A Stock Return Decomposition Using Observables
Prof. Annette Vissing-Jorgensen
Professor of Finance
University of California, Berkeley
We propose a method to decompose stock returns period by period. First, we argue that one can directly estimate expected stock returns from securities available in modern financial markets (using the real yield curve and the Martin (2017) equity risk premium). Second, we derive a return decomposition which is based on stock price elasticities with respect to expected returns and expected dividends. We calculate elasticities from dividend futures. Our decomposition is an alternative to the Campbell-Shiller log-linearization which relies on an assumption about the log linearization constant (ρ). An application to the COVID crisis in 2020 reveals that risk premium changes drove much of the crash and rebound in the S&P500 while a fall in long-term real yields drove a strong positive return for 2020 as a whole.