10 1111@acfi 12656 PDF

Accounting & Finance
The effect of stock liquidity on investment efficiency

under financing constraints and asymmetric
information: Evidence from the United States
Heidi Quah, Janto Haman , Dharmendra Naidu

Department of Accounting, Monash Business School, Monash University, Caulfield East,
VIC, Australia
Abstract
We examine whether the association between stock liquidity and investment

efficiency is more pronounced for firms with more financial constraints and
information asymmetry problems. The results show that the effect of higher
stock liquidity on lowering under (over)-investment is more pronounced for
firms with more financial constraints and information asymmetry problems as
proxied by younger and higher business risk firms, respectively. We also find
similar results for firms with lower institutional ownership, more external
financing dependence and higher idiosyncratic risks. The findings collectively
suggest that the effect of stock liquidity in our setting is more pervasive for
firms with more financial constraints and information asymmetry problems.
Key words: Stock liquidity; Investment efficiency; Financial constraint;

Information asymmetry; Institutional ownership
JEL classification: M40, M41, M48, G12, G32, G38
doi: 10.1111/acfi.12656
1. Introduction
Firm investments are a major factor through which firms increase

shareholders’ wealth. However, over (under)-investment by firms is likely to
hamper this overriding objective. Over-investments can create excess capacity
while under-investments can cause firms to lose a business opportunity in
generating more income. Further, investment efficiency has major ramifications
for economic growth. Hence, it is important to understand factors that can
promote an optimal level of investments in firms.
Please address correspondence to Janto Haman via email: janto.haman@monash.edu
© 2020 Accounting and Finance Association of Australia and New Zealand

2 H. Quah et al./Accounting & Finance
This study examines the effect of one such factor, the liquidity of firm stocks,
under two conditions: financial constraints and information asymmetry. Extant
studies (Levine and Zervos, 1998; Beck and Levine, 2003; Caporale et al., 2004)
show that stock liquidity is positively associated with current and future rates
of capital accumulation, productivity and economic growth. Demirgüç-Kunt
and Levine (1996) suggest that stock liquidity increases incentives for investors
to acquire more information about firms. They state:
To the extent that larger, more liquid stock markets increase incentives to research
firms, the improved information will improve resource allocation and accelerate
economic growth’ (Demirgüç-Kunt and Levine, 1996, p. 230).
Investors are more willing to participate in capital raising of a firm that has high
liquid stock. In this context, it is easier for managers to obtain equity funding to
finance investments (Demirgüç-Kunt and Levine, 1996). Hence, due to lower
trading costs, higher stock liquidity can attract investors to become large
shareholders who are then incentivised to promote corporate governance via voice
threats through direct intervention (Maug, 1998) or exit threats through indirect
intervention by selling their shares (Admati and Pfleiderer, 2009; Edmans, 2009).
Both types of interventions can encourage managers to invest efficiently.
Another mechanism through which stock liquidity can promote efficient
investments is through improving firm transparency and making stock prices
more informative (Holmström and Tirole, 1993; Subrahmanyam and Titman,
2001; Easley and O’Hara, 2004; Khanna and Sonti, 2004). Transparency and
informative stock prices are likely to increase market scrutiny on firm
performance (Wurgler, 2000; Durnev et al., 2003), which can help improve
the quality and velocity of market feedback on managers’ investment decisions
(Subrahmanyam and Titman, 2001; Durnev et al., 2004). As such, higher stock
liquidity should reduce adverse selection and moral hazard problems, and thus
increase investment efficiency.
While we expect that higher stock liquidity improves investment efficiency,
we also expect this effect to be more pronounced for firms with more financial
constraints and information asymmetry problems. Our study examines how
stock liquidity affects these firms’ investment behaviours. We use young firms
to proxy for higher financial constraints and high business risk firms to proxy
for more information asymmetry problems.1 This is worthy of investigation for
the following reasons.
First, to the best of our knowledge, no study investigates the beneficial effect
of stock liquidity for young and high business risk firms. This is important
1
While the KZ index is a common measure for financial constraints, we use firm age to
proxy for financial constraints because Hadlock and Pierce (2010, p. 1909) caution
researchers ‘on the validity of the KZ index as a measure of financial constraints’. They
find that ‘firm size and age are particularly useful predictors of financial constraint
levels’. Therefore, we use firm age to measure financial constraints and we control for
firm size in our empirical models.

H. Quah et al./Accounting & Finance 3
because these firms play a strategic role in creating jobs and stimulating growth
and innovation (OECD, 2015)2. Also, Fazzari et al. (1988) show that financial
constraints in capital markets influence investments. They document that
investment–cash flow sensitivity increases in financing constraints measured as
firm age. Mature firms are unconstrained because they have enough cash
reserves or less information asymmetry problems (see Lang and Lundholm,
1993) due to their established reputation. Thus, mature firms appear to face
fewer obstacles in sourcing funds for their investments.
Second, younger firms generally have higher growth rate (Schreyer, 2000;
Arellano et al., 2012), but they also face higher rates of failure (Haltiwanger et
al., 2013; Collier et al., 2016). Such firms typically take much higher business
risk and have earnings uncertainty (Zhang, 2006; Fang et al., 2009). High
business risk firms also tend to exhibit greater opacity in their information
environment (Zhang, 2006; Fang et al., 2009). Hence, the presence of financial
constraints and information asymmetry is more pervasive for younger and
higher business risk firms, respectively. As such, we expect that stock liquidity
will play a more beneficial role in promoting investment efficiency of these firms
through facilitating financing, information transparency and market feedback
to managers.
We examine the relation between stock liquidity and investment efficiency
using a large sample of US listed firms from 1992 to 2014. We commence by
showing that higher stock liquidity is associated with lower investment
inefficiency in the form of lower under (over)-investment. Further, we examine
the effect of stock liquidity on investment efficiency for firms with more
financial constraints and information asymmetry problems, in particular, for
young firms and high business risk firms.
We find that the relationship between stock liquidity and investment
efficiency is more pronounced for younger and higher business risk firms,
consistent with the beneficial effects of higher stock liquidity being more
prominent in situations where a firm is more financially constrained or its
information environment is more opaque. These results are consistent with the
view that higher stock liquidity has a beneficial effect to increase the
information content of stock prices, which improves market feedback and
transparency so as to influence managers of younger and higher business risk
firms to make value-creation decisions (Wurgler, 2000; Subrahmanyam and
Titman, 2001; Durnev et al., 2003) such as efficient investment decisions.
We confirm our main findings using two-stage least-squares regression and
our findings are robust to a series of alternative measures of stock liquidity and
investment efficiency. In additional tests, we show that the relation between
stock liquidity and investment efficiency is more pronounced for firms with
2
https://www.oecd.org/sti/young-SME-growth-and-job-creation.pdf.

lower institutional shareholdings, higher external financing dependence and

higher idiosyncratic risk.3
We expect that our findings can provide insightful information to investors,
regulators and managers of firms in the following ways. First, we document
that the positive association between stock liquidity and investment efficiency is
stronger for firms with more financial constraints and information asymmetry
problems, in particular, for younger firms and higher business risk firms. Our
results suggest that higher stock liquidity plays an important role in increasing
transparency for these firms that results in more market feedback, thereby
inducing managers to invest efficiently. Second, our findings can inform
managers on the important role of stock liquidity in attracting investors to
invest in younger and higher business risk firms. Third, our findings can
provide thought-provoking investment insights to assist investors understand
the beneficial effects of higher stock liquidity, generally for firms with more
financial constraints and information asymmetry problems, and specifically for
younger and higher business risk firms. These insights can assist investors to
make better-informed investment decisions. Fourth, our findings can be useful
to stock exchange regulators by highlighting the benefits of stock liquidity for
younger and higher business risk firms. This can encourage policy makers and/
or regulators to establish a policy and regulatory framework that can enhance
stock liquidity of firms with more financial constraints and information
asymmetry problems and promote greater investment efficiency.
We next discuss the literature review and hypotheses development. Section 3
discusses the research method, data and sample selection. Section 4 presents the
summary statistics, empirical, robustness and sensitivity test results. Section 5
provides results from additional analyses and Section 6 concludes.
2. Literature review and hypotheses development
2.1. Literature review
Prior research suggests that investment inefficiency could arise from the
separation of ownership and control, which creates information asymmetry
problems between managers and shareholders (Jensen and Meckling, 1976;
Fama and Jensen, 1983). Managers have the incentives to distort investment
(e.g., Shleifer and Vishny, 1997) in order to maximise their utility (Williamson,
1985; Grossman and Hart, 1986) through perquisite consumption and empire
building (Jensen, 1986; Avery et al., 1998), career concerns (Fama, 1980;
Narayanan, 1985; Bebchuk and Stole, 1993; Holmström, 1999) and preference
for a quiet life (Bertrand and Mullainathan, 2003). In addition, managers may
be overconfident in taking on risky projects (Roll, 1986; Heaton, 2002;
3
We thank the anonymous reviewer for the suggestions.

Malmendier and Tate, 2008). As such, it is important to identify factors that

can mitigate inefficient investment problems in firms.
Extant studies have investigated the effect of stock liquidity on firm value,
accruals quality, innovation, corporate governance and stock price informa-
tiveness. Fang et al. (2009), Bharath et al. (2013) and Cheung et al. (2015) find
that an exogenous increase (decrease) in stock liquidity improves (reduces) firm
value. Chen et al. (2015) find that firms with higher stock liquidity are less likely
to undertake accrual-based and real earnings management.
Higher stock liquidity can enhance corporate governance via voice threats
(direct intervention) and exit threats (indirect intervention). Maug (1998)
argues that higher stock liquidity enhances monitoring and control. Without
causing significant price volatility, higher stock liquidity allows share accumu-
lation at a low transaction cost. These new large shareholders are then
incentivised to collect information and intervene directly (i.e., threat of voice)
with the objective of improving firm value and price efficiency. Appel et al.
(2016) find that even now passive institutional investors increasingly play an
important role in US stock ownership since they can influence firm decision
making through large voting blocs and enhance long-term performance.
Admati and Pfleiderer (2009) and Edmans (2009) show that higher stock
liquidity strengthens corporate control by allowing large shareholders (e.g.,
institutional investors) to exit with low transaction costs when they are
disappointed with underperforming firms. This form of indirect intervention
(i.e., threat of exit) creates ex-ante incentives for managers to invest efficiently
to preserve their wealth as such exits can depress stock prices.
Extant studies also document that higher stock liquidity increases the
informativeness of stock prices (Grossman and Stiglitz, 1980; Kyle, 1985;
Easley and O’Hara, 2004; Khanna and Sonti, 2004). Hence, higher stock
liquidity can also improve investment efficiency through facilitating market
feedback to firm managers via informative stock prices. Informative stock
prices convey meaningful signals to the financial market about the need to
intervene when managerial investment decisions are of poor quality. Curative
mechanisms, such as shareholders’ lawsuits, institutional investor pressure and
the market for corporate control, depend on stock prices (Durnev et al., 2004).
Khanna and Sonti (2004) argue that investors have incentives to influence
prices to induce firms to undertake certain investments. Hence, the movement
of stock prices can act as a feedback device by conveying investor sentiments to
managers, allowing managers to make more informed decisions. Higher stock
liquidity attracts more informed investors who trade more intensively based on
their information on managerial behaviour, thus making stock prices more
informative to firm managers and other stakeholders. This is expected to
increase the importance of the market feedback effect (Subrahmanyam and
Titman, 2001) on managers’ decisions. These mechanisms induce better
monitoring so that managers are more likely to invest efficiently (Durnev et
al., 2004). Hence, higher stock liquidity can increase transparency and the

