Alexander Ljungqvist is Professor of Finance at New York University’s Stern School of Business.
Is greater trading liquidity good or bad for corporate governance? In the paper, Liquidity and Governance, which was recently made publicly available on SSRN, my co-authors (Kerry Back and Tao Li) and I address this question both theoretically and empirically. A liquid secondary market in shares facilitates capital formation but may be deleterious for corporate governance. Bhide (1993) argues that greater liquidity reduces the cost to a blockholder of selling her stake in response to managerial problems (‘taking the Wall Street walk’), resulting in too little monitoring by large shareholders. Bhide’s work has spawned an active literature on the effects of liquidity on governance. The present paper makes two contributions to that literature: (i) we solve a theoretical model consisting of an IPO followed by a dynamic Kyle (1985) market in which the large investor’s private information concerns her own plans for taking an active role in governance and show that greater liquidity leads to lower blockholder activism, and (ii) we verify the negative theoretical relation between liquidity and activism using three distinct natural experiments.
Liquidity has opposing effects on governance because it facilitates both block acquisition and block disposition (Maug, 1998). We show that this is true in the dynamic Kyle model: the probability of blockholder activism increases (decreases) with the amount of liquidity trading if the blockholder’s initial block is smaller (larger) than a certain critical value. We assume that the initial block is determined in an IPO. Following the analysis of IPO mechanisms in Stoughton and Zechner (1998), we show that optimal mechanisms lead to a block of sufficient size so that liquidity is harmful for governance (a conclusion opposite to that of Maug). A novel aspect of the dynamic Kyle model we study is that the realized sign and magnitude of liquidity trading affect the blockholder’s choice about becoming active and so affect the ultimate value of the stock. If liquidity traders happen to sell shares, the blockholder is likely to buy shares and become active; conversely, if liquidity traders buy shares, the blockholder is likely to take the Wall Street walk.
Our empirical results support our theoretical results. Establishing the causal effect of liquidity on governance is empirically challenging because, as Edmans, Fang, and Zur (2013) note, liquidity and governance are likely jointly determined by a firm’s unobserved characteristics. To address this challenge, we use three natural experiments: brokerage closures (Kelly and Ljungqvist, 2012), market maker closures (Balakrishnan et al., 2013), and mergers of retail with institutional brokerage firms (Kelly and Ljungqvist, 2012). Events of the first two types exogenously reduce liquidity and events of the third type exogenously increase liquidity. For two of them, we can even sign the direction of the resulting change in liquidity trading: as Kelly and Ljungqvist show, liquidity traders sell in response to brokerage-closure shocks and buy in response to retail brokerage-mergers.
In all three experiments, we find that blockholder activity, as measured by hedge fund activism and the number of shareholder proposals submitted in opposition to management, increases when liquidity decreases and vice versa. These findings suggest that for the average stock market-listed firm in the U.S., greater trading liquidity is harmful for governance. They stand in contrast to prior empirical work that treats the level of a firm’s trading liquidity as exogenous (for example, Norli et al. (2010)) or that uses decimalization as a shock to liquidity. A potential explanation for the difference in results is that decimalization, which undoubtedly improved some aspects of liquidity, coincided with some other aggregate shock that independently improved governance (such as Regulation Fair Disclosure). The staggered nature of the 43 brokerage closures, the 50 market maker closures, and the six retail brokerage-mergers we use makes it highly unlikely that our results are confounded in a similar way.
The full paper is available for download here.