Excerpted from John Hussman’s Weekly Comment,

The atmosphere is getting thin up here, and every ounce counts triple when you’re climbing in rarefied air. While near-term market dynamics are more likely to be impacted by Friday’s employment report than any other factor, our broad view remains that stocks are in the late-stage top formation of the second most extreme episode of equity market overvaluation in U.S. history, second only to the 2000 peak, and already beyond the 1929, 1937, 1972, and 2007 episodes, not to mention lesser extremes across history.

On the economic front, much of the uncertainty about the current state of the economy can be resolved by distinguishing between leading indicators (such as new orders and order backlogs) and lagging indicators (such as employment). It’s not clear whether the weakness we’ve observed for some time in leading indicators will make its way to the employment figures in time to derail a Fed rate hike in December, but as we’ve demonstrated before, the market response to both overvaluation and Fed actions is highly dependent on the state of market internals at the time. Presently, we observe significant divergence and internal deterioration on that front. If we were to observe shift back to uniformly favorable internals and narrowing credit spreads, our immediate concerns would ease significantly, even if longer-term risks would remain.

Having reviewed the divergences we observe across leading economic indicators and market internals last week (see Dispersion Dynamics), a few additional notes on current valuations may be useful.

As I’ve noted before, the valuation measures that have the strongest and most reliable correlation with actual subsequent market returns across history are those that mute the impact of cyclical variations in profit margins. If one examines the deviation of various valuation measures from their historical norms, those deviations are rarely eliminated within a span of a year or two, but are regularly eliminated within 10-12 years (the autocorrelation profile drops to zero at that point). As a result, even the best valuation measures have little relationship to near-term returns, but provide strong information about subsequent market returns on a 10-12 year horizon. Among the most reliable valuation measures we identify, those with the strongest relationship with subsequent 12-year nominal S&P 500 total returns are:

• Shiller P/E: -84.7% correlation with actual subsequent 12-year S&P 500 total returns
• Tobin’s Q: -84.6% correlation
• Nonfinancial market capitalization/GDP: -87.6%
• Margin-adjusted forward operating P/E (see my 8/20/10 weekly comment): -90.7%
• Margin-adjusted CAPE (see my 5/05/14 weekly comment): -90.7%
• Nonfinancial market capitalization/GVA (see my 5/18/15 weekly comment): -91.9%

MarketCap/GVA is presented below to provide a variety of perspectives on current valuation extremes. The chart below shows this measure since 1947. We know by the relationship between MarketCap/GVA and other measures (with records preceding the Depression) that the current level of overvaluation would easily exceed those of 1929 and 1937, making the present the most extreme point of stock market overvaluation in history with the exception of 2000. In hindsight, the only portion of 2000 when stocks were still in a bull market was during the first quarter of that year.

To be as clear as possible: Over the near term, broad improvement in market internals and credit spreads would suggest a return to risk-seeking speculation that might defer the unwinding of this obscene Fed-induced speculative bubble, or could even extend it. But with market internals presently negative and credit spreads continuing to widen, the market remains vulnerable to an air-pocket, panic or crash, here and now. In either case, our expectation is that the completion of the current market cycle will involve a market loss of at least 40-55%; a loss that would merely take the most historically reliable valuation measures to run-of-the-mill pre-bubble norms, not materially below them.

Investors should remember from the 2000-2002 and 2007-2009 collapses that in the absence of investor risk-seeking – as conveyed by market internals – even aggressive Fed easing does not support stocks. The reason is that once investors become risk-averse, safe, low-interest liquidity is a desirable asset rather than an inferior one. So creating more liquidity fails to achieve what the Fed does so successfully and perniciously during a risk-seeking bubble: drive investors to chase yield and speculate in risk-assets.

The chart above places MarketCap/GVA on an inverted log scale (blue line, left scale), along with the actual S&P 500 nominal annual total return over the following 12-year period (red line, right scale). Note that current valuations imply a 12-year total return of only about 1% annually. Given that all of this return is likely to come from dividends, current valuations also support the expectation that the S&P 500 Index will be lower 12 years from now than it is today. While that outcome may seem preposterous, recall that the same outcome was also realized in the 12-year period following the 2000 peak.

Given both obscene valuations and clearly unfavorable market internals and credit spreads at present, we see extreme market losses as not only an immediate risk, but also as the predominant likelihood over the next few years.