feedback effect by making prices more informative for market participants to

take disciplinary actions.
The price of a firm’s stock affects and reflects managerial decisions.
Information-laden stock prices provide valuable feedback about the quality
of managerial decisions to managers (Durnev et al., 2004; Khanna and Sonti,
2004; Fang et al., 2009). For example, it is plausible that a lower price indicates
investors’ dissatisfaction with poor managerial investment decisions, whereas a
higher price correctly reflects investors’ expectations of managerial efficiency in
the use of corporate capital. As prices change in response to firm investment
decisions, managers seek to continuously glean information from market prices
and then take actions that affect the value of the security (Subrahmanyam and
Titman, 2001; Bond et al., 2010). To attract investors’ interest in firm stocks,
Subrahmanyam and Titman (2001) argue that the feedback from stock prices
can influence managerial incentives to increase firm value such as by investing
efficiently. This explains why managers pay a great deal of attention to stock
liquidity and the movements of stock prices to gauge how they should respond.
For example, the market expects that CEOs must maintain higher stock prices
to retain their positions and avoid shareholder activists’ scrutiny (Bond et al.,
2010). As such, it is widely believed that managerial investment decisions are
influenced by their firms’ stock prices and liquidity (Bond et al., 2010). This
argument is supported by numerous studies (Durnev et al., 2001, 2004; Luo,
2005; Chen et al., 2007; Kau et al., 2008; Bakke and Whited, 2010).
Given that corporate investment efficiency is sought by investors, the
feedback from informative stock prices that incorporate investors’ expectations
can influence managers (Subrahmanyam and Titman, 2001) to invest efficiently.
Thus, it is likely that higher stock liquidity enhances investment efficiency
through a feedback effect, as higher stock liquidity attracts informed investors
who increase the information content of the stock price. Xiong and Su (2014)
find that stock liquidity improves investment efficiency of Chinese listed firms.
Ha and Vinh (2017) assert that corporate managers in Vietnam use stock
liquidity as a feedback for firm investment decisions. Cheung et al. (2019) find
that higher stock liquidity of Chinese listed firms post-reform increases
institutional ownership and improves price efficiency thereby inducing man-
agers to invest optimally.
2.2. Hypotheses development
We specifically examine whether the association between stock liquidity and

investment efficiency is more pronounced for firms with more financial
constraints and information asymmetry problems. We use young firms to
proxy for higher financial constraints (Fazzari et al., 1988; Lang and
Lundholm, 1993) and high business risk firms to proxy for more information
asymmetry problems (Zhang, 2006; Fang et al., 2009). Therefore, we develop
the following hypotheses 1a, 1b, 2a and 2b.

2.2.1. Hypotheses 1a and 1b
Extant studies show that younger firms are more susceptible to financing
constraints due to asymmetric information problems (Rauh, 2006; Brown et al.,
2009; Fee et al., 2009; Hadlock and Pierce, 2010) and moral hazards (Sanders
and Boivie, 2004; Park and Patel, 2015) because they lack performance history
in the capital market. Their limited record of accomplishment makes it more
difficult for outside investors to assess their quality and potential. Extant
studies also show that younger firms use more external funds to finance capital
expenditures because those firms have not had time to accumulate their
retained earnings (Rajan and Zingales, 1998; Pang and Wu, 2009; Brown et al.,
2012).
However, younger firms usually have trouble in raising equity funding and/or
obtaining external financing due to the higher uncertainty about future
prospects and value of these firms (Stiglitz and Weiss, 1981). Hence, they are
exposed to higher funding costs and/or external financing costs due to the
lemon’s premium. For such firms, financing constraints can cause their
investments to deviate from the optimal level.
Higher stock liquidity is likely to be more beneficial to younger firms through
increasing transparency and market feedback to managers, which should relax
financing constraints in younger firms (Subrahmanyam and Titman, 2001). On
the other hand, we expect that the effect of stock liquidity on investment
efficiency is weaker for more mature firms, since these firms benefit from a
richer information environment due to their longer operational history. These
firms can finance their investments using internally generated funds or through
raising equity funds, because outside investors are able to value these firms’
potential more accurately. We therefore predict that the effect of higher stock
liquidity on investment efficiency is more pronounced for firms with more
financial constraints as proxied by younger firms. This leads to the development
of hypotheses 1a and 1b:
H1a: The effect of higher stock liquidity on lowering under-investment is more

pronounced for younger firms.
H1b: The effect of higher stock liquidity on lowering over-investment is more

pronounced for younger firms.
2.2.2. Hypotheses 2a and 2b
Firms with higher business risk are likely to experience uncertainty in their
future cash flows. Such firms tend to exhibit greater opacity in their
information environment (Zhang, 2006; Fang et al., 2009), which incentivises

managers to act opportunistically by investing suboptimally. Further, due to

the lack of information on managers’ decisions, investors may be reluctant to
invest in these firms for fear of being disadvantaged (Myers and Majluf, 1984).
This could make it difficult for firms to raise sufficient funds for financing
positive net present value (NPV) projects and thus result in lower investments.
On the other hand, information asymmetry problems can also prevent
investors from assessing firm value accurately, thus causing temporary
mispricing. If such firms are overvalued, this can lead to managers over-
investing using the excessive funds obtained from temporary stock mispricing
(Baker et al., 2003). Conversely, if those firms are undervalued, managers may
opt to forego investment with a positive NPV opportunity due to the inability
to raise enough external capital.
Given that stock prices of firms with liquid stock are more informative, the
information environment of firms with higher business risk is also likely to
improve if their stocks are more liquid. Specifically, the problems of moral
hazard and adverse selection in those firms can be mitigated by increased
transparency and feedback effects that are associated with high stock liquidity.
As such, it is possible that high stock liquidity can enhance investment
efficiency to a greater extent in high business risk firms through increasing
transparency.
Furthermore, information-laden stock prices can accelerate market feedback
to firm managers on their investment decisions if their firm’s stock liquidity is
greater, influencing them to take value-increasing investment decisions. The
feedback effects from firm stock prices and liquidity to managers become more
important when their firms’ information environment is opaque. Subrahma-
nyam and Titman (2001) argue that feedback is more valuable to firms with
high cash flow uncertainty with respect to existing projects. In a similar vein,
Fang et al. (2009) document that the feedback effect of stock liquidity on firm
performance is stronger for firms with higher operating income volatility, as
these firms exhibit greater uncertainty in their information environment. As
such, it is plausible that higher stock liquidity is particularly more beneficial to
firms inflicted with information opacity such as firms with high business risk
through accelerating feedback from stock prices to managers. Therefore, we
predict that the effect of higher stock liquidity on investment efficiency is
more pronounced for firms with more information asymmetry problems as
proxied by higher business risk firms. This leads to development of hypotheses
2a and 2b:
H2a: The effect of higher stock liquidity on lowering under-investment is more

pronounced for higher business risk firms.
H2b: The effect of higher stock liquidity on lowering over-investment is more

pronounced for higher business risk firms.

3. Sample selection and research method
3.1. Sample selection
Panel A of Table 1 outlines our sample selection process. We obtain data for
firms listed on the NYSE, AMEX and NASDAQ from Compustat between
1992 and 2014 (inclusive) for constructing investment levels and control
variables. Next, we exclude financial firms (SIC codes 6000–6999) due to their
unique nature of investments and financing.4 Finally, we require data for
constructing our test and control variables from four sources: (1) stock liquidity
and firm age data from CRSP; (2) analyst data from IBES; (3) institutional
investor data from Thomson Reuters 13F; and (4) investor protection rights
data from Andrew Metrick’s website.5 This further reduces our sample,
resulting in a final sample of 51,576 firm-year observations.
Panel B of Table 1 reports the industry distribution of our final sample based
on the Fama and French (1997) 48-industry classification. The business services
industry represents the largest industry, constituting about 13.58 percent of the
whole sample. The next highest represented industries are the electronic
equipment, pharmaceutical products, and retail industries, which each
represent between 6 and 8 percent of the sample. Panel C of Table 1 shows
that all years are generally evenly distributed in our sample, contributing
between 1,855 and 2,559 observations.
3.2. Research method
Before moving on to testing our hypotheses, we first examine whether higher

stock liquidity in the United States is associated with lower under (over)-
investment. Next, we test how the effect of stock liquidity on investment
efficiency is influenced by firm age and business risk through executing our
main analyses for (1) young and mature firms separately and; (2) high and low
business risk firms separately.
We adapt the methodology of Biddle et al. (2009) to investigate the
relationship between stock liquidity and investment efficiency using pooled
ordinary least squares (OLS) regressions. To reduce potential model
misspecification bias, we cluster the standard errors by firm and year to
account for serial- and cross-sectional correlation (Petersen, 2009). We also
include industry fixed effects using the Fama and French (1997) 48-industry
4
The composition and nature of investments significantly differ in financial firms relative
to non-financial firms. Financial firms usually do not invest heavily in capital and R&D
in view of the nature of their business. Further, high leverage is common in financial
firms but denotes financial distress for non-financial firms.
5
http://faculty.som.yale.edu/andrewmetrick/data.html

Table 1
Sample selection and distribution
Panel A: Sample selection Observations
Compustat data for years between 1992 and 2014 143,610

Less: Firms operating in Financial industries (SIC 6000–6999) (42,464)
101,146
Less: Missing data after merging with data from CRSP, IBES (49,570)
and Thomson Reuters 13F
Final firm-year data for years between 1992 and 2014 51,576
Panel B: Industry distribution of sample firms
Fama and French industry name Number of firms % of sample
Business Services 7,004 13.58

Electronic Equipment 4,168 8.08
Pharmaceutical Products 3,570 6.92
Retail 3,150 6.11
Medical Equipment 2,452 4.75
Machinery 2,405 4.66
Computers 2,337 4.53
Utilities 2,332 4.52
Petroleum & Natural Gas 2,315 4.49
Wholesale 1,984 3.85
Measuring & Control Equipment 1,588 3.08
Communication 1,364 2.64
Chemicals 1,351 2.62
Transportation 1,243 2.41
Construction Materials 1,172 2.27
15 Industries 38,435 74.51
24 Other Industries 13,141 25.49
Total Sample 51,576 100.00
Panel C: Year distribution of sample firms
Year Number of firms % of sample
1992 1,855 3.60

1993 2,198 4.26
1994 2,288 4.44
1995 2,387 4.63
1996 2,432 4.72
1997 2,426 4.70
1998 2,530 4.91
1999 2,438 4.73
2000 2,377 4.61
2001 2,559 4.96
2002 2,482 4.81
2003 2,389 4.63
2004 2,324 4.51
(continued)

Table 1 (continued)
Panel C: Year distribution of sample firms
Year Number of firms % of sample
2005 2,266 4.39

2006 2,178 4.22
2007 2,143 4.16
2008 2,151 4.17
2009 2,097 4.07
2010 2,055 3.98
2011 2,018 3.91
2012 1,981 3.84
2013 1,991 3.86
2014 2,011 3.90
Total sample 51,576 100.00
This table presents sample selection procedure (Panel A), industry distribution of sample
firms (Panel B) and year distribution of sample firms (Panel C).
classification to control for industry-specific shocks to investment. Our

investment efficiency regression model is specified as follows:
INVi,tþ1 ¼ α þ β1 LIQi,t þ β2 LIQi,t OverFirmi,t þ β3 OverFirmi,t

(1)
þ∑γj Controls j,i,t þ ɛi,tþ1
INV is the sum of capital expenditures, R&D expenditures and acquisitions,

minus sales of property, plant and equipment, scaled by lagged total assets
(Richardson, 2006). While this measure captures overall corporate investment
behaviour, we undertake additional tests to examine whether our findings vary
for investments in tangible- and non-tangible assets given the higher levels of
uncertainty associated with intangible assets (see Section 4.4.2).
We employ several measures of stock liquidity (LIQ). For our main analyses,
we use a composite measure of stock liquidity (LIQindex) to reduce
measurement errors. We compute LIQindex as the standardised average of
LIQFHT and LIQBAS. The higher (lower) values of LIQindex indicate higher
(lower) stock liquidity. LIQFHT measures the cost of trading as a percentage of
the share price based on Fong et al. (2017), while LIQBAS is the closing percent
quoted spread introduced by Chung and Zhang (2014).6 We also examine our
hypotheses using alternative measures of stock liquidity as sensitivity tests (see
Section 4.4.2).
6
These liquidity proxies are calculated from daily data and are highly correlated with
intraday-based liquidity benchmarks (Goyenko et al., 2009; Fong et al., 2017). The
measures of stock liquidity derived from high-frequency data include percent effective
spread, percent quoted spread, percent realised spread, and percent price impact.

