Scholarlink

<p>We reassess the main conclusions of Boehmer et al. (J Finance 76:2249–2305, 2021) on the predictive power of retail order imbalance (ROI) for future stock returns. First, we examine whether their findings for 2010–2015 persist in the 2016–2021 period. We find that ROI’s predictive power weakens considerably: past ROI no longer predicts weekly returns on large-cap stocks, and the associated long–short strategy is no longer profitable. Second, we assess the impact of using the quote midpoint (QMP) method instead of BJZZ’s original algorithm. While QMP may better identify individual trades, it does not materially alter BJZZ’s conclusions for the original period. However, in the 2016–2021 sample, the QMP method provides stronger support for BJZZ’s key results than their own algorithm. Differences in findings across the two methods are largely driven by bid-ask spreads and become more pronounced in the later period, when spreads are, on average, wider.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Revisiting Boehmer et al. (2021): recent period, alternative method, different conclusions

  • 期刊论文

  • 发表日期：

    2025年9月18日

查看全文
  • David Ardia,
  • Clément Aymard,
  • Tolga Cenesizoglu

摘要

We reassess the main conclusions of Boehmer et al. (J Finance 76:2249–2305, 2021) on the predictive power of retail order imbalance (ROI) for future stock returns. First, we examine whether their findings for 2010–2015 persist in the 2016–2021 period. We find that ROI’s predictive power weakens considerably: past ROI no longer predicts weekly returns on large-cap stocks, and the associated long–short strategy is no longer profitable. Second, we assess the impact of using the quote midpoint (QMP) method instead of BJZZ’s original algorithm. While QMP may better identify individual trades, it does not materially alter BJZZ’s conclusions for the original period. However, in the 2016–2021 sample, the QMP method provides stronger support for BJZZ’s key results than their own algorithm. Differences in findings across the two methods are largely driven by bid-ask spreads and become more pronounced in the later period, when spreads are, on average, wider.