COMMONALITY IN LIQUIDITY OF NEARBY FIRMS

• Author: Xiaoqiong Wang,Chengcheng Li
• Published date: 01 December 2019
• Date: 01 December 2019
• DOI: http://doi.org/10.1111/jfir.12193

The Journal of Financial Research Vol. XLII, No. 4 Pages 675–711 Winter 2019

DOI: 10.1111/jfir.12193

COMMONALITY IN LIQUIDITY OF NEARBY FIRMS

Chengcheng Li

Dongbei University of Finance and Economics

Xiaoqiong Wang

Indiana University Kokomo

Abstract

In this article, we examine commonality in liquidity of firms headquartered in the

same states and how the local liquidity commonality is influenced by firm‐and state‐

level characteristics. We document strong liquidity comovement of nearby firms.

Moreover, firms that change headquarters location experience a decrease in their

liquidity commonality with firms in the old states and an increase in their liquidity

commonality with firms in the new states. Our findings show that both firm‐and

state‐level characteristics determine local liquidity comovement. Local liquidity

commonality is stronger for firms with smaller size and lower level of institutional

ownership. Our results also suggest that state‐level volatility, state personal income,

state investment income, and state turnover commonality explain the local

component of liquidity commonality. We further document that the four state‐level

factors perform differently during volatile market periods.

JEL Classification: G10, G12, G23

I. Introduction

The liquidity of a stock and its associated liquidity risk are essential determinants in

expected stock returns and investors’decisions (Amihud and Mendelson 1986; Amihud

2002; Pástor and Stambaugh 2003; Acharya and Pedersen 2005; Liu 2006; Lee 2011).

Commonality in liquidity has been widely documented in prior studies. It refers to the

impact of a common factor on the liquidity of firms. Extensive research has shown

significant commonality in liquidity of stocks traded on the U.S. market (Chordia, Roll,

and Subrahmanyam 2000; Coughenour and Saad 2004; Kamara, Lou, and Sadka 2008;

Hameed, Kang, and Viswanathan 2010; Koch, Ruenzi, and Starks 2016; Hasbrouck and

Seppi 2001; Huberman and Halka 2001). In addition, commonality in liquidity has been

shown to exist in other markets, such as Hong Kong (Brockman and Chung 2002), Japan

(Stahel 2005), and the United Kingdom (Stahel 2005). Other studies document liquidity

commonality in an international setting, showing that liquidity commonality exists

worldwide at the country level (Karolyi, Lee, and Van Dijk 2012; Deng, Li, and Li 2018)

and the stock exchange level (Brockman, Chung, and Pérignon 2009). In addition to

We thank the editor (Murali Jagannathan) and an anonymous referee for valuable suggestions and

comments. We also thank Qinghai Wang for providing firm headquarters location data.

675

© 2019 The Southern Finance Association and the Southwestern Finance Association

market liquidity commonality, the literature has documented that stock liquidity comoves

with the liquidity of firms in the same industries (Chordia, Roll, and Subrahmanyam

2000). These studies focus on either the market component or the industry component of

liquidity commonality. However, little is known about the comovement in liquidity of

firms headquartered in nearby locations.

In this article, we investigate whether the liquidity of nearby firms comoves,

namely, the local component of liquidity commonality. Specifically, we examine the

liquidity comovement of firms headquartered in the same states. Firm location matters

for both asset pricing and corporate finance. Previous studies suggest that investors are

locally biased in their portfolio choices and that the markets are segmented both

internationally and within a country. The literature has documented that firm location is

important for stock return comovement (Pirinsky and Wang 2006), acquisitions and

leverage (Almazan et al. 2010), dividend policy (Becker, Ivković, and Weisbenner

2011; John, Knyazeva, and Knyazeva 2011), and investment policy (Dougal, Parsons,

and Titman 2015). The evidence on the importance of firm location suggests that the

U.S. financial market is not fully integrated and that geographic segmentation exists.

These studies indicate that the fundamental supply and demand sources for firms vary

across states. As a result, firm liquidity may comove with the local liquidity portfolio

and the comovement may vary across states.

Our article focuses on the local liquidity commonality of the U.S. stock market.

