We're Witnessing the Stock Market Do Something for Only the 3rd Time in 156 Years, and History Is Crystal Clear About What's Next for Stocks

The S&P 500's Shiller CAPE ratio stood at 41.60 on Sept. 21, versus its 156-year average of 17.42, marking only the third period in history in which the valuation measure has exceeded 40. Prior episodes preceded major bear markets, including 49% and 78% declines in the S&P 500 and Nasdaq after the dot-com peak, respectively. The article argues that AI infrastructure spending, strong earnings, buybacks and IPO activity have supported record highs, but elevated valuations create meaningful downside risk; it nonetheless frames any bear market as a long-term buying opportunity.

Analysis

CAPE is a poor short-horizon timing tool, but it matters as an asymmetry signal: at elevated starting valuations, forward index returns become more dependent on earnings delivery and real-rate stability than on further multiple expansion. The more actionable vulnerability is concentration. If AI infrastructure spending normalizes, the earnings revision risk is concentrated in NVDA and the hyperscaler-capex complex, while passive SPY/QQQ flows transmit that shock broadly despite limited fundamental exposure in most constituents.

The key second-order issue is that buybacks can cushion per-share earnings but cannot offset a de-rating if financing costs rise or free cash flow is redirected toward AI capex. A weakening labor or consumption backdrop would also expose the gap between index-level earnings resilience and cyclically sensitive breadth; equal-weighted exposure should outperform cap-weighted indices in a benign rotation, but not in a recessionary drawdown.

Over the next days to weeks, this article itself is not a catalyst and should not prompt directional de-risking. Over 1-3 months, watch long-end Treasury yields, upward versus downward S&P 500 EPS revisions, and hyperscaler capex guidance. The bearish setup is falsified by continued broadening of earnings revisions beyond mega-cap technology, falling real yields, and evidence that AI monetization—not merely infrastructure spend—is supporting software and enterprise demand.

Contrarian view: the historical comparison overstates precision because today’s index has a larger share of high-margin, asset-light businesses and CAPE embeds a decade of unusually low-margin pandemic and pre-AI earnings. That does not justify chasing QQQ; it argues for replacing expensive beta with selective quality rather than treating a valuation threshold as an imminent crash signal.

Key Decisions for Investors

• Do not add broad QQQ beta into strength; for 1-3 month downside protection, buy 3-6 month QQQ put spreads approximately 5%/15% out-of-the-money, funded where possible by trimming concentrated NVDA exposure. The hedge is attractive only if implied volatility remains below levels seen during prior growth scares; risk is a continued melt-up driven by falling yields.
• Implement a 3-6 month relative-value rotation: long RSP versus short an equal-dollar QQQ position. Thesis is that any earnings broadening or multiple compression reduces the return premium from cap-weight concentration; exit if mega-cap EPS revisions continue to outperform the rest of the index by the next earnings cycle.
• Maintain NVDA as a tactical hold rather than a new core add until customer capex guidance validates incremental demand beyond current build-outs. A guidance reset from major cloud customers or evidence of rising accelerator lead times/inventory would be the trigger to reduce; upside risk is another step-up in sovereign or enterprise demand.
• For defensive equity exposure over 6-18 months, favor cash-generative, lower-duration sectors through XLV and XLP rather than raising portfolio cash indiscriminately. This is a volatility-management allocation, not a recession call; reverse if real yields decline materially and cyclical earnings revisions broaden.