What Daily Stock Returns Tell Us About the Economy

Daily stock returns contain a meaningful real-economy signal once their short-term noise is separated from their persistent, low-frequency component.

• Fabrizio Ghezzi’s “microfiltering” method combines daily S&P 500 returns with nonfarm payrolls, unemployment, industrial production, and capacity utilization through a real-time Kalman filter that respects actual release dates and preliminary data.
• Estimated from 1950 through 2019, the factor falls sharply in every NBER-dated recession, has an autoregressive coefficient near 0.99, and displays left skew and fat tails consistent with asymmetric downturn risk.
• The factor correlates 0.80 monthly with the Chicago Fed National Activity Index and 0.50 quarterly with real GDP growth; it explains about 50% of next-month CFNAI variation and 18–30% of next-quarter GDP growth.
• It also correlates 0.74 with the Philadelphia Fed’s daily Aruoba-Diebold-Scotti Business Conditions Index, has an adjusted R² of roughly 55% at one-day and five-day horizons, and correlates 0.64 with the ISM Manufacturing PMI.
• The factor predicts dividend growth, net equity payouts, realized market variance, credit spreads, and the term spread, but has no reliable predictive power for stock returns or inflation; it updates daily rather than only when official macroeconomic data are released.

Takeaway: Investors should not infer economic conditions from raw daily market moves, but a properly filtered version of equity returns could provide earlier recession warnings and a more timely view of economic activity, while highlighting the downside asymmetry of macroeconomic risk.