Trade Size Based Order Imbalance and the Cross-Section of Stock Returns

Trade size-based order imbalance and its impact on market returns, with findings indicating a shift in trading behavior around 2000.

Paper Metadata

• Publication Date: 2024-11-08

• Source: SSRN

• Link: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4968938

• Authors:

  • David A. Lesmond, Tulane University - A.B. Freeman School of Business;

Keywords

• Market Microstructure

• Trade Size

• Order Imbalance

Notes for Review

Recommendation: 98%

This study provides significant insights into the predictive power of trade size-based order imbalance on future security and market returns. The authors conduct a comprehensive analysis of trade size and returns, using a dataset spanning from 1980 to 2023. Their findings indicate that medium-sized trades (1000-4999 shares) have significant predictive power in predicting future abnormal returns. However, the authors also find a structural break in 2000, where prior to the break, traders realized positive abnormal returns, indicative of informed trading, while post-break returns were negative, suggesting a shift to uninformed trading. The study's findings are robust to various tests and controls, and have important implications for market microstructure and high-frequency trading strategies. The authors' use of a habitat-based approach to studying asset prices and their ability to bridge the gap between earlier research on trade size and the evolving landscape of pricing behavior make this study a valuable contribution to the field.

The dataset for this study spans January 1980 to December 2023, including all ordinary common stocks listed on the NYSE, AMEX, NASDAQ, and NYSE Arca. Primary data sources include the Institute for the Study of Security Markets (ISSM) from 1983–1992, the Trade and Quote (TAQ) dataset for 1993–2023, and intraday trade data from Francis Emory Fitch for NYSE and AMEX-listed firms from 1980–1982. These datasets provide comprehensive trade size information, including clustering around specific sizes such as 1,000 to 4,999 shares, which proved critical to the study's insights. Key data processing steps included employing the Lee and Ready (1991) algorithm to classify trades, eliminating retail trades, and filtering trades exceeding a daily cap of 55,000 shares to refine the sample for institutional trades.

This study reveals that trade size-based order imbalance, particularly in the 1,000–4,999 share range, has significant predictive power over future security returns. A key finding is a structural break in 2000: pre-2000, medium-sized trades positively correlated with abnormal returns, indicating informed trading. Post-2000, the relationship reversed, reflecting uninformed trading behaviors, potentially driven by market evolutions like decimalization and Regulation FD. These results are robust across multiple tests, such as the Fama-French models and Fama-MacBeth regressions. Notably, the predictive power of medium-sized trades extends to broader market indicators like S&P 500 premiums. The findings contribute meaningfully to understanding market microstructure, emphasizing the changing dynamics of informed versus uninformed trading strategies across evolving regulatory landscapes

---

Abstract

This study conducts a comprehensive analysis of trade size-based order imbalance and its impact on future security and market returns. It particularly highlights the predictive power of medium-sized trades (1000 to 4999 shares) in predicting future abnormal returns that is significantly influenced by a structural break in 2000. Prior to the break, traders realized positive abnormal returns, indicative of informed trading. Conversely, post-break returns were negative, suggesting a shift to uninformed trading. These findings are robust to a battery of tests and controls known to affect predictions of future security and market returns.