Return Auto-Correlation as Implied by Option Prices

This paper estimates auto-correlation beliefs for equity returns. It builds on the statistical concept of variance scaling and uses the implied variance term-structure as its sole input. Hence, return persistence is quantified in a forward-looking manner. The linkage is derived in a non-parametric fashion, utilizing the stylized fact of long-range dependent volatility. On empirical data of the S\&P 500 index I observe that investors believe in trending returns during bull markets, and anti-persistence in bearish times. The deviation from full-randomness - that is, an absence of serial-dependence - frequently exceeds economic significance. Therefore, expected auto-correlation is fluctuating, which means that returns are of non-linear dynamics. While a heavy-tailed distribution causes short-termed crashes, I emphasize that the non-linear dynamic is a major amplifier of market meltdowns. Expected auto-correlation is thus a crucial metric for understanding the stability of financial markets. After comparing expectations with realizations, I detect a strong predictive potential from ex-ante implied on future realized return persistence. This means that the degree of random walk becomes predictable, which rises a broad variety of economic questions. Implications are discussed in the context of investor behavior, market efficiency, the anatomy of meltdowns and investment opportunities.

Dissertation Paper