In order to assess whether stock liquidity can alleviate under (over)-

investment problems, we follow Biddle et al. (2009) and construct a variable
(OverFirm) that partitions the sample firms into those that are likely to over
(under)-invest based on cash balance and leverage. To construct OverFirm, we
first rank firms into deciles based on their cash balance and negative leverage.
We then take the average of the two ranks for each observation and rescale
them so that the values for OverFirm range between zero and one, whereby
OverFirm is increasing in the likelihood of over-investment. We formally test
our hypotheses after interacting OverFirm with stock liquidity.
We examine whether the coefficient on stock liquidity (β1 in Equation (1))
increases with the investment levels in firms that are more likely to under-
invest. In addition, we concurrently examine whether the sum of the coefficients
on stock liquidity and its interactions with OverFirm (β1 + β2) decreases with
the investment levels in firms that are more likely to over-invest. In other
words, we expect the sum of these coefficients to be negative.
Our control variables in Equation (1) consist of firm-specific characteristics
and corporate governance variables. Consistent with Biddle et al. (2009), we
include all control variables and their related interaction with OverFirm.7 Prior
research shows that corporate governance has an impact on investment efficiency
(Biddle et al., 2009). As such, we control for anti-takeover protection index
(G_Score) (Gompers et al., 2003),8 analysts coverage (Analyst) and institutional
ownership (IO) (Bushee, 1998; Gillan and Starks, 2000; Hartzell and Starks,
2003; Velury and Jenkins, 2006; Chung and Zhang, 2011; Ramalingegowda and
Yu, 2012). Further, Biddle et al. (2009) find that firms with higher financial
reporting quality are more likely to be positively associated with investment
efficiency. Therefore, we include accruals quality (AQ) to proxy for financial
reporting quality, measured using the Francis et al. (2005) approach.
In addition, we control for firm age (Firm_Age) and dividend payout (DIV)
to capture investment–cash flow sensitivity as suggested by Fazzari et al.
(1988). Following Biddle and Hilary (2006), we also control for market-to-
book ratio (MB) and the ratio of operating cash flow to sales (CFOsale) to
proxy for investment opportunities as firms with more investment opportuni-
ties are likely to invest more. We also include bankruptcy risk (Z_Score) and
industry leverage (Ind_Kstruct) as proxies for financial pressure as both
measures are found to be negatively related to investment levels in Biddle et al.
(2009). Finally, we control for other innate firm characteristics that may affect
7
Consistent with Biddle et al. (2009) and for the brevity of our tables, we only include the
interaction term of main control variables with OverFirm. Nevertheless, we re-estimate
the regressions in Tables 3, 5 and 7 after including the interaction term of all control
variables with OverFirm and find qualitatively and quantitatively consistent results.
8
We multiply the Gompers et al. (2003) score by −1, so that the G_Score variable is
increasing in corporate governance. Following Biddle et al. (2009), we include an
indicator variable, G_Dummy, coded as 1 if G_Score is missing and zero otherwise.

investment levels (Verdi, 2006; Biddle et al., 2009). These controls include firm
size (LogAsset), cash flow volatility (σ(CFO)), sales volatility (σ(Sales)), log
operating cycle (OCycle), losses (Loss), investment volatility (σ(INV)), and
tangibility of firm assets (Tangibility). We provide the definition and measure-
ment of all variables in the Appendix. To minimise the effect of outliers, we
winsorise all continuous control variables at the top and bottom 1 percentile.
In order to examine whether the effect of stock liquidity on investment
efficiency is more pronounced for firms with more financial constraints and
information asymmetry problems, we split our main sample into quintiles
based on firm age (Firm_Age) and operating income volatility (IncVol),
respectively.
Hypotheses 1a and 1b predict that the effect of higher stock liquidity on
lowering under (over)-investment is more pronounced for young firms relative
to mature firms. We split our main sample into quintiles based on firm age
(Firm_Age) and classify young (mature) firms as those in the bottom (top)
quintile of Firm_Age.
Firm_Age is defined as the number of years since the firm first appears in
CRSP. We then estimate Equation (1) separately within the bottom and top
quintiles of Firm_Age. We estimate the regression models for young and
mature firms separately rather than using three-way interactions to allow the
parameter estimates on the control variables to also vary across the two sub-
samples of firms and to simplify the interpretation of the empirical results.
However, we also perform the robustness checks using three-way interactions
to confirm the results of sub-samples.
Furthermore, hypotheses 2a and 2b expect that the effect of higher stock
liquidity on lowering under (over)-investment is more pronounced for high
business risk firms relative to low business risk firms. We undertake a similar
approach as discussed above. In this context, we measure higher or lower level
of business risk using operating income volatility (IncVol). Specifically, we
define IncVol as the standard deviation of annual income before extraordinary
items divided by total assets for the 5-year window from t–4 to t (Bates, 2005;
Fang et al., 2009). We then split the sample firms into quintiles based on the
values of IncVol with the top (bottom) quintile representing firms with high
(low) business risk. To test H2a and H2b, we execute our analysis using
Equation (1) for the sub-samples of high and low business risk firms separately.
For the robustness checks, we use three-way interactions to confirm the results
of sub-samples.
4. Empirical results
4.1. Summary statistics
Table 2 presents the descriptive statistics for the variables employed in our
main analysis. On average, firms in our sample invest 15.32 percent of their

lagged total assets (median = 10.13 percent), which is similar to the investment
levels reported in prior studies (e.g., Biddle et al., 2009). The descriptive
statistics of stock liquidity imply that the average (median) cost of trading is
around 0.75 (0.35) percent of stock price and the daily closing bid-ask spread is
1.85 (0.98) percent greater than the quote midpoint. The mean (median)
OverFirm is 0.545 (0.556). The summary statistics of governance and other
control variables are similar to those reported in other studies (e.g., Biddle et
al., 2009; Cheng et al., 2013), which suggests that our sample is comparable to
those employed in prior studies.
4.2. Stock liquidity and investment efficiency
Table 3 reports the regression results of investment on stock liquidity

conditional on firm’s propensity to over (under)-invest using three measures of
stock liquidity (i.e., LIQFHT, LIQBAS and LIQindex). The coefficient on LIQ,
i.e., LIQFHT (LIQBAS), [LIQindex] is positive and statistically significant
(p < 0.01), suggesting that when the likelihood of under-investing is greater,
firms with higher stock liquidity have significantly higher levels of investments
than firms with lower stock liquidity. These results are also economically
significant. For instance, one standard deviation increase in LIQFHT
Table 2
Descriptive statistics
n Mean Std. dev. Min. Median Max.
INVt+1 (%) 51,576 15.323 16.714 −2.271 10.132 146.860

LIQFHT 51,576 −0.747 0.968 −5.322 −0.347 0.000
LIQBAS 51,576 −1.851 2.368 −15.325 −0.978 −0.016
LIQindex 51,576 0.038 0.786 −3.885 0.341 0.666
OverFirm 51,576 0.545 0.270 0.000 0.556 1.000
Total Assets in $m 51,576 2,826.817 11,745.218 0.207 311.633 349,493.000
logAsset 51,576 5.859 1.974 1.128 5.742 11.422
MB 51,576 2.002 1.563 0.447 1.496 19.611
σ(Sales) 51,576 0.192 0.190 0.008 0.133 1.466
σ(CFO) 51,576 0.076 0.080 0.005 0.052 0.807
σ(INV) 51,576 7.728 16.672 0.116 2.961 252.816
Z_Score 51,576 1.303 1.316 −7.723 1.374 4.996
Tangibility 51,576 0.279 0.233 0.003 0.204 0.907
Ind_Kstruct 51,576 0.170 0.110 0.029 0.139 0.588
CFOsale 51,576 −0.075 1.231 −27.239 0.081 0.698
DIV 51,576 0.399 0.490 0.000 0.000 1.000
Firm_Age 51,576 19.366 16.769 1.000 14.000 90.000
OCycle 51,576 4.642 0.741 1.936 4.715 6.636
Loss 51,576 0.265 0.441 0.000 0.000 1.000
AQ 51,576 −0.054 0.045 −0.299 −0.041 −0.003
(continued)

Table 2 (continued)
n Mean Std. dev. Min. Median Max.
IO 51,576 0.497 0.286 0.000 0.515 1.000

Analyst 51,576 6.461 7.214 0.000 4.000 55.000
G_Score 51,576 −1.231 3.281 −18.000 0.000 0.000
G_Dummy 51,576 0.866 0.340 0.000 1.000 1.000
This table presents descriptive statistics for the variables used in the analyses. Investment
(INV) is a measure of one-year-ahead total investment scaled by lagged total assets. The main
test variables include LIQFHT, LIQBAS and LIQindex. LIQFHT, formulated by Fong et al.
(2017), implies the cost of trading as a percentage of stock price. LIQBAS is a measure of the
closing percent quoted spread developed by Chung and Zhang (2014). Both LIQFHT and
LIQBAS are multiplied by minus 100 so that higher values indicate greater stock liquidity.
LIQindex is a continuous variable computed as the standardised average of LIQFHT and
LIQBAS. OverFirm is a ranked variable based on the average of a ranked (deciles) measure of
cash (scaled by total assets) and leverage (multiplied by minus one). LogAsset is the log of
total assets. MB is the ratio of the market value to the book value of total assets. σ(Sales),
σ(CFO) and σ(INV) is the standard deviation of sales, CFO and investment, respectively.
Z_Score is a measure of distress computed following the methodology in Altman (1968).
Tangibility is the ratio of PPE to total assets. Ind_Kstruc is the mean market leverage for firms
in the same SIC 3-digit industry. Market leverage is computed as the ratio of long-term debt
to the sum of long-term debt to the market value of equity. CFOsale is the ratio of CFO to
sales. DIV is an indicator variable that takes the value of one if the firm paid a dividend, and
zero otherwise. Firm_Age is the difference between the first year when the firm appears in
CRSP and the current year. OCycle is a measure of the log of operating cycle of the firm. Loss
is an indicator variable that takes the value of one if net income before extraordinary item is
negative, and zero otherwise. AQ is accruals quality, followingthe model of Francis et al.
(2005), defined as the standard deviation of the firm-level residuals during the years t−5 to
t−1 and multiplied by minus one. IO is the percentage of firm shares held by institutional
investors. Analyst is the number of analysts following the firm. G_Score is the measure of
investor protection rights created by Gompers et al. (2003), multiplied by minus one.
G_Dummy is an indicator variable that takes the value of one if G_Score is missing, and zero
otherwise.
(LIQBAS) [LIQindex] is associated with an increase of under-investment by 1.2

(1.2) [1.4] percent of total assets, which represents 8.1 (8.1) [9.2] percent
increment in investment levels in our sample.
Further, the sum of the coefficients (the joint test) on LIQFHT and LIQFHT
× OverFirm (LIQBAS and LIQBAS × OverFirm) [LIQindex and LIQindex ×
OverFirm] is negative and statistically significant {p < 0.01 (p < 0.05)
[p < 0.01]}. This result suggests that firms with higher stock liquidity are
associated with lower over-investment. In terms of the economic significance,
an increase in one standard deviation in LIQFHT (LIQBAS) [LIQindex]
suggests a decrease in over-investment by 1.87 (0.77) [1.45] percent of total
assets. This represents a decrease in investment of 12.2 (5.0) [9.5] percent
relative to the mean investment levels in our sample.

