When Recessions Officially End — The Rally Comes Early
S&P 500 performance from each NBER-dated recession end
Recessions officially end long before it feels that way — unemployment is typically still high and headlines still grim when the economy quietly turns. The S&P 500 has often bottomed four to six months before the NBER-declared end of a recession, meaning the strongest gains arrived while conditions still looked terrible. This chart anchors market performance to each official recession end month from 1970 through 2020 and shows what came next.
| Date | 1M return | 1Y return | 5Y return |
|---|---|---|---|
| 1970-11-02 | +6.0% | +11.1% | +7.0% |
| 1975-03-03 | -0.7% | +20.5% | +37.3% |
| 1980-07-01 | +5.9% | +12.9% | +65.1% |
| 1982-11-01 | +2.4% | +20.6% | +68.1% |
| 1991-03-01 | +2.4% | +11.4% | +77.9% |
| 2001-11-01 | +4.2% | -16.9% | +26.1% |
| 2009-06-01 | -2.5% | +13.6% | +104.1% |
| 2020-04-01 | +14.6% | +62.7% | +104.9% |
What history says
Editorial commentary written by ALAN analysts. Figures cited below are analyst-authored context — they are not derived from the chart above and may reflect different windows or sources.
Markets have historically turned up months before recessions officially ended, pricing the recovery while the economic data was still deteriorating. Investors who waited for confirmation missed the sharpest part of the rebound.
At each official recession end, unemployment remained elevated and sentiment depressed. The disconnect between how the economy felt and what markets did next is the core lesson of this overlay.
Because the turn cannot be identified in real time, capturing recession-end rallies historically required already holding equities through the downturn — not predicting the inflection.
Rather than trying to time the recovery, consider maintaining a rebalancing discipline through downturns: reviewing the portfolio on a fixed schedule and restoring equity weights that drawdowns pushed below target positions it to participate in rebounds that begin before the news improves.