Worried About the September Effect? Here’s What History Tells Us About Investing During What’s Generally Been the Worst Month for Stocks.

The article highlights the historical September Effect: since 1928, September has been the only month in which the S&P 500 has closed lower more often than higher, with an average decline of 1.1%. Recent weak Septembers included declines of 9.3% in 2022, 4.9% in 2023, 4.8% in 2021, and 3.9% in 2020. Despite seasonal downside risk, the article argues investors should continue buying quality, reasonably valued stocks because longer-term returns have historically outweighed September volatility; supportive factors include broadening market participation and Q2 EPS and revenue beats from 87% and 77% of S&P 500 companies, respectively.

Analysis

The calendar effect is not independently tradeable without confirming positioning, implied volatility, and breadth. A predictable seasonal pattern is typically arbitraged away; its practical relevance is that it can amplify an existing de-risking impulse if CTA exposure is extended, dealer gamma turns negative, or macro data reprice the Fed path. The more relevant near-term vulnerability is index concentration: a modest derating in NVDA and GOOG can outweigh improving participation elsewhere, creating a superficially broad selloff even if earnings revisions remain intact.

For the next 1-3 months, distinguish a flow-driven correction from an earnings-led drawdown. A 3-5% SPX decline with stable 2027 EPS estimates, narrowing credit spreads, and resilient equal-weight relative performance would be an opportunity to add cyclical and quality factor exposure; it would not justify abandoning AI leaders. Conversely, rising real yields, a widening HY OAS, or negative hyperscaler capex commentary would turn a seasonal pullback into a multiple-compression event, with long-duration AI equities carrying disproportionate downside.

The contrarian point is that broad-market dip buying can be crowded precisely when retail and systematic investors interpret a routine pullback as seasonal. Better risk-adjusted exposure is likely through dispersion: own companies with identifiable earnings revision support while hedging index beta, rather than adding unhedged SPY. NVDA and GOOG remain structurally supported only as long as AI monetization and cloud demand convert capex into sustained revenue growth; capex persistence alone is insufficient for multiple expansion over 6-18 months.

Key Decisions for Investors

• Do not initiate a standalone SPY short based solely on seasonality. Set an alert for a 3-5% SPX drawdown combined with VIX below its 12-month median; if realized volatility rises while implied volatility remains inexpensive, buy 1-3 month SPY put spreads rather than outright puts, targeting roughly 2:1 payoff with defined premium risk.
• Maintain a barbelled AI position: long NVDA and GOOG only against an SPY or QQQ hedge sized to reduce net beta by 30-50%. Reassess after the next hyperscaler earnings cycle; exit the relative long if aggregate cloud/AI revenue guidance weakens or if AI-related capex is reduced without offsetting monetization evidence.
• On an index pullback, favor equal-weight exposure via RSP over adding cap-weight SPY, provided equal-weight earnings revisions remain positive and HY spreads do not widen materially. This captures continued breadth improvement while reducing dependence on a small set of high-duration mega-cap names.
• Treat a sustained rise in 10-year real yields above the recent three-month range, or a 50bp+ widening in HY OAS, as thesis falsifiers for dip-buying. Under those conditions, rotate the hedge from SPY put spreads toward QQQ puts because duration-sensitive mega-cap technology should bear the larger valuation adjustment.