Table 3
Conditional relation between investment and stock liquidity (2-way clustering)
LIQFHT LIQBAS LIQindex

Variable Pred. Sign (1) (2) (3)
LIQ (1) Under-invest + 1.283*** 0.524*** 1.790***

(3.340) (4.312) (3.900)
LIQ × OverFirm (2) − −3.218*** −0.847*** −3.634***
(−5.663) (−4.265) (−5.391)
(1)+(2) Over-invest − −1.934*** −0.323** −1.844***
(−5.307) (−2.015) (−3.796)
Control variables
IO + 1.613* 1.922** 1.331
(1.402) (1.752) (1.189)
IO × OverFirm − 4.304** 2.426* 4.151**
(2.422) (1.355) (2.281)
Analyst + 0.127*** 0.143*** 0.130***
(2.594) (2.767) (2.604)
Analyst × OverFirm − 0.013 0.005 0.016
(0.137) (0.051) (0.164)
G_Score − −0.212*** −0.227*** −0.219***
(−2.337) (−2.432) (−2.392)
G_Score × OverFirm + 0.045 0.053 0.049
(0.355) (0.402) (0.379)
G_Dummy 1.850** 1.900** 1.873**
(2.274) (2.294) (2.274)
AQ + 16.688*** 17.986*** 17.321***
(2.678) (2.891) (2.783)
AQ × OverFirm − −24.162*** −26.526*** −25.427***
(−2.432) (−2.590) (−2.519)
OverFirm −3.202** −1.765 −0.817
(−2.067) (−1.193) (−0.577)
LogAsset − −1.137*** −1.285*** −1.201***
(−7.688) (−7.554) (−7.343)
MB + 2.365*** 2.334*** 2.357***
(10.918) (10.554) (10.668)
σ(Sales) −1.234 −1.178 −1.189
(−1.347) (−1.286) (−1.296)
σ(CFO) 5.216** 5.213** 5.291**
(2.471) (2.435) (2.496)
σ(INV) 0.051** 0.052** 0.052**
(1.976) (1.983) (1.982)
Z_Score − −1.718*** −1.701*** −1.715***
(−6.231) (−6.162) (−6.215)
Tangibility + 7.887*** 8.053*** 7.959***
(9.320) (9.350) (9.312)
Ind_Kstruct −9.267*** −8.691*** −8.988***
(−5.110) (−4.502) (−4.762)
(continued)

Table 3 (continued)
LIQFHT LIQBAS LIQindex

Variable Pred. Sign (1) (2) (3)
CFOsale −0.800*** −0.775*** −0.789***

(−2.924) (−2.857) (−2.896)
DIV − −1.209*** −1.234*** −1.237***
(−4.970) (−5.034) (−5.061)
Firm_Age − −0.051*** −0.052*** −0.052***
(−6.317) (−6.451) (−6.449)
OCycle −1.873*** −1.812*** −1.839***
(−8.106) (−7.772) (−7.920)
Loss − −3.350*** −3.299*** −3.309***
(−10.528) (−10.411) (−10.416)
Constant 28.440*** 28.822*** 27.801***
(9.695) (10.325) (9.930)
Industry FE Yes Yes Yes
Firm & Year clusters Yes Yes Yes
Adjusted R2 (%) 24.06 23.92 23.98
R2old (%) 23.88 23.88 23.88
R2new (%) 24.09 23.95 24.01
Gujarati (2003) ΔR2 F-statistic 71.34*** 23.74*** 44.12***
Sample size 51,576 51,576 51,576
This table presents the results of regression estimates for models that examine the relation
between stock liquidity and investment levels. LIQFHT, formulated by Fong et al. (2017),
implies the cost of trading as a percentage of stock price. LIQBAS is a measure of the closing
percent quoted spread developed by Chung and Zhang (2014). Both LIQFHT and LIQBAS
are multiplied by minus 100 so that higher values indicate greater stock liquidity. LIQindex is
a continuous variable computed as the standardised average of LIQFHT and LIQBAS. The
first row reports the coefficient estimate and the second row reports the two-way clustered t-
statistic based on standard error clustered by firm and year (in parentheses). Industry fixed
effects based on Fama and French’s classification scheme are included in all regressions but
the coefficients are not reported. The Gujarati (2003) ΔR2 F-statistics test the null hypothesis
that the inclusion of LIQ and LIQ × OverFirm as explanatory variables do not affect the
ðR2new R2old Þ=n
explanatory power of the regression analyses. The F-statistic is given by ð1R , where
new Þ=df
2
R2new (R2old) is the R2 value of the regression model with the inclusion (exclusion) of LIQ and
LIQ × OverFirm, n equals the number of new regressors being considered (i.e., n = two; the
new regressors include LIQ and LIQ × OverFirm), and df is the number of observations
minus the number of parameters in the regression model that includes LIQ and LIQ ×
OverFirm (Gujarati, 2003, pp. 260–264). ***, **, * denote statistical significance at the 1%,
5% and 10% levels, respectively, based on one-tailed tests if the coefficient has a predicted
sign, and two-tailed tests otherwise. See the Appendix for other variable definitions.
The control variables across the three specifications are generally significant
(p < 0.05 or better) and the signs on the coefficients along with the adjusted R2
values are similar to those reported in prior studies (e.g., Biddle et al., 2009).
The Gujarati (2003) F-statistics ranging between 23.74 and 71.34 across the

Table 4
Descriptive statistics and univariate tests for young and mature firms
Young firms (top quintile) (1) Mature firms (bottom quintile) (2)
N = 10,274 N = 10,274
t-test
Variable Mean Median SD Mean Median SD t
INVt+1 (%) 20.228 14.166 20.206 10.408 7.539 11.117 43.16***

LIQindex −0.120 0.202 0.887 0.368 0.527 0.452 −49.674***
OverFirm 0.618 0.667 0.284 0.429 0.389 0.220 53.269***
Total Assets 817.642 171.822 3,319.046 8,423.900 2,068.368 20,738.215 −36.709***
in $M
logAsset 5.223 5.146 1.575 7.420 7.635 1.994 −87.65***
MB 2.470 1.758 2.094 1.590 1.327 0.880 39.25***
σ(Sales) 0.285 0.202 0.261 0.132 0.092 0.129 53.402***
σ(CFO) 0.119 0.081 0.115 0.041 0.032 0.036 65.877***
σ(INV) 18.368 7.121 30.453 2.529 1.626 3.279 52.417***
Z_Score 0.965 1.112 1.539 1.508 1.448 0.875 −31.117***
Tangibility 0.236 0.148 0.226 0.352 0.291 0.235 −36.119***
Ind_Kstruct 0.150 0.112 0.101 0.216 0.191 0.123 −41.594***
CFOsale −0.280 0.059 1.682 0.098 0.098 0.486 −21.891***
DIV 0.151 0.000 0.799 1.000
Firm_Age 3.866 4.000 1.231 46.648 41.000 15.512 −278.688***
OCycle 4.585 4.643 0.821 4.649 4.717 0.581 −6.48***
Loss 0.397 0.000 0.121 0.000
AQ −0.060 −0.045 0.050 −0.038 −0.030 0.032 −37.38***
IO 0.443 0.420 0.284 0.579 0.630 0.239 −37.28***
Analyst 5.629 4.000 5.427 9.179 7.000 8.384 −36.032***
G_Score −0.482 0.000 1.993 −2.331 0.000 4.423 38.634***
G_Dummy 0.940 1.000 0.770 1.000
This table presents the descriptive statistics of variables used in regression analyses for sub-
samples of young (firms in the top quintile of Firm_Age) and mature firms (bottom quintile of
Firm_Age). Young firms are characterised as financially constrained whereas mature firms are
likely to be non-financially constrained. Univariate analysis results for comparing the means
of variables of these two sub-samples are reported. ***, **, * denote statistical significance at
the 1%, 5% and 10% levels, respectively. See the Appendix for other variable definitions.
three regression specifications are statistically significant (p < 0.01), suggesting

that LIQ and LIQ × OverFirm significantly increase the explanatory power of
the regression models.9
9
We re-estimate the three regressions without our test variables and find that the decline
in R2 after excluding the test variables is significant at the one percent level. The F-
statistic is computed using ðR2NEW R2OLD Þ=n=ð1 R2NEW Þ=df, where R2NEW (R2OLD) is the
R2 value of the regression model with (without) the test variable, n refers to the number
of new predictors being considered, and df refers to the difference between the number of
observations and the number of parameters in the regression model that includes the test
variables (Gujarati, 2003).

Table 5
Conditional relation between investment and stock liquidity: young and mature firms
Young firms Mature firms

Variables Pred. Sign (top quintile) (1) (bottom quintile) (2)
Test variables
LIQ (1) H1a: Under-invest + 3.547*** −0.084
(4.578) (−0.123)
LIQ × OverFirm (2) − −4.820*** −1.204
(−4.341) (−0.987)
(1)+(2) H1b: Over-invest − −1.274** −1.288
(−1.947) (−1.205)
Control variables
IO + −0.284 4.054**
(−0.127) (2.106)
IO × OverFirm − 5.497* 1.100
(1.553) (0.297)
Analyst + 0.285*** −0.007
(2.393) (−0.132)
Analyst × OverFirm − −0.176 0.254**
(−0.874) (2.193)
G_Score − −0.159 −0.262***
(−0.459) (−2.751)
G_Score × OverFirm + 0.144 0.129
(0.421) (1.046)
G_Dummy 1.911 2.101**
(0.694) (2.170)
AQ + 5.687 −0.662
(0.317) (−0.051)
AQ × OverFirm − −4.681 −3.285
(−0.196) (−0.138)
OverFirm −0.692 0.374
(−0.219) (0.154)
LogAsset − −2.448*** −0.471***
(−7.869) (−2.778)
MB + 2.235*** 2.112***
(8.842) (5.420)
σ(Sales) −4.625*** 0.409
(−4.633) (0.237)
σ(CFO) −0.876 11.263
(−0.352) (1.421)
σ(INV) 0.020 0.418***
(1.356) (5.497)
Z_Score − −1.405*** 0.259
(−2.975) (0.672)
Tangibility + 7.968*** 9.014***
(4.792) (6.313)
Ind_Kstruct −11.603*** −7.810***
(−3.180) (−4.233)
(continued)

Table 5 (continued)

CFOsale −0.663** −1.936**

(−2.478) (−2.392)
DIV − −1.185** −1.127***
(−1.850) (−2.383)
OCycle −2.106*** 0.161
(−5.357) (0.432)
Loss − −3.168*** −2.894***
(−5.183) (−6.681)
Constant 37.972*** 4.773
(7.234) (1.461)
Industry FE Yes Yes
Firm & Year clusters Yes Yes
Adjusted R2 (%) 0.246 0.165
Sample size 10,274 10,274
Columns (1) and (2) present the regression analysis results for young firms (firms in the top
quintile) and mature firms (bottom quintile). Firm_Age is measured as the number of years
since the firm first appears in Compustat with a stock price. The first row reports the
coefficient estimate and the second row reports the two-way clustered t-statistic based on
standard errors clustered by firm and year (in parentheses). Industry fixed effects based on
Fama and French’s classification scheme are included in all regressions, but the coefficients
are not reported. ***, **, * denote statistical significance at the 1%, 5% and 10% levels,
respectively, based on one-tailed tests if the coefficient has a predicted sign, and two-tailed
tests otherwise. See the Appendix for other variable definitions.
4.3. Main analyses and results
4.3.1. Results of testing hypotheses 1a and 1b
To test H1a and H1b on how the effect of stock liquidity on investment
efficiency is influenced by firm age, we execute our main analyses for young
firms and mature firms separately.
Table 4 presents the descriptive statistics and the differences between mean
values of stock liquidity proxy and other variables across young and mature
firms. The t-test results reveal statistical differences (p < 0.01) between the
mean values of stock liquidity proxy, firm growth (MB) and other firm
characteristics across the two groups.10 Overall, the mean (median) of firm age
is 4 years (4 years) for young firms and 47 years (41 years) for mature firms.
Specifically, younger firms exhibit lower stock liquidity (LIQindex), higher
growth (MB) and invest more (INV) than mature firms. These firms also have a
higher likelihood of over-investing (OverFirm).
10
We repeat the two-sample test using the non-parametric Mann-Whitney-Wilcoxon test
for medians and find that the results are consistent with the t-test results.

