Short-Duration Equity Return Puzzle

Abstract

Short-duration dividend strips have high conditional Sharpe ratios during crises, far beyond the theoretical upper bound. The finding is called the short-duration equity return puzzle. Using dividend prices and analyst dividend forecasts, this paper constructs the forward-looking required rate of return and the conditional Sharpe ratio for S&P 500 dividend strips over the period between 2002 and 2021. This paper also verifies that analyst dividend forecasts are reasonably accurate. During the COVID-19 pandemic crisis, the required rate of return of the 1-year dividend peaked at 55%, and its conditional Sharpe ratio rose to 14. This paper simulates the Hansen-Jagannathan theoretical upper bounds of the conditional Sharpe ratio for four mainstream macro-finance models—the preferred habit model (Campbell & Cochrane 1999), the long-run risk model (Bansal & Yaron 2004), the rare disaster model (Gabaix 2012), and the term structure model (Lettau & Wachter 2011). These predicted upper bounds are 2 to 17 times lower than the observed maximum. The finding is robust to construction methods of dividend prices, measurement errors, and transaction costs.

Keywords: Dividend Strip; Conditional Sharpe Ratio; Stochastic Discount Factor

JEL Classification: G12, G17, G01