The Economic Risk of Ignoring Arithmetic

“Let us not, in the pride of our superior knowledge, turn with contempt from the follies of our predecessors. The study of errors into which great minds have fallen in the pursuit of truth can never be uninstructive... Men, it has been well said, think in herds; it will be seen that they go mad in herds, while they only recover their senses slowly, one by one... Truth, when discovered, comes upon most of us like an intruder, and meets the intruder’s welcome... Nations, like individuals, cannot become desperate gamblers with impunity. Punishment is sure to overtake them sooner or later.”

Charles MacKay, Extraordinary Popular Delusions and The Madness of Crowds, 1841

When we observe the greatest follies of our predecessors, the episodes of speculative madness that come most immediately to mind are the pre-crash bubble peaks of 1929, 2000, and to a lesser extent, 2007. Unfortunately, we are in the midst of yet another episode of equivalent speculative madness, but one that will only be recognized in hindsight, and in the recollections of our children. They too are likely to take pride in a feeling of superior knowledge, forgetting the same lessons, and eventually creating another bubble and collapse of their own. The herd mentality is human nature. As in 2000-2002 and 2007-2009, when the S&P 500 collapsed by 50% and 55% respectively, we’ll likely see that herd mentality expressed on the downside soon enough. That also is human nature.

The stock market bubble that ended with the September 1929 peak began in August 1921, running just a few days beyond 8 years in duration. The bubble that ended with the March 2000 peak began in October 1990, running fully 9 years and 5 months in duration. Those two episodes represent the longest bull markets in U.S. history. The current half-cycle began at the March 2009 low, and has now run 7 years and 10 months in duration, making it the third-longest advance in history, placing it just 2 months short of the 1929 instance, but a full year and 7 months short of the 2000 instance.

What’s notable here is that by the time the bubbles that ended in 1929 and 2000 reached a duration similar to the present, they were already experiencing a significant increase in volatility and the frequency of corrections. From a time-perspective, for example, 7 years and 10 months into the bull market that began in October 1990, it was August 1998, about the point that the S&P 500 took a nearly 20% dive. By July 1999, the S&P 500 had eclipsed its mid-1998 high, followed by a correction of over -12% during the next 13 weeks (taking the net gain from mid-1998 to just 5%). The S&P 500 then rallied again into December 1999, followed by a nearly -10% correction over the next 8 weeks. Immediately after the final market high in March 2000, the S&P 500 quickly gave up more than 11% in the following month, wiping out a year of gains, and the bear market had hardly even started.

Likewise, by the time the bull market from the 1921 low extended into late-1928, the market became much more susceptible to corrections. Following a peak in November 1928, the Dow Jones Industrial Average lost nearly -13% in a month. The next peak in February 1929 was followed by a -8% correction over the following two weeks. A fresh high in May 1929 was followed by a -10% correction over the following four weeks. The final high in September 1929 was followed by an initial drop of nearly -15% over the next 6 weeks. Put simply, once the bull market was as mature as the present one has become, further market gains were not smooth. Nor were they ultimately durable.

From a valuation perspective, the current run is similarly mature. The most reliable market valuation measure we’ve identified across history (having the strongest correlation with actual subsequent market returns) is the ratio of nonfinancial market capitalization to corporate gross value-added (MarketCap/GVA), with Warren Buffet’s old favorite, the ratio of market capitalization to gross domestic product, just slightly behind, followed somewhat further by a variety of measures that normalize earnings in one way or another (by comparison, neither the ratio of price/forward operating earnings nor the Fed Model even come close). By the time valuations reached levels similar to recent extremes during the advances toward the 1929 and 2000 bubble peaks, it was already mid-1929 and late-1999, respectively. The market losses over the subsequent 10-12 year periods directly reflected the extent of the corresponding overvaluation. At present, we estimate S&P 500 nominal total returns averaging just 0.8% annually over the coming 12-year horizon. If our measures of market internals were to improve materially, particularly across interest-sensitive sectors, we would be inclined to allow leeway for valuations to become even more extreme. But in the end, richer valuations would only imply steeper losses over the completion of the current market cycle, and even weaker returns on a 10-12 year horizon. For a review of current valuation extremes, see Economic Fancies and Basic Arithmetic.