Ding Granger Engle 1993
“A Long Memory Property of Stock Market Returns and a New Model” (Journal of Empirical Finance 1:83-106, 1993) is the canonical empirical study of long-range dependence in equity returns. Examining the power transformations |r_t|^d of S&P 500 daily returns, Ding, Granger & Engle document that while raw returns contain little linear autocorrelation, absolute and squared returns are strongly and persistently autocorrelated — volatility clustering — with a dependence that decays slowly (hyperbolically), not geometrically. It appears in this vault as the load-bearing evidence under the First-Order Memory Assumption failure mode: a memoryless first-order Markov Chain Trading Model cannot represent this slow-decaying persistence, because its transition law conditions only on the current state and discards the path. The paper also introduced the Asymmetric Power ARCH (A-PARCH) volatility model.
Connections
- First-Order Memory Assumption — contradicts, documents the long-memory dependence a first-order chain discards, source: https://www.cambridge.org/core/books/abs/essays-in-econometrics/long-memory-property-of-stock-market-returns-and-a-new-model/1AC2DDC7C61CEC07C2C653C25A592F4D
- Markov Chain Trading Model — relates, the memoryless assumption omits documented volatility persistence, source: https://www.cambridge.org/core/books/abs/essays-in-econometrics/long-memory-property-of-stock-market-returns-and-a-new-model/1AC2DDC7C61CEC07C2C653C25A592F4D