To start with, we define the local liquidity portfolio as the value‐weighted average

liquidity of stocks headquartered in the same states. Our methodology follows Chordia,

Roll, and Subrahmanyam (2000) and adds variation in the local liquidity portfolio in

addition to the market and industry liquidity portfolios. We find a positive and significant

local liquidity beta, with an average of 0.249 during 1980–2016. The evidence indicates

that the local liquidity component plays an imperative role in explaining stock liquidity

variation even after controlling for the market and industry components. To ensure that the

results are not driven by unobservable factors, we perform placebo tests where we replace

the state liquidity portfolio for each firm each year with that of a randomly selected state.

The distribution of the coefficients for the placebo state liquidity portfolio shows that the

liquidity commonality of a firm with a randomly selected state liquidity portfolio is

significantly smaller in magnitude. Our findings suggest that local liquidity commonality

is more prevalent between the liquidity of a stock and its own state liquidity portfolio.

We also show that our results are robust across different measures and liquidity

comovement specifications. First, we find that the local liquidity factor still matters

when we use equal‐weighted liquidity portfolios, or employ a bid–ask spread‐based

measure (realized spread) rather than the Amihud (2002) illiquidity measure. Second,

we show that local liquidity commonality is still significant when using the stock

pairwise correlations in liquidity to proxy for the stock liquidity comovement. We find

that being in the same state increases the liquidity correlation of the stock pair. Finally,

we show that our results are unlikely to be the artifacts of using state borders as proxies

for geographic proximity. In particular, the results still stand when we use the actual

distance between two firms’zip codes rather than the same state dummy to estimate the

stock liquidity pairwise correlation regression.

676 The Journal of Financial Research

We further find that local liquidity commonality decreases over time.

Specifically, the average local liquidity beta estimates decrease from 0.305 during

1980–1992 to 0.232 during 2005–2016. The decreasing local liquidity beta may be the

result of growing market integration (Akbari, Ng, and Solnik forthcoming). The local

liquidity component tends to explain less than both the market liquidity and industry

liquidity components over time. We then examine the relation between local liquidity

commonality and firm size. Kamara, Lou, and Sadka (2008) document an increasing

market liquidity commonality divergence between small firms and large firms from

1963 to 2005. They show that greater institutional ownership is the reason large firms

comove more with the market. Presumably, large firms are less likely to be limited to

local resources (e.g., local order flows; Hasbrouck and Seppi 2001). We find that the

liquidity of small firms tends to comove more with the local liquidity portfolio than

that of large firms. The average local liquidity betas of the bottom and top size quintiles

are 0.404 and 0.124, respectively. Although the liquidity of small firms tends to

comove less with the market liquidity portfolio (Kamara, Lou, and Sadka 2008), it is

more subject to the local liquidity component.

The primary concern of our findings is that firms located nearby may have

similar unobservable characteristics and the comovement in these unobservable factors

drives local liquidity commonality. We employ a firm relocation sample to provide a

rigorous control that our results are not driven by the comovement in omitted variables.

The relocation sample allows us to examine whether there is any short‐term immediate

liquidity change covariance pattern when firms relocate, which is less likely to be

affected by the comovement in unobservable characteristics. We require a firm to have

at least three years and at most five years of data to be included in the relocation

sample, which consists of 172 firms. We find that firm liquidity exhibits significantly

positive sensitivity to the state liquidity portfolios of their contemporaneous locations.

In particular, after a firm relocates, its liquidity sensitivity to the old state liquidity

portfolio significantly reduces, whereas that to the new state liquidity portfolio

substantially increases. We also include interaction terms of the liquidity portfolios and

a Post dummy, which equals 1 if a firm year is after the relocation year, and 0

otherwise, in the regression model. We find consistent evidence that the commonality

of a firm’s liquidity with its old state liquidity portfolio decreases and that with its new

state liquidity portfolio increases after the relocation. A potential concern is that firm

headquarters relocation is not completely exogenous and may be a response to changes

within a firm, such as its intention to enter another industry. To address this concern,

we construct two additional industry liquidity portfolios of a firm based on its two

mostly related industries. Including changes in liquidity of the two additional industry

liquidity portfolios does not alter our results.

Moreover, a three‐year window may increase concerns of correlated omitted

variables. To resolve this issue, we restrict the sample to one year before and after the

relocation to examine the instant liquidity change covariance pattern of firm

headquarters relocation. The findings present consistent patterns that the liquidity

commonality of a firm with its old and new state liquidity portfolios changes

substantially after the relocation. Finally, we match the moving firms with nonmoving

firms in the old and new states separately. For both control samples, we show that after

677Commonality in Liquidity of Nearby Firms