Earlier in the year, I wrote a couple of articles about why those who argued that weak interest rates were failing to confirm the stock market rally were wrong in insisting on a connection between the two. A glance at the last forty years of data shows that stock markets can rise quite well without interest rates rising. Over the course of the last fifteen years, it was often true that interest rates and stocks were correlated, but there was nothing to say that they either should or should not be. Sometimes they correlate, and sometimes they don't.
Rather, I argued that the most important thing to nail down first was the secular disposition of the market, chiefly by the link between commodity prices and the earnings yield (the inverse of the P/E ratio). In other articles, I have written about the tight correlation between commodity prices and the earnings yield for as far back as we have data on the earnings yield (140 years), and why this link probably goes back to the early 1700s. Over the last forty years, since the fall of Bretton Woods during the Nixon administration and the rise of the unfettered dollar, this correlation has only grown stronger.
Moreover, using monthly commodity price data, we can identify how particular commodities interact with the earnings yield. A secular rise in equity yields begins with a surge in energy prices, especially oil, along with weakness in precious metals. When the surge in equity yields concludes, precious metals surge and oil peaks or falls. The ratio between the prices of oil and gold provide a simple history of the vagaries of the P/E ratio since the 1970s.
In 2008-2009, many commodities crashed, oil most spectacularly. That was the first indication that the secular disposition of the markets was changing. In 2011, just as in 1980, silver and gold went parabolic, the death knell of a commodity "supercycle."
The particular relationships between given commodities and the earnings yield will probably not endure over the long run, but the stability of the relationship between the commodity sector as a whole and the earnings yield suggests that commodities can provide a reliable indicator of the secular disposition of the market. The current stock market rally has been confirmed by weakness in commodities generally and by the sometimes startling way in which precious metals have fallen this year.
Having outlined the relationship between stocks and the supercycles, we can now start to integrate how interest rates interact with stocks.
Treasury yields, when extracted from their long-term trends, exhibit a fairly regular cyclicality. At the mid-point and end of every decade (e.g., 1974 and 1980), rates rally dramatically and peak. Since short-term rates tend to move more dramatically, throwing the yield curve upside down, it is not surprising that we have recessions at the end of every decade.
There is a similar sort of cyclicality in employment data. A fall in unemployment tends to reverse or halt at these points and employment intensity as measured by hours worked per week (AWHMAN) tends to fall. (This relationship is discussed in detail in Edward Leamer's Macroeconomic Patterns and Stories).
The same is true for the earnings yield. John Hussman has argued that the correlation between interest rates and the earnings yield over the last four or five decades is an historical anomaly, but the clearly demonstrated historical relationship between commodity prices and the earnings yield, along with the equally well demonstrated correlation between commodity prices and interest rates from 1730-1914, suggests that interest rates are genetically predisposed to correlate with the earnings yield. The breakdown in that relationship between World War I and the beginning of the Cold War seems to have been an artifact of the transition from Pax Britannica and its gold standard to Pax Americana and its dollar standard.
As you can see in the chart below, since the decisive victory of the dollar standard, both the earnings yield and interest rates have experienced the same sort of cyclicality. [Note: Throughout this article, I will be calculating the earnings yield using concurrent earnings rather than the Shiller CAPE. Similar results can be found using the trailing twelve-month yield (TTM)].
(Source: Unless otherwise indicated, all data in charts are derived from Robert Shiller's Irrational Exuberance).
And since the earnings yield is nothing more than the ratio of corporate earnings to stock prices, we can take a look at those two factors to see how they relate to that cyclicality.
The relationship is interesting. It appears that profits experience waves of accelerated and decelerated growth over five years, just as the earnings yield and interest rates do. That is fairly interesting and surprising, I think. But, what is really startling is how stock prices react to those shifts in earnings. In simple terms, we can say that when the stock market (more specifically, P/E ratios) is in a secular bull market, stocks rally strongly when profits decelerate, but when stocks are in a secular bear market, stocks correlate with profits.
Below, I have detrended earnings using moving averages and the post-1950 trend in earnings. Although the magnitudes vary, the story remains the same with respect to the trends.
If you did not already know, you can see that earnings have been moving relentlessly upwards for the last sixty years. Although earnings can run a bit above trend (roughly speaking, 1970s and 2000s) or below trend (roughly, 1980s and 1990s), they do not seem to experience the same sort of volatility as do equity prices.
But, that sort of depends on the secular disposition of the market. In the 2000s, for example, equities, earnings, and the earnings yield were all correlated with one another. The only way this is possible is if earnings were more volatile. In the 1980s and 1990s, equities, earnings, and the earnings yield did not correlate. The earnings yield was way down, equity prices were way up, and earnings were up but below the post-war trend.
In other words, the stock market likes great earnings, but what it really loves, what really makes it go bananas, is good earnings. What it hates is sudden drops in earnings.
