Good Logic and Bad Facts

One of the benefits of historically-informed investing is that it allows various investment perspectives to be evaluated from the standpoint of evidence rather than verbal argument. That’s particularly important during periods like today, when much of financial commentary on Wall Street can be filed into a folder labeled “it’s hard to argue with your logic, if only your facts were actually true.”

The inconvenience of historically-informed investing, at least during episodes of reckless speculation, is that it’s sometimes rewarding to follow the crowd in its embrace of invalid ideas, at least for a while. That has certainly been the case for a longer period during the recent half-cycle than in most market cycles across history. “In the short run,” as the legendary value investor Ben Graham once wrote, “the market is a voting machine,” and it’s quite clear that the votes aren’t always cast in favor of value or even fact. Still, Graham continued, “in the long run, it is a weighing machine.”

Despite our strong focus on value-conscious investing, we do place some of our attention on tracking that “voting machine,” largely through the uniformity or divergence of market internals across a broad range of stocks and security-types (when investors are inclined to speculate, they tend to be indiscriminate about it). The most effective information from market internals comes when steep overvaluation is joined by deteriorating internals (a warning sign), or when a substantial retreat in valuation is joined by early improvement in market internals (a constructive sign).

With market valuations obscene, extreme "overvalued, overbought, overbullish" syndromes in place, interest rates off the zero-bound, and market internals still showing evidence of dispersion, we remain convinced that investors face very steep downside risks over the completion of the current market cycle.

As is typical at market extremes, the focus of Wall Street’s narrative is on minimizing the consideration of these risks. Among the arguments we hear quite a bit is the notion that large market losses only occur when a recession unfolds, so until a recession unfolds, the risk of large market losses can be dismissed. Again, it’s hard to argue with that logic, if only the facts were actually true.

In any given week since 1928, there has been a 21% chance of an S&P 500 loss in excess of -20% within the following 18 months, as measured from the prevailing level of the S&P 500 at the time. Yet for more than 40% of those points, the U.S. economy was not in a recession within the preceding or following 6 months, and 33% saw no recession over the entire subsequent 18-month period. If we look at individual bear markets since 1940, nearly 40% them were associated with no recession at all.

Even when market losses have been associated with a recession, much of the damage has been exerted before the recession was widely recognized. By the time the typical U.S. recession starts, the S&P 500 is already down by more than -9%, extending that loss to an average of -12% within 5 weeks of the start of the recession, and -16% within the first 10 weeks. Even those figures rely on the back-dating of recessions by the National Bureau of Economic Research (NBER), as U.S. recessions are almost never recognized until months or quarters after they begin.

The chart below shows the S&P 500 Index (blue line, left scale), along with the percentage loss of the S&P 500 at each point in time, relative to its prior 2-year high (red line, right scale). The chart uses weekly closing data, so actual market losses on a daily closing basis were somewhat deeper. Recessions are shaded. Along with the general historical perspective, pay particular note to the steepness of market losses once they get underway. It’s certainly true that market losses tend to be deeper when they are accompanied by a recession, but again, substantial losses can emerge even before a recession is recognized, and even without one occurring at all.