Table 6
Descriptive statistics and univariate tests for high and low business risk firms
High business risk firms (top Low business risk firms (bottom
quintile) (1) quintile) (2)
N = 10,198 N = 10,198
t-test
Variable Mean Median SD Mean Median SD t
INVt+1 (%) 22.371 16.585 21.141 11.483 8.220 11.862 −45.358***

LIQindex −0.202 0.179 1.007 0.255 0.455 0.565 39.965***
OverFirm 0.708 0.778 0.244 0.406 0.333 0.237 −89.721***
Total Assets 506.038 85.521 3,691.605 5,671.046 1,074.948 15,711.005 32.319***
in $M
logAsset 4.573 4.449 1.582 7.022 6.980 1.823 102.475***
MB 2.705 1.917 2.233 1.666 1.353 0.972 −43.074***
σ(Sales) 0.294 0.222 0.248 0.107 0.072 0.119 −68.745***
σ(CFO) 0.160 0.126 0.123 0.033 0.025 0.028 −102.224***
σ(INV) 19.046 8.274 30.124 2.737 1.645 4.654 −54.034***
Z_Score 0.232 0.522 1.792 1.613 1.512 0.950 68.782***
Tangibility 0.167 0.104 0.179 0.383 0.319 0.259 69.003***
Ind_Kstruct 0.114 0.083 0.080 0.231 0.212 0.124 80.048***
CFOsale −0.658 0.011 2.386 0.121 0.106 0.168 32.889***
DIV 0.106 0.000 0.693 1.000 0.461
Firm_Age 11.214 8.000 9.541 27.298 24.000 19.300 75.443***
OCycle 4.658 4.729 0.862 4.517 4.564 0.682 −13.000***
Loss 0.617 1.000 0.026 0.000
AQ −0.092 −0.075 0.062 −0.031 −0.025 0.025 92.751***
IO 0.386 0.332 0.285 0.553 0.581 0.257 43.828***
Analyst 4.100 2.000 5.109 8.598 6.000 7.707 49.134***
G_Score −0.496 0.000 2.028 −1.861 0.000 3.931 −31.16***
G_Dummy 0.939 1.000 0.803 1.000
IncVol 0.344 0.209 0.964 0.009 0.010 0.004 −35.076***
This table presents the descriptive statistics of variables used in regression analyses for sub-
samples of high business risk firms (top quintile of income volatility (IncVol)) and for low
business risk firms (bottom quintile of income volatility (IncVol)). IncVol is defined as the
standard deviation of quarterly operating income before depreciation divided by quarterly
book value of assets. It is measured over 20 quarters prior to the end of fiscal year t with a
minimum of eight quarterly observations. Unvariate analysis results for comparing the means
of variables of these two sub-samples are reported. ***, **, * denote statistical significance at
The results reported in column (1) of Table 5 indicate that the coefficient on
LIQindex is positive and statistically significant (p < 0.01) and the sum of the
coefficients on LIQindex and LIQindex × OverFirm (the joint test) is negative
and statistically significant (p < 0.05) for young firms. In contrast, the results
reported in column (2) of Table 5 indicate that these coefficients are statistically
insignificant for mature firms.

Additionally, we conduct tests to confirm that the main effect and joint effect of
stock liquidity on investment efficiency are significantly different from the two
regressions. We re-estimate Equation (1) by adding a dummy variable (Mature-
Firm) and its interactions with our test variable. That is, we estimate a three-way
interaction model where MatureFirm is coded 1 for mature firms and 0 for young
firms. Our untabulated results show a negative [positive] coefficient (−2.163)
[4.837] on the interaction between LIQindex and MatureFirm [LIQindex,
OverFirm and MatureFirm] that is statistically significant at the five [one] percent
level (t-stat = −2.040) [t-stat = 2.926]. These results suggest that the coefficients
for our test variables in column (1) are significantly different from those in column
(2) of Table 5. The results collectively support H1a and H1b that the effect of
stock liquidity on investment efficiency is more pronounced for young firms.
4.3.2. Results of testing hypotheses 2a and 2b
To test H2a and H2b on how the effect of stock liquidity on investment
efficiency is influenced by the level of business risk, we execute our main
analyses for high business risk firms and low business risk firms separately.
Table 6 presents the descriptive statistics and the differences between mean
values of stock liquidity and other variables across high and low business risk
firms. The t-test results reveal statistical differences (p < 0.01) between the
mean values of stock liquidity, firm growth and other firm characteristics across
the two groups.11 Overall, the mean (median) of IncVol is 0.344 (0.209) for high
business risk firms and 0.009 (0.010) for low business risk firms. Specifically,
higher business risk firms exhibit lower stock liquidity (LIQindex), higher
growth (MB) and invest more (INV) than lower business risk firms. These firms
tend to be younger (Firm_Age) and over-invest (OverFirm).
Column (1) of Table 7 presents the results from estimating Equation (1)
within the high business risk sub-sample. The results indicate that the
coefficient on LIQindex is positive and statistically significant (p < 0.01) while
the sum of coefficients on LIQindex and LIQindex × OverFirm (the joint test) is
negative and statistically significant (p < 0.05). In contrast, these coefficients
are insignificant, when we replicate this analysis using the sub-sample of low
business risk firms (reported in column (2) of Table 7).
Moreover, we conduct tests to confirm that the main effect and joint effect of
stock liquidity on investment efficiency are significantly different for high and
low business risk firms. We re-estimate Equation (1) by adding a dummy
variable (HighBusRisk) and its interactions with our test variable. That is, we
estimate a three-way interaction model where HighBusRisk is coded 1 for high
business risk firms and 0 for low business risk firms. Our untabulated results
11
We repeat the two-sample test using the non-parametric Mann-Whitney-Wilcoxon test
for medians and find that the results are consistent with the t-test results.

Table 7
Conditional relation between investment and stock liquidity: high and low business risk firms
High business risk Low business risk

Test variables
LIQ (1) H2a: Under-invest + 2.441*** −0.264
(3.430) (−0.407)
LIQ × OverFirm (2) − −3.591*** −0.177
(−3.395) (−0.155)
(1)+(2) H2b: Over-invest − −1.150** −0.442
(−1.660) (−0.538)
Control variables
IO + 6.221** 4.620***
(1.993) (2.963)
IO × OverFirm − −1.560 −2.785
(−0.352) (−0.788)
Analyst + 0.299** 0.089*
(1.646) (1.554)
Analyst × OverFirm − 0.001 0.073
(0.002) (0.581)
G_Score − 0.064 −0.032
(0.164) (−0.305)
G_Score × OverFirm + −0.113 −0.234
(−0.282) (−1.208)
G_Dummy 2.006 0.873
(0.710) (0.903)
AQ + 8.710 −5.726
(0.747) (−0.309)
AQ × OverFirm − −20.024 39.184
(−1.272) (1.275)
OverFirm 3.501 2.232
(1.283) (0.850)
LogAsset − −2.551*** −0.845***
(−7.010) (−4.434)
MB + 2.309*** 1.786***
(8.238) (6.181)
σ(Sales) −3.075** 4.767***
(−2.327) (2.767)
σ(CFO) −1.262 5.687
(−0.375) (0.672)
σ(INV) 0.027 0.137*
(1.479) (1.926)
Z_Score − −1.744*** −0.305
(−3.994) (−0.987)
Tangibility + 8.779*** 9.227***
(4.304) (6.953)
Ind_Kstruct −12.133*** −8.158***
(−2.889) (−3.764)
(continued)

Table 7 (continued)
High business risk Low business risk

CFOsale −0.644** −0.417

(−2.217) (−0.224)
DIV − −0.479 −2.150***
(−0.656) (−4.594)
Firm_Age − −0.103*** −0.031***
(−2.916) (−3.149)
OCycle −2.458*** −0.100
(−7.067) (−0.242)
Loss − −3.003*** −3.201***
(−4.272) (−2.854)
Constant 25.846*** 15.821***
(3.837) (4.940)
Industry FE Yes Yes
Adjusted R2 (%) 0.309 0.134
Columns (1) and (2) present the regression analysis results for high business risk firms (firms
in the top quintile of income volatility) and low business risk firms (bottom quintile). Income
volatility (IncVol) is defined as the standard deviation of quarterly operating income before
depreciation divided by quarterly book value of assets. It is measured over 20 quarters prior
to the end of fiscal year t with a minimum of eight quarterly observations. The first row
reports the coefficient estimate and the second row reports the two-way clustered t-statistic
based on standard error clustered by firm and year (in parentheses). Industry fixed effects
based on Fama and French’s classification scheme are included in all regressions, but the
coefficients are not reported. ***, **, * denote statistical significance at the 1%, 5% and 10%
levels, respectively, based on one-tailed tests if the coefficient has a predicted sign, and two-
tailed tests otherwise. See the Appendix for other variable definitions.
show a positive [negative] coefficient (2.012) [−3.686] on the interaction between

LIQindex and HighBusRisk [LIQindex, OverFirm and HighBusRisk] that is
statistically significant at the five percent level (t-stat = 2.286) [t-stat = −2.066].
These results suggest that the coefficients for our test variables in column (1) are
significantly different from those in column (2) of Table 7. The results support
H2a and H2b that the effect of stock liquidity on investment efficiency is more
pronounced for high business risk firms.
4.4. Robustness and sensitivity tests
4.4.1. Two-stage least-squares estimation
We supplement our analyses on the association between stock liquidity and

investment efficiency for the sub-samples of young, mature, high business risk
and low business risk firms by employing two-stage instrumental variables (IV)

Table 8
Endogeneity and reverse causality: two-stage least-squares (2SLS) regression for young and mature
firms
2nd Stage 2nd Stage

2nd Stage 2nd Stage high business low business
1st Stage young firms mature firms risk firms risk firms
Variable LIQindex (1) INV (2) INV (3) INV (4) INV (5)
TwoLIQ 0.474***
(7.977)
PrLIQ (1) 2.268* −1.830 3.638*** −1.862
Under-invest (1.965) (−1.264) (3.363) (−1.387)
PrLIQ × −4.493*** −0.688 −7.488*** 0.211
OverFirm (2) (−3.316) (−0.345) (−4.498) (0.101)
(1)+(2) Over-invest −2.226** −2.518 −3.850*** −1.651
(−2.778) (−1.593) (−3.199) (−1.414)
Control variables
IO 0.775*** 1.983 5.607** −0.085 7.002***
(10.066) (0.979) (2.344) (−0.026) (3.688)
IO × OverFirm 3.570 −0.971 8.132* −2.464
(1.300) (−0.219) (1.784) (−0.599)
Analyst −0.015*** 0.246** −0.062 0.377* 0.039
(−8.846) (2.602) (−1.144) (2.067) (0.764)
Analyst × OverFirm −0.166 0.310** −0.120 0.147
(−0.945) (2.361) (−0.468) (1.165)
G_Score 0.006* 0.456* −0.273*** 0.030 −0.059
(1.903) (1.845) (−3.358) (0.076) (−0.927)
G_Score × OverFirm 0.174 0.083 −0.053 −0.122
(0.628) (1.130) (−0.200) (−0.859)
G_Dummy −0.043 −3.557* 2.499** 1.306 0.763*
(−1.023) (−1.739) (2.618) (0.423) (1.738)
AQ 0.732*** 9.996 1.051 12.045 −22.469
(4.393) (0.877) (0.089) (1.122) (−1.095)
AQ × OverFirm −14.298 −12.760 −31.634** 50.992*
(−0.958) (−0.477) (−2.344) (1.766)
OverFirm 0.401*** 1.709 1.860 −1.063 3.781
(6.897) (0.920) (0.659) (−0.392) (1.446)
LogAsset 0.164*** −1.560*** −0.114 −1.756*** −0.522**
(6.565) (−6.803) (−0.420) (−4.797) (−2.540)
MB 0.060*** 1.882*** 2.610*** 2.397*** 1.801***
(7.624) (7.976) (5.422) (8.042) (5.415)
σ(Sales) −0.013 −4.320*** −0.531 −4.252*** 4.217**
(−0.357) (−5.739) (−0.259) (−4.093) (2.484)
σ(CFO) −0.129 1.313 5.028 −0.358 5.149
(−0.763) (0.790) (0.547) (−0.097) (0.508)
σ(INV) 0.001* 0.021 0.682*** 0.028 0.140*
(1.905) (1.472) (6.091) (1.390) (2.070)
Z_Score 0.007 −1.202*** 0.167 −1.672*** −0.261
(1.023) (−2.875) (0.353) (−3.779) (−0.873)
(continued)

Table 8 (continued)
2nd Stage 2nd Stage

2nd Stage 2nd Stage high business low business
1st Stage young firms mature firms risk firms risk firms
Variable LIQindex (1) INV (2) INV (3) INV (4) INV (5)
Tangibility 0.043 9.392*** 7.562*** 7.925*** 10.277***