It is perhaps best to look at this from a cyclical perspective: in the chart below is a comparison of the S&P 500 index and its earnings divided by their respective five-year moving averages. From 1980 to 1999, the biggest rallies began when profits began to decelerate. They almost look inversely correlated, although that is not quite the case (the correlation is 0.03). The most infamous rallies of the period (1987 and the period of "irrational exuberance" in the late 1990s) coincided with these mid-cycle (counting from peak to peak) decelerations.
For the period from 2000 to 2010, however, the correlation between detrended earnings and stocks is 0.70. Despite the surge in profits and the stock market's positive response, the rallies were actually rather subdued.
Another way to state this, then, is to say that in a secular bull market, the best rallies occur during earnings decelerations, while in a secular bear market, the best rallies occur during earnings accelerations.
Now, since stock market rallies coincide with decelerations or outright declines in earnings during secular bull markets, one can imagine how this impacts P/E ratios (sharply up!) and, therefore, the earnings yield (sharply down!). On this cyclical basis, therefore, stocks are inversely correlated with the earnings yield during bull markets and, because stocks react positively but feebly to earnings in secular bear markets, stocks and yields are positively correlated during secular bear markets.
And, because the earnings yield and Treasury yields tend to (or at least have tended to) correlate with one another, we would expect a similar sort of relationship between the stock market and interest rates.
One can see that the earnings yield and interest rates do not peak and trough at the exact same times. Peaks and troughs for the respective yields might be off up to a year and a half.
In previous discussions of interest rate cycles, I have written about the powerful snapback interest rates experience after hitting a cyclical bottom. In the 1987 instance, interest rates troughed in March, well before the earnings yield, and then snapped up violently into the fall. The earnings yield was still bottoming out, therefore, while interest rates and the stock market briefly rose in tandem in the middle of 1987.
This seems to be what has happened in the current cycle. Stocks rallied from 2011 to 2012 while long yields dropped. Stocks continued to rally as Treasury rates rose, and it is possible that those rates are now at or near their cyclical tops (setting aside the question of whether or not the Fed will raise short-term rates in the next couple of years). Another similarity between this cycle and the 1987 cycle is that both were preceded by a compression of the yield curve due to a rapidly falling long rate but without inverting. The difference, however, is that there has been no significant rally in the price of crude oil so far.
The rally, therefore, remains intact.
But, for the purposes of this article, it should be clear that due to the relationship between the stock market and corporate earnings on the one hand, and the earnings yield and interest rates on the other (at least on an historical basis), that the relationship between stocks and interest rates in the dollar era hangs entirely on the secular disposition of the market.
For those who have been staying out of the market because they have been waiting for a downturn in profits, not only might there be no more than a deceleration in profits, but even if there is a significant downturn, as long as it is not overly precipitous, it is more likely than not that stocks will respond even more bullishly than they have so far. If profits were to accelerate, stocks would probably continue to rise, although perhaps at a more deliberate pace.
It should be pointed out that this is definitely a post-war phenomenon most strongly expressed since the early 1970s. As you can see, prior to the establishment of the Federal Reserve, equity prices moved in lock-step with earnings (somewhat contrary to Shiller's depiction of this period in Irrational Exuberance).
It remains for us to discuss the 1970s, which I have avoided up until now, because it is both more complicated than the period after the 1970s and because it seems to confirm everything I have said here and elsewhere about the relationship between commodities, stocks, profits, interest rates, and the earnings yield.
In previous articles about the 1970s, I have argued that the decade was primarily defined by the secular bear market in equities and P/E valuations and the secular bull in commodities. I have also, however, argued that the mid-1970s experienced a brief and sudden pole-reversal. After the oil shock of 1973-1974, the oil/gold ratio collapsed, commodity prices fell dramatically (remember gold collapsing?), the stock market surged, P/E ratios rose, inflation halved, and Gerald Ford almost beat Jimmy Carter. In late 1976, the oil/gold ratio was back at historic highs, stocks flattened out, the commodity rally was back on, and inflation was back in style.
The behavior of stocks relative to earnings coincides with that narrative perfectly. Prior to 1973, stocks and earnings were positively correlated, but throughout 1973-1975, they were negatively correlated. Earnings peaked in typical, cyclical fashion in late 1974, while stocks tanked. But then, as earnings decelerated, stocks rallied. By 1976, however, and up until the early 1980s, stocks and earnings were correlating once again, with equities struggling to keep up.
Understanding the secular mode of the market and how it interacts with each asset class is key to improved returns. Correctly identifying it in a given timeframe helps one to adjust from playing last decade's game (commodities and BRICs) to playing this one's (U.S. stocks).
So, keep an eye on profits and interest rates, but do not be surprised if the markets fail to respond how orthodoxy says they "should." Commodities and the relationship between stocks and interest rates strongly suggest that this is a market powered by P/E multiple expansion, and history suggests that bull markets respond differently with respect to profits and interest rates than do secular bear markets such as we became accustomed to in the 2000s. Weak profits might be just what this market is looking for. It is best to adhere to the secular trend, avoid the quarterly guessing game, and stay long (NYSEARCA:SPY) until you can see the whites of the next commodity bull's eyes.