Working Paper: NBER ID: w30366
Authors: Yacine Atsahalia; Jianqing Fan; Lirong Xue; Yifeng Zhou
Abstract: This paper studies the predictability of ultra high-frequency stock returns and durations to relevant price, volume and transactions events, using machine learning methods. We find that, contrary to low frequency and long horizon returns, where predictability is rare and inconsistent, predictability in high frequency returns and durations is large, systematic and pervasive over short horizons. We identify the relevant predictors constructed from trades and quotes data and examine what determines the variation in predictability across different stock's own characteristics and market environments. Next, we compute how the predictability improves with the timeliness of the data on a scale of milliseconds, providing a valuation of each millisecond gained. Finally, we simulate the impact of getting an (imperfect) peek at the incoming order flow, a look ahead ability that is often attributed to the fastest high frequency traders, in terms of improving the predictability of the following returns and durations.
Keywords: high-frequency trading; predictability; machine learning; stock returns; market efficiency
JEL Codes: C45; C53; C58; G12; G14; G17
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| predictors derived from trades and quotes data (C58) | predictability of returns (G17) |
| predictors derived from trades and quotes data (C58) | predictability of durations (C41) |
| characteristics of stocks (G10) | predictability of returns (G17) |
| market conditions (P42) | predictability of returns (G17) |
| imbalance in the limit order book (C69) | predictability of returns (G17) |
| recent transaction imbalances (F32) | predictability of returns (G17) |
| predictability of returns (G17) | predictability vanishes to 0 after 5 minutes and 2000 transactions (C69) |