(1.020) (6.833) (4.740) (3.919) (7.667)
Ind_Kstruct −0.372*** −13.120*** −8.620*** −12.977*** −8.929***
(−2.818) (−4.107) (−4.063) (−3.568) (−3.674)
CFOsale −0.024*** −0.713*** −1.666 −0.807*** −0.898
(−3.958) (−3.437) (−1.551) (−3.242) (−0.523)
DIV 0.037** −2.033*** −1.258** −1.344** −1.863***
(2.348) (−4.082) (−2.545) (−2.216) (−3.704)
Firm_Age −0.002*** −0.088*** −0.029**
(−3.711) (−3.844) (−2.667)
OCycle −0.028*** −1.614*** 0.029 −2.355*** −0.188
(−2.869) (−5.787) (0.083) (−5.701) (−0.524)
Loss −0.105*** −2.761*** −2.881*** −3.513*** −3.357***
(−5.482) (−5.041) (−5.950) (−5.742) (−3.236)
Constant −1.320*** 27.617*** −3.552 31.445*** 5.567
(−5.619) (7.017) (−0.814) (4.828) (1.670)
Industry FE Yes Yes Yes Yes Yes
Firm & Year Yes Yes Yes Yes Yes
clusters
Adjusted R2 (%) 62.8 29.6 15.43 30.47 13.11
Sample size 38,213 10,063 9,380 9,494 9,494
F-statistics F(1, 38,
(1st stage 2SLS) 154) = 63.63,
p = 0.0000
This table presents the two-stage least-squares (2SLS) estimates of the effect of stock liquidity
on firm investment efficiency. In the first-stage regression analysis, stock liquidity (LIQindex
as proxy for LIQ) is estimated using instrument variable, which measures the mean stock
liquidity of the two firms in firm i’s industry that have the closest size (market value of equity)
to firm i (TwoLIQ). In the second-stage regression, we use the predicted values of stock
liquidity (PrLIQ) and its interaction with Overfirm (PrLIQ × OverFirm) from the first-stage
regressions to estimate their effect on the dependent variable, one-year-ahead investment
levels (INV). The first row reports the coefficient estimate and the second row reports the two-
way clustered t-statistic based on standard errors clustered by firm and year (in parentheses).
Industry fixed effects based on Fama and French’s classification scheme are included in all
regressions, but the coefficients are not reported. ***, **, * denote statistical significance at
regressions. This approach attempts to address the possibility of our outcome

and explanatory variables being simultaneously determined and our results
affected by correlated omitted variables (Larcker and Rusticus, 2010). We
follow Fang et al. (2009) and use the mean stock liquidity of two firms in firm

Table 9
Conditional relation between investment and stock liquidity: alternative stock liquidity measures
Variable LIQFHT (1) LIQBAS (2) LIQAM (3) LIQHL (4) LIQindex4 (5)
Panel A: Results from sub-sample of young firms
LIQ (1) Under-invest 2.556*** 1.060*** 1.004** 3.541*** 2.063***

(3.618) (3.841) (2.196) (4.553) (3.360)
LIQ × OverFirm (2) −4.044*** −1.731*** −1.489** −2.439** −5.027***
(−3.994) (−3.282) (−2.734) (−2.127) (−2.927)
(1)+(2) Over-invest −1.488** −0.671* −0.485 1.102 −2.964**
(−2.747) (−1.862) (−1.225) (1.496) (−2.136)
Controls Yes Yes Yes Yes Yes
Firm & Year clusters Yes Yes Yes Yes Yes
Adjusted R2 (%) 24.63 25.26 26.52 24.03 27.98
Sample size 10,274 10,274 7,872 7,872 7,872
Panel B: Results from sub-sample of mature firms
LIQ (1) Under-invest 0.052 −0.182 0.223 1.055 0.192

(0.085) (−0.677) (0.844) (1.092) (0.150)
LIQ × OverFirm (2) −1.543 −0.088 −0.110 −0.816 −1.090
(−1.443) (−0.257) (−0.238) (−0.465) (−0.567)
(1)+(2) Over-invest −1.491* −0.270 0.113 0.238 −0.898
(−2.026) (−1.018) (0.303) (0.203) (−0.662)
Adjusted R2 (%) 16.58 16.59 15.40 15.44 24.63
Sample size 10,274 10,274 9,631 9,631 7,872
Panel C: Results from sub-sample of high business risk firms
LIQ (1) Under-invest 2.028*** 0.996*** 1.341** 3.975*** 2.962***

(3.414) (4.511) (2.675) (3.506) (3.075)
LIQ × OverFirm (2) −3.672*** −1.787*** −1.677** −4.095** −5.225***
(−3.888) (−4.041) (−2.386) (−2.241) (−3.809)
(1)+(2) Over-invest −1.644** −0.792** −0.336 −0.120 −2.263**
(−2.790) (−2.223) (−0.678) (−0.108) (−2.241)
Adjusted R2 (%) 31.02 31.74 28.46 29.08 29.10
Sample size 10,198 10,198 5,632 5,632 5,632
(continued)

Table 9 (continued)
Variable LIQFHT (1) LIQBAS (2) LIQAM (3) LIQHL (4) LIQindex4 (5)
Panel D: Results from sub-sample of low business risk firms
LIQ (1) Under-invest −0.745 0.027 0.248 1.028 0.032

(−1.226) (0.146) (0.846) (1.589) (0.078)
LIQ × OverFirm (2) 0.327 −0.211 −0.727** −0.422 −0.634
(0.347) (−0.603) (−2.337) (−0.766) (−0.879)
(1)+(2) Over-invest −0.418 −0.184 −0.479 0.606 −0.603
(−0.749) (−772) (−1.451) (1.204) (−1.136)
Adjusted R2 (%) 13.58 13.53 13.68 13.70 24.63
Sample size 10,198 10,198 9,862 9,862 10,274
This table presents regression analysis results with additional proxies of stock liquidity as test
variables. LIQAM is the price impact measure introduced by Amihud (2002). LIQHL is the
price-based spread proxy employed by Corwin and Schultz (2012). LIQindex4 is a composite
liquidity index computed by standardising the individual measures of stock liquidity
(LIQFHT, LIQBAS, LIQAM, LIQHL) and then calculating the unweighted mean of these
four standardised measures. The first row reports the coefficient estimate and the second row
reports the two-way clustered t-statistic based on standard errors clustered by firm and year
(in parentheses). Constant term, control variables and industry fixed effect based on Fama
and French’s classification scheme are included in all regressions, but the coefficients are not
reported. ***, **, * denote statistical significance at the 1%, 5% and 10% levels, respectively.
See the Appendix for other variable definitions.
i’s industry that are the closest in size (i.e., market value of equity) to firm i
(TwoLIQ) as instruments for stock liquidity (for brevity, we use LIQindex as
proxy for LIQ). Fang et al. (2009) propose that the component of firm i’s stock
liquidity that is correlated with its competitors’ stock liquidity is less likely to be
correlated with unobservable variables that affect firm i’s investment levels.12
The first-stage results from regressing the endogenous variable (LIQ) on
exogenous variable (TwoLIQ) and the control variables employed in Equa-
tion (1) are presented in column (1) of Table 8. The coefficient on TwoLIQ is
positive and highly significant (p < 0.01), suggesting that the instrument is
highly correlated with stock liquidity (LIQ). The partial F-statistic from the
first stage is 63.63, which is far greater than the critical F-value of 8.96
12
Lennox et al. (2012) emphasise the importance of identifying an exogenous variable
that affects the test variable (i.e., stock liquidity) but not the dependent variable (i.e.,
investment efficiency). A limitation of the two-stage least squares is that it is difficult to
identify exogenous variables. In our setting, we follow Fang et al. (2009) and use the
average liquidity of two firms in firm i’s industry that are closest in size to firm i as the
instrumental variable.

Table 10
Conditional relation between investment and stock liquidity: alternative investment measures
Variable Capex (1) Non_Capex (2) Capex (3) Non_Capex (4)
Panel A: Results from sub-samples of young and mature firms
LIQ (1) Under-invest 4.078** 1.718** −1.779 0.400

(2.209) (2.600) (−1.187) (0.503)
LIQ × OverFirm (2) −8.653*** −2.926*** −1.797 −0.585
(−3.221) (−2.975) (−0.455) (−0.497)
(1)+(2) Over-invest −4.575*** −1.208** −3.575 −0.185
(−2.909) (−2.160) (−1.146) (−0.265)
Controls Yes Yes Yes Yes
Industry FE Yes Yes Yes Yes
Firm & Year clusters Yes Yes Yes Yes
Adjusted R2 (%) 2.26 2.13 24.90 14.18
Sample size 10,784 10,564 10,538 10,538
High business risk firms Low business risk firms
Variable Capex (1) Non_Capex (2) Capex (3) Non_Capex (4)
Panel B: Results from sub-samples of high and low business risk firms
LIQ (1) Under-invest 3.583** 1.242** −2.611 0.682

(2.669) (2.695) (−1.189) (1.093)
LIQ × OverFirm (2) −7.424*** −2.442*** −1.279 −1.668
(−3.202) (−3.284) (−0.337) (−1.052)
(1)+(2) Over-invest −3.841** −1.200** −3.890 −0.986
(−2.668) (−2.375) (−1.586) (−0.818)
Controls Yes Yes Yes Yes
Industry FE Yes Yes Yes Yes
Firm & Year clusters Yes Yes Yes Yes
Adjusted R2 (%) 5.37 3.88 12.74 8.52
Sample size 10,413 10,623 10,407 10,623
This table presents regression analysis results with additional proxies of investment, Capex
and Non_Capex, as dependent variables. Capex is a measure of capital expenditure scaled by
lagged total assets. Non_Capex is calculated as the sum of R&D expenditures and
acquisitions scaled by lagged total assets. The first row reports the coefficient estimate and
the second row reports the two-way clustered t-statistic based on standard errors clustered by
firm and year (in parentheses). Constant term, control variables and industry fixed effect
based on Fama and French’s classification scheme are included in all regressions, but the
coefficients are not reported. ***, **, * denote statistical significance at the 1%, 5% and 10%
levels, respectively. See the Appendix for other variable definitions.
suggested by Stock et al. (2002). This result suggests that TwoLIQ is not a weak
instrument, and hence a Wald test is not required to test for weak instrument
(e.g., Rego and Wilson, 2012).

The results from the second-stage regressions are reported in columns (2)–(5)
of Table 8. The results from the second-stage regression for young (column 2)
and high business risk (column 4) firms show a positive coefficient on the fitted
values of stock liquidity (PrLIQ) (p < 0.10 or better). The joint significance test
(PrLIQ + PrLIQ × OverFirm) show a negative and significant (p < 0.05)
coefficient. Overall, the relationship between stock liquidity and investment
efficiency for young and high business risk firms is robust to controlling for
endogeneity using an instrumental variable approach.
Conversely, the results from the second-stage regression for mature (column
3) and low business risk (column 5) firms show an insignificant coefficient on
the fitted values of stock liquidity (PrLIQ) and the joint coefficient (PrLIQ +
PrLIQ × OverFirm), consistent with our main results.
4.4.2. Other sensitivity tests
We conduct other tests to evaluate the sensitivity of our results. First, we

explore whether our main results are driven by a particular year by repeating
our main regression after excluding one fiscal year at a time. We use a similar
approach of excluding one industry at a time to assess whether our results are
driven by a single industry. The results from these regressions continue to
return similar results, suggesting that they are not driven by any particular year
or industry.
We also assess the sensitivity of our results to five alternative measures of
stock liquidity. Our first two measures are LIQFHT and LIQBAS that are
defined in Section 3.2 and also in the Appendix . Our third measure of stock
liquidity (LIQAM) is based on Amihud (2002) and captures the lack of stock
liquidity, defined as the yearly average of daily absolute return divided by
dollar trading volume.13 We employ the natural logarithm of the Amihud
measure as it is highly positively skewed (Edmans et al., 2013). The fourth
additional proxy of stock liquidity (LIQHL) is the bid-ask spread measure
developed by Corwin and Schultz (2012). This measure is computed from daily
high and low prices based on the notion that daily high (low) prices are almost
always buy (sell) trades. We obtain Corwin and Schultz’s (2012) high-low
spread from Professor Corwin’s website (http://www3.nd.edu/~scorwin/). We
multiply both of these two additional proxies of stock liquidity by −1 so that
greater values of LIQAM and LIQHL reflect higher stock liquidity. Our fifth
proxy for stock liquidity considers the collective effects of the four individual
liquidity proxies. Specifically, we develop a new composite stock liquidity index
(LIQindex4) by standardising these four individual measures of stock liquidity
(LIQAM, LIQHL, LIQFHT and LIQBAS) and then calculating the
13
Stock selected in our sample must be listed at the end of its fiscal year t; have at least
200 days of return and volume data available; and have a minimum price of $5 at the
end of fiscal year t (Amihud, 2002; Fang et al., 2009).

