Why a 60-65% Market Loss Would Be Run-Of-The-Mill

"In the summer of 1929 the surface of Wall Street was a mixture of placidity and mania – stock averages at record highs and still headed upward, the dissenters momentarily routed … Roger Babson said to an audience at a routine New England financial luncheon, ‘I repeat what I said at this time last year and the year before, that sooner or later a crash is coming.’ As Babson implied, his earlier warnings had been roundly ignored… When the crash finally came, it came with a kind of surrealistic slowness – so gradually that, on the one hand, it was possible to live through a good part of it without realizing it was happening, and, on the other hand, it was possible to believe that one had experienced and survived it when in fact it had no more than just begun."

– John Brooks, Once in Golconda, 1969

From September 3, 1929 to July 8, 1932, the Dow Jones Industrial Average fell by -89.2%, though certainly not in one fell-swoop. In fact, the decline known as the “1929 Crash” took the Dow down by an initial -47.9%, setting a trough on November 13, 1929. That initial decline was followed by a 48.0% recovery that peaked on April 17, 1930, leaving the Dow still -22.9% below its bull market high, because that’s how compounding works. By the 1932 low, the Dow had plunged -86.0% below its April 1930 peak, and -79.3% below even the “bottom” it set in November 1929 after losing nearly half of its value.

One might view the very comparison of present stock market conditions to 1929 market peak as exaggerated and preposterous, but then, one would be wrong. The fact is that on the valuation measures we find most strongly correlated with actual subsequent long-term and full-cycle market returns across history (and even in recent decades), current market valuations match or exceed those observed at the 1929 peak.

Likewise, valuations for nearly every decile of stocks presently exceed those observed at the 2000 market peak. As we’ll see below, the extreme valuation of capitalization-weighted indices like the S&P 500 at the 2000 peak was driven by single decile of stocks, largely represented by large-cap technology stocks that collapsed by -83% during the subsequent bear market. At present, every decile of stocks, without exception, is sufficiently overvalued to allow market losses on the order of -59% to -71%, without even breaching their respective valuation norms.

Still, as we’ll also discuss, the behavior of market internals continues to suggest that investors have a speculative bit in their teeth, though tenuously enough that it could drop out on even a few sessions of weak or divergent market action. Still, we try to align with prevailing internals rather than forecasting shifts, so despite the likelihood of absurdly steep market losses over the completion of this cycle, our very near-term outlook is rather neutral, and will remain so until we observe broader divergence and fresh deterioration in our measures of internals.

Why a 60-65% market loss would be run-of-the-mill

We begin with the Hussman Margin-Adjusted P/E (MAPE), which is among the most reliable valuation measures we’ve tested in market cycles across history, based on its correlation with full-cycle and 10-12 year market returns. The MAPE is second in reliability only to MarketCap/GVA – the ratio of nonfinancial market capitalization to nonfinancial corporate gross value-added (including estimated foreign revenues). Both measures essentially act as broad, apples-to-apples market price/revenue ratios, and significantly outperform popular earnings-based measures like price/forward operating earnings, the Fed Model, and the Shiller P/E. While both are at similar extremes, the MAPE has a longer data history, which allows direct comparison with 1929 valuation levels.

One might imagine that valuation levels might have simply “shifted higher” in recent decades, but that would miss the fact that associated subsequent returns have also “shifted lower” in recent decades, leaving the mapping between valuations and subsequent returns unaffected.