Search-Based Peer Groups and Commonality in Liquidity
Paul Brockman, Dennis Y Chung, Neal M Snow
Review of Finance, Volume 27, Issue 1, February 2023, Pages 33–77, https://doi.org/10.1093/rof/rfab033
In this study, we examine the relationship between EDGAR-initiated, search-based peer (SBP) groups proposed by (Lee, Ma, and Wang, 2015) as a potential determinant of commonality in liquidity. Understanding the determinants of commonality is important since the liquidity risk associated with comovement can have a significant impact on a firm’s expected return. Our central hypothesis is that correlated searches (i.e., chronologically adjacent searches) within SBP groups will lead to correlated trading within these same groups and, consequently, to increased demand-side commonality in liquidity.
Our empirical analyses confirm that SBP membership is a statistically and economically significant source of commonality for both bid-ask spreads and volume depths. We also show that SBP-induced commonality has been increasing in significance over the past two decades. When we compare supply-side commonality from specialist-portfolio membership to demand-side commonality from SBP-group membership, we find that demand-side SBP membership has a stronger impact on both spread and depth commonality. This is a significant result, especially in light of Tham, Sojli, and Skjeltorp’s (2018, p. 2919) observation that, “Resiliency and the endogenous relation between liquidity demanders and suppliers are understudied but important in today’s electronic markets.” Finally, we track the IP addresses of EDGAR searches to distinguish between retail and institutional investor-initiated searches. The results show that retail investors play a major role in both SBP formation, and in SBP-induced commonality.
Overall, our study contributes to the market microstructure and disclosure literature in two main areas. First, we provide new evidence on the determinants of liquidity comovement by showing that membership in an SBP group increases a firm’s spread and depth commonality. In addition to identifying a new source of commonality, our findings also provide new evidence on the relative importance of demand-side versus supply-side commonality. By tracking the locations of the IP addresses that generate SBP groups, we show that retail investors are responsible for roughly 85% (54%) of the EDGAR searches that generate SBP groups (top-five SBP groups). Second, our study contributes to an emerging literature that uses internet searches to examine potential implications for investor behavior and stock market trading patterns. Our findings show that the identification of a significant source of comovement in liquidity is only possible through the application of a co-search algorithm to the SEC’s EDGAR database. Without access to investor-search data and SBP-forming algorithms, these liquidity patterns within SBP groups would have remained hidden from view. We expect that future research projects will identify other latent patterns through the application of SBPs to novel research questions.