unweighted mean. Again, higher values of LIQindex4 signify higher stock

liquidity.
The results based on the five additional proxies of stock liquidity, reported in
columns (1)–(5) of Table 9, show that the coefficients on all alternative stock
liquidity measures are positive and significant (p < 0.01) for young (Panel A)
and high business risk (Panel C) firms. The results also reveal that the sum of
the coefficients on additional stock liquidity proxies and their interactions with
OverFirm are negative and statistically significant (p < 0.10 or better) for three
out of five measures for young and high business risk firms. These results are
also consistent with our main results and suggest that stock liquidity promotes
investment efficiency by alleviating under (over)-investment problems for
young and high business risk firms. On the other hand, we do not find
significant coefficients on the main and joint effect for mature (Panel B) and low
business risk (Panel D) firms.
Moreover, prior studies (e.g., Biddle et al., 2009; Cheng et al., 2013) consider
two distinct types of investments, namely investments in intangible (Non_-
Capex) and tangible (Capex) assets. Stein (1988) argues that market partici-
pants find it more challenging to appraise firm investments in intangible assets
due to the return uncertainty of these assets. Given that intangible assets are
associated with higher agency costs arising from higher information asymme-
try, the need for monitoring becomes increasingly important in the presence of
more intangible assets (Gompers, 1995). As such, it is possible that stock
liquidity may play a particularly important role in monitoring and improving
the efficiency of intangible investments (Non_Capex) in young and high
business risk firms. We formally investigate this by repeating our main tests in
Tables 5 and 7 using dependent variables representing these two separate
components of investment (Capex and Non_Capex).14
Interestingly, the results from this analysis, reported in Table 10, show that
the relationship between stock liquidity and investment efficiency is applicable
to both tangible (Capex) and intangible investments (Non_Capex). Consistent
with our main results, we find a positive coefficient on the main effect and
negative coefficient on the joint effect that are statistically significant (p < 0.05
or better), for young (Panel A) and high business risk (Panel B) firms. We do
not find significant coefficients on the test variables for mature and low business
risk firms. Our findings suggest that higher stock liquidity improves the
efficiency of not only investments that are more difficult to monitor but also
plays an incremental role in improving the efficiency of tangible investments
that are less afflicted by informational opacity.
14
Following Baker et al. (2003) and Lang et al. (2012), Capex is computed as the capital
expenditure in year t + 1 scaled by lagged total assets, whereas Non_Capex is calculated
as the sum of R&D expenditures and acquisitions in year t + 1 scaled by lagged total
assets.

5. Additional analyses
In our additional analyses, we consider additional conditions or alternative

proxies for financial constraints or information asymmetry problems where the
relation between stock liquidity and investment efficiency is likely to be more
pronounced.
We first test the relation conditional on a firm’s institutional shareholding.15
One mechanism through which stock liquidity enhances investment efficiency is
better equity financing. More equity financing indicates more institutional
ownership compared to debt financing. It is possible that institutional
ownership would strengthen institutional monitoring (Jiang et al., 2014; Appel
et al., 2016) and corporate governance. On the contrary, firms with lower
institutional ownership have weak governance. For example, Chung and Zhang
(2011) show that institutional ownership increases with the quality of a firm’s
governance structure. Agency problems are likely to be higher for firms with
lower institutional ownership and/or weak governance. Hence, managers of
such firms are likely to extract rents by under (over)-investing. We conjecture
that higher stock liquidity is more beneficial for firms with low institutional
shareholdings because it is likely to increase transparency and market feedback
to managers. Therefore, we test the conditional relation between stock liquidity
and investment efficiency for low and high institutional shareholding firms and
report the results in Panel A of Table 11.
Institutional shareholding is measured as the percentage of shares held by
institutional investors. In Panel A, firms with low (high) institutional
shareholding represent firms in the bottom (top) quintile of institutional
shareholding. Our results support our conjecture and show a positive
coefficient on stock liquidity and a negative coefficient on the sum of stock
liquidity and its interaction with OverFirm, that are both statistically significant
(p < 0.01), for firms with low institutional shareholding (column 1). On the
contrary, we find insignificant coefficients on the main and joint effect for firms
with high institutional shareholding (column 2).
Next, we test the relation between stock liquidity and investment efficiency
conditional on an alternative proxy of financial constraints – external financing
dependence. We follow Ben-Nasr and Alshwer (2016) and measure external
financing dependence as the industry-median value of the difference between
capital expenditure and cash flow from operations scaled by capital expendi-
ture. High (low) financial constraints represent firms in the top (bottom)
quintile of external financing dependence. We report the results from this test in
Panel B of Table 11. The results show a positive coefficient on stock liquidity
and a negative coefficient on the sum of stock liquidity and its interaction with
OverFirm, which are both statistically significant (p < 0.05 or better), for firms
with high financial constraints (column 1). On the contrary, we find statistically
15
We thank the anonymous reviewer for the suggestion.

Table 11
Conditional relation between investment and stock liquidity: alternative sub-sample analysis
Low institutional shareholding High institutional shareholding

Variable INV (1) INV (2)
Panel A: Results from sub-samples of low and high institutional shareholding
LIQ (1) Under-invest 1.440*** 0.222

(2.981) (0.069)
LIQ × OverFirm (2) −3.284*** −1.808
(−4.155) (−0.411)
(1)+(2) Over-invest −1.844*** −1.586
(−3.299) (−0.496)
Controls Yes Yes
Industry FE Yes Yes
Adjusted R2 (%) 26.91 21.15
High financial constraints Low financial constraints

Panel B: Results from sub-samples of high and low financial constraints
LIQ (1) Under-invest 2.397*** 1.634

(3.206) (1.328)
LIQ × OverFirm (2) −4.158*** −0.379
(−3.734) (−0.478)
(1)+(2) Over-invest −1.761** 1.255
(−2.272) (1.411)
Controls Yes Yes
Industry FE Yes Yes
Adjusted R2 (%) 34.95 17.15
High idiosyncratic risk Low idiosyncratic risk
Panel C: Results from sub-samples of high and low idiosyncratic risk
LIQ (1) Under-invest 1.180** −1.291

(2.227) (−0.904)
LIQ × OverFirm (2) −2.734*** 1.588
(−2.906) (0.575)
(1)+(2) Over-invest −1.554** 0.297
(continued)

Table 11 (continued)
High idiosyncratic risk Low idiosyncratic risk

(−2.661) (0.148)
Controls Yes Yes
Industry FE Yes Yes
Adjusted R2 (%) 33.25 13.26
This table presents regression analysis results using alternative sub-samples to proxy for
conditions where high and low information asymmetry and/or equity financing are
significantly different. Institutional shareholding (IO) is the percentage of shares held by
institutional investors. Low (high) institutional shareholding represents firms in the bottom
(top) quintile of institutional shareholding. We use external financing dependence (EF) to
proxy for financial constraints. EF is measured as the industry-median value of the difference
between capital expenditure and cash flow from operations scaled by capital expenditure.
High (low) financial constraints represent firms in the top (bottom) quintile of EF.
Idiosyncratic risk (IdioRisk) is the standard deviation of the residuals from the firm-specific
market model regression. High (low) idiosyncratic risk represents firms in the top (bottom)
quintile of idiosyncratic risk. The first row reports the coefficient estimate and the second row
reports the two-way clustered t-statistic based on standard errors clustered by firm and year
(in parentheses). Constant term, control variables and industry fixed effect based on Fama
and French’s classification scheme are included in all regressions, but the coefficients are not
reported. ***, **, * denote statistical significance at the 1%, 5% and 10% levels, respectively.
See the Appendix for other variable definitions.
insignificant coefficients on the main and joint effect for firms with low financial
constraints (column 2).
In our final test, we consider the conditional relation using an alternative
proxy of business risk – idiosyncratic risk. Amit and Wernerfelt (1990) examine
the rationale for firms to reduce business risk and use idiosyncratic risk to
proxy for a firm’s business risk. We follow Amit and Wernerfelt (1990) and
measure idiosyncratic risk as the standard deviation of the residuals from the
firm-specific market model regression. High (low) business risk represents firms
in the top (bottom) quintile of idiosyncratic risk. Our results reported in
Panel C of Table 11 show a positive coefficient on stock liquidity and a negative
coefficient on the sum of stock liquidity and its interaction with OverFirm, that
are both statistically significant (p < 0.05), for firms with high business risk
(column 1). On the contrary, we do not find significant coefficients on the main
and joint effect for firms with low business risk (column 2). Our results from
Table 11 collectively suggest that stock liquidity is beneficial for enhancing
investment efficiency for firms with low institutional shareholdings, high
external financing dependence and high idiosyncratic risk.

6. Conclusion
This study examines the association between stock liquidity and investment
efficiency of firms listing on US stock exchanges, in particular, whether the
effect of stock liquidity on mitigating under (over)-investment problems is more
pronounced for firms with more financial constraints and information
asymmetry problems. The results show that the effect of higher stock liquidity
on lowering under (over)-investment is more pronounced for firms with more
financial constraints and information asymmetry problems as proxied by
younger and higher business risk firms, respectively. Our results are robust to a
series of robustness checks including endogeneity tests and alternative measures
of stock liquidity and investment efficiency. In additional analyses, we consider
other conditions or alternative proxies of financial constraints and business
risk. We find that the effect of higher stock liquidity on investment efficiency is
beneficial for firms with low institutional shareholding, high external financing
dependence and high idiosyncratic risk.
Our study has limitations. First, there is no single proxy developed to date
that can perfectly capture all three elements of stock liquidity, namely (i) depth
(the number of shares that can be traded at given bid and ask prices); (ii)
breadth (the size of market participants who are unable to exercise significant
market power regardless of their size) and; (iii) resiliency (how quickly prices
return to previous levels after experiencing price changes) (Hasbrouck, 2007).
Second, our findings are robust to a wide range of additional tests, but we
cannot completely rule out the possibility of our findings being driven by
omitted variables. Hence, our empirical results should be interpreted with some
caution. Furthermore, future research can extend our study to emerging
markets by examining the effect of stock liquidity on investment efficiency
under financing constraints and asymmetric information.
References
Admati, A. R., and P. Pfleiderer, 2009, The ‘Wall Street Walk’ and shareholder activism:
exit as a form of voice, The Review of Financial Studies 22(7), 2645–2685.
Altman, E., 1968, Financial ratios, discriminant analysis and the prediction of corporate
bankruptcy, The Journal of Finance 23(4), 589–609.
Amihud, Y., 2002, Illiquidity and stock returns: cross-section and time-series effects,
Journal of Financial Markets 5(1), 31–56.
Amit, R., and B. Wernerfelt, 1990, Why do firms reduce business risk?, Academy of
Management Journal 33(3), 520–533.
Appel, I. R., T. A. Gormley, and D. B. Keim, 2016, Passive investors, not passive
owners, Journal of Financial Economics 121(1), 111–141.
Arellano, C., Y. Bai, and J. Zhang, 2012, Firm dynamics and financial development,
Journal of Monetary Economics 59(6), 533–549.
Avery, C., J. A. Chevalier, and S. Schaefer, 1998, Why do managers undertake
acquisitions? An analysis of internal and external rewards for acquisitiveness, Journal
of Law, Economics, and Organization 14(1), 24–43.

