened. A side effect of a strong real estate market during the recession was an even stronger real estate market after the recession, which resulted in a bubble in which real estate prices got about as overvalued as stocks had been in 2000. The aftermath of this bubble was much more severe on both the stock market and the economy. At the market bottom in 2009 P/R had fallen by two thirds from its high, and the US economy had experienced its first financial panic in 75 years.

Today the stock market is again approaching the valuations seen in 2000 and 1929. Unlike these peaks, today's peak does not show a particularly high level in terms of P/E or other conventional measures of valuation. Stocks don't look particularly high because the impact of the drop in capital productivity over the last decade has not yet filtered through to the bottom line. Rising profit margins since 2000 have worked to offset falling capital productivity so as to maintain ROC at pre-2000 levels (Figure 3).

Figure 3. Trends in capital productivity, profit margin and ROC since 1980

Concerns over the sustainability of current margins, have been expressed:

You have to pay close attention to Mauldin's reasoning, which I'm obliged to disclose are based on the opinions of investor John Hussman, a closely followed investment adviser. He says "while price/earnings multiples appear only moderately elevated, those multiples themselves reflect earnings that embed record profit margins that stand about 70% above their historical norms." Get that? Stocks are priced too high because expected record profits are too far above normal corporate earning power. By this rule of thumb, Maudlin reckons that "market valuations... are well above any point prior to the late-1990s market bubble." Meaning that it's 2000 all over again, as profit margins must revert to mean sooner or later.

Mauldin's analysis concludes that stocks are again at levels similar to 2000, which agrees with Figure 3. Using a different methodology John Hussman has also concluded that the market as of 2014 was extremely overvalued:

The upshot is that equities are likely to produce total returns close to zero over the coming decade. But they still present something of an "inventory" problem. The basic inventory problem is to accumulate inventory prior to advances in price, to hold that inventory as long as it appreciates in price, and to release that inventory when prices are elevated. What we observe at present is a market where the inventory now fetches record prices and is likely to enjoy little return for long-term holders, and suffer severe losses over the completion of the present cycle. But should short-term demand become even greater, one can't rule out a move to even higher prices and even more dismal long-term prospective returns - something to be celebrated by those who hold out long enough to sell at that point, but tragic for those who actually buy the inventory in the hope that it will be rewarding over time.

If we accept the likelihood of a major decline in the next few years how far is the market likely to fall? If the stock market returns to the normal secular bear pattern of a downward trend in P/R, then the coming bear market bottom should see P/R below 0.49. Even if the bear market were the start today it would likely take a couple of years to reach the bottom, as was the case in 2000-2002 and 2007-2009. Let us assume a bottom in 2018. Over those three years R can be expected to gain another 10% and be at around 2200. Multiplying this by 0.49 gives about 1080, about half of current levels. This is the best case.

Worst case would be if the coming bear market/recession were associated with another financial crisis. Should this happen there is little the Fed can do, interest rates are already near zero and three rounds of quantitative easing have already been tried and shown to have little positive impact on growth. Crises are all about confidence, specifically the loss of it (which is why historically they were called panics). With the recent experience of limited Fed effectiveness, it is unlikely Fed action along will be able to significantly moderate the extent of the decline.

This leaves direct government intervention. In the last crisis the market found a bottom shortly (thirteen market days) after the passage of the stimulus bill. And this had following the TARP bill four months earlier. Serious deflation never appeared and unemployment never approached the 25% level seen in the 1932-33 crisis. So it would appear policymakers were successful in preventing a worst-case outcome in 2009. Worst case for the coming bear market would be if no legislative action is taken this time. In this case the relevant examples would be come from the period after WW I (when modern secular cycles really began) to the period after 1950 when government economic intervention became routine. This period includes bear market bottoms in 1920, 1932, 1938 1942 and 1949. The P/R values for these are 0.27, 0.21, 0.45, 0.28, and 0.31, respectively. The average of these is 0.30, which when multiplied by 2200 yields 670, about the same value as that reached in 2009. Thus a range of lows between around 650 and 1050 is projected for the coming decline.

The consensus of the chart at the top of the page suggests a P/R value around 0.4 as most typical, which translates to a bottom around 900. So 900 would be the median forecast for the bottom. In terms of the DJIA, this translates to about 8000, that is, about a ten thousand point drop in the Dow.

References:
1. Alexander, Michael A. (2000) Stock cycles: why stocks won't beat money markets over the next twenty years, Writers Club Press, New York, p 44.