Baker, M., J. C. Stein, and J. Wurgler, 2003, When does the market matter? Stock prices
and the investment of equity-dependent firms, The Quarterly Journal of Economics 118
(3), 969–1005.
Bakke, T.-E., and T. M. Whited, 2010, Which firms follow the market? An analysis of
corporate investment decisions, The Review of Financial Studies 23(5), 1941–1980.
Balakrishnan, K., M. B. Billings, B. Kelly, and A. Ljungqvist, 2014, Shaping liquidity:
on the causal effects of voluntary disclosure, The Journal of Finance 69(5), 2237–2278.
Bates, T. W., 2005, Asset sales, investment opportunities, and the use of proceeds, The
Journal of Finance 60(1), 105–135.
Bebchuk, L. A., and L. A. Stole, 1993, Do short-term objectives lead to under- or
overinvestment in long-term projects?, The Journal of Finance 48(2), 719–729.
Beck, T., and R. Levine, 2003, Stock markets, banks and growth: panel evidence,
Journal of Banking and Finance 28(3), 423–442.
Ben-Nasr, H., and A. A. Alshwer, 2016, Does stock price informativeness affect labor
investment efficiency?, Journal of Corporate Finance 38, 249–271.
Bertrand, M., and S. Mullainathan, 2003, Enjoying the quiet life? Corporate governance
and managerial preferences, Journal of Political Economy 111(5), 1043–1075.
Bharath, S. T., S. Jayaraman, and V. Nagar, 2013, Exit as governance: an empirical
analysis, The Journal of Finance 68(6), 2515–2547.
Biddle, G. C., and G. Hilary, 2006, Accounting quality and firm-level capital
investment, The Accounting Review 81(5), 963–982.
Biddle, G. C., G. Hilary, and R. S. Verdi, 2009, How does financial reporting quality
relate to investment efficiency?, Journal of Accounting and Economics 48(2/3), 112–131.
Bond, P., I. Goldstein, and E. S. Prescott, 2010, Market-based corrective actions, The
Review of Financial Studies 23(2), 781–820.
Brown, J. R., S. M. Fazzari, and B. C. Petersen, 2009, Financing external innovation
and growth: cash flow, equity, and the 1990s R&D boom, The Journal of Finance 64
(1), 151–185.
Brown, J. R., G. Martinsson, and B. C. Petersen, 2012, Do financing constraints matter
for R&D?, European Economic Review 56(8), 1512–1529.
Bushee, B. J., 1998, The influence of institutional investors in myopic R&D investment
behavior, The Accounting Review 73(3), 305–333.
Caporale, G. M., P. G. Howelles, and A. M. Soliman, 2004, Stock market development
and economic growth: causal linkage, Journal of Economic Development 29(1), 33–50.
Chen, Q., I. Goldstein, and W. Jiang, 2007, Price informativeness and investment
sensitivity to stock price, The Review of Financial Studies 20(3), 619–650.
Chen, Y., S. G. Rhee, M. Veeraraghavan, and L. Zolotoy, 2015, Stock liquidity and
managerial short-termism, Journal of Banking and Finance 60, 44–59.
Cheng, M., D. Dhaliwal, and Y. Zhang, 2013, Does investment efficiency improve after
the disclosure of material weaknesses in internal control over financial reporting?,
Journal of Accounting and Economics 56(1), 1–18.
Cheung, W. M., R. Chung, and S. Fung, 2015, The effects of stock liquidity on firm
value and corporate governance: endogeneity and the REIT experiment, Journal of
Corporate Finance 35, 211–231.
Cheung, W. M., R. W. Faff, H. J. Im, and S. Selvam, 2019, Stock liquidity and the
under-investment problem: a quasi-natural experiment in China, Working paper.
Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3335116.
Chung, K. H., and H. Zhang, 2011, Corporate governance and institutional ownership,
Journal of Financial and Quantitative Analysis 46(1), 247–273.
Chung, K. H., and H. Zhang, 2014, A simple approximation of intraday spreads using
daily data, Journal of Financial Markets 17, 94–120.

Collier, B., A. Haughwout, M. Stewart, H. Kunrether, and E. Michel-Kerjan, 2016,

Firm age and size and the financial management of infrequent shocks. Risk
Management and Decision Processes Center, The Wharton School, University of
Pennsylvania, Working Paper # 2016-09.
Corwin, S. A., and P. Schultz, 2012, A simple way to estimate bid-ask spreads from daily
high and low prices, The Journal of Finance 67(2), 719–760.
Dechow, P. M., and I. D. Dichev, 2002, The quality of accruals and earnings: the role of
accrual estimation errors, The Accounting Review 77(S1), 35–59.
Demirgüç-Kunt, A., and R. Levine, 1996, Stock markets, corporate finance and
economic growth, The World Bank Economic Review 10(2), 223–239.
Durnev, A., R. Morck, and B. Yeung, 2001, Does firm-specific information in stock
prices guide capital allocation?, NBER Working Paper No. 8093. Available at: http://
www.nber.org/papers/w8093.
Durnev, A., R. Morck, B. Yeung, and P. Zarowin, 2003, Does greater firm-specific
return variation mean more or less informed stock pricing?, Journal of Accounting
Research 41(5), 797–836.
Durnev, A., R. Morck, and B. Yeung, 2004, Value-enhancing capital budgeting and
firm-specific stock return variation, The Journal of Finance 59(1), 65–105.
Easley, D., and M. O’Hara, 2004, Information and the cost of capital, The Journal of
Finance 59(4), 1553–1583.
Edmans, A., 2009, Blockholder trading, market efficiency, and managerial myopia, The
Edmans, A., V. W. Fang, and E. Zur, 2013, The effect of liquidity on governance, The
Review of Financial Studies 26(6), 1443–1482.
Fama, E. F., 1980, Agency problems and the theory of the firm, Journal of Political
Economy 88(2), 288–307.
Fama, E. F., and K. R. French, 1997, Industry costs of equity, Journal of Financial
Economics 43(2), 153–193.
Fama, E. F., and M. C. Jensen, 1983, Separation of ownership and control, Journal of
Law and Economics 26(2), 301–326.
Fang, V. W., T. H. Noe, and S. Tice, 2009, Stock market liquidity and firm value,
Journal of Financial Economics 94(1), 150–169.
Fazzari, S. M., R. G. Hubbard, B. C. Petersen, A. S. Blinder, and J. M. Poterba, 1988,
Financing constraints and corporate investment, Brookings Papers on Economic
Activity 1, 141–206.
Fee, C. E., C. J. Hadlock, and J. R. Pierce, 2009, Investment, financing constraints, and
internal capital markets: evidence from the advertising expenditures of multinational
firms, The Review of Financial Studies 22(6), 2361–2392.
Fong, K. Y. L., C. W. Holden, and C. A. Trzcinka, 2017, What are the best liquidity
proxies for global research?, Review of Finance 21(4), 1355–1401.
Francis, J., R. LaFond, P. Olsson, and K. Schipper, 2005, The market pricing of
accruals quality, Journal of Accounting and Economics 39(2), 295–327.
Gillan, S. L., and L. T. Starks, 2000, Corporate governance proposals and shareholder
activism: the role of institutional investors, Journal of Financial Economics 57(2), 275–305.
Gompers, P. A., 1995, Optimal investment, monitoring, and the staging of venture
capital, The Journal of Finance 50(5), 1461–1489.
Gompers, P., J. Ishii, and A. Metrick, 2003, Corporate governance and equity prices,
Quarterly Journal of Economics 118(1), 107–156.
Goyenko, R. Y., C. W. Holden, and C. A. Trzcinka, 2009, Do liquidity measures
measure liquidity?, Journal of Financial Economics 92(2), 153–181.
Grossman, S. J., and O. D. Hart, 1986, The costs and benefits of ownership: a theory of
vertical and lateral integration, Journal of Political Economy 94(4), 691–719.

Grossman, S. J., and J. E. Stiglitz, 1980, On the impossibility of informationally efficient

markets, American Economic Review 70(3), 393–408.
Gujarati, D., 2003, Basic Econometrics, 4th edn (McGraw-Hill, New York, NY).
Ha, D. T., and V. X. Vinh, 2017, Stock market liquidity and firm investment: evidence
from Vietnam, Journal of International Business and Economics 5(1), 67–74.
Hadlock, C. J., and J. R. Pierce, 2010, New evidence on measuring financial constraints:
moving beyond the KZ index, The Review of Financial Studies 23(5), 1909–1940.
Haltiwanger, J., R. S. Jarmin, and J. Miranda, 2013, Who creates jobs? Small versus
large versus young, Review of Economics and Statistics 95(2), 347–361.
Hartzell, J. C., and L. T. Starks, 2003, Institutional investors and executive compen-
sation, The Journal of Finance 58(6), 2351–2374.
Hasbrouck, J., 2007, Empirical Market Microstructure: The Institutions, Economics and
Econometrics of Securities Trading (Oxford University Press, Oxford).
Heaton, J. B., 2002, Managerial optimism and corporate finance, Financial Management
31(2), 33–45.
Holmström, B., 1999, Managerial incentive problems: a dynamic perspective, The
Review of Economic Studies 66(1), 169–182.
Holmström, B., and J. Tirole, 1993, Market liquidity and performance monitoring,
Journal of Political Economy 101(4), 678–709.
Jensen, M. C., 1986, Agency costs of free cash flow, corporate finance, and takeovers,
American Economic Review 76(2), 323–329.
Jensen, M. C., and W. H. Meckling, 1976, Theory of the firm: managerial behavior,
agency costs and ownership structure, Journal of Financial Economics 3(4), 305–360.
Jiang, L., J.-B. Kim, and L. Pang, 2014, The influence of ownership structure, analyst
following and institutional infrastructure on stock price informativeness: international
evidence, Accounting and Finance 54(3), 885–919.
Kau, J. B., J. S. Linck, and P. H. Rubin, 2008, Do managers listen to the market?,
Journal of Corporate Finance 14(4), 347–362.
Khanna, N., and R. Sonti, 2004, Value creating stock manipulation: feedback effect of
stock prices on firm value, Journal of Financial Markets 7(3), 237–270.
Kyle, A. S., 1985, Continuous auctions and insider trading, Econometrica 53(6),
1315–1335.
Lang, M., and R. Lundholm, 1993, Cross-sectional determinants of analyst ratings of
corporate disclosures, Journal of Accounting Research 31(2), 246–271.
Lang, M., K. V. Lins, and M. Maffett, 2012, Transparency, liquidity, and valuation:
international evidence on when transparency matters most, Journal of Accounting
Research 50(3), 729–774.
Larcker, D. F., and T. O. Rusticus, 2010, On the use of instrumental variables in
accounting research, Journal of Accounting and Economics 49, 186–205.
Lennox, C. S., J. R. Francis, and Z. Wang, 2012, Selection models in accounting
research, The Accounting Review 87(2), 589–616.
Levine, R., and S. Zervos, 1998, Stock markets, banks, and economic growth, American
Economic Review 88(3), 537–558.
Luo, Y., 2005, Do insiders learn from outsiders? Evidence from mergers and
acquisitions, The Journal of Finance 60(4), 1951–1982.
Malmendier, U., and G. Tate, 2008, Who makes acquisitions? CEO overconfidence and
the market’s reaction, Journal of Financial Economics 89(1), 20–43.
Maug, E., 1998, Large shareholders as monitors: is there a trade-off between liquidity
and control?, The Journal of Finance 53(1), 65–98.
Myers, S. C., and N. S. Majluf, 1984, Corporate financing and investment decisions
when firms have information that investors do not have, Journal of Financial
Economics 13(2), 187–221.

Narayanan, M. P., 1985, Managerial incentives for short-term results, The Journal of
Finance 40(5), 1469–1484.
OECD, 2015, Young SMEs, growth and job creation. Available at: https://www.oecd.
org/sti/young-SME-growth-and-job-creation.pdf
Pang, J., and H. Wu, 2009, Financial markets, financial dependence, and the allocation
of capital, Journal of Banking and Finance 33(5), 810–818.
Park, H. D., and P. C. Patel, 2015, How does ambiguity influence IPO underpricing?
The role of the signalling environment, Journal of Management Studies 52(6),
796–818.
Petersen, M. A., 2009, Estimating standard errors in finance panel data sets: comparing
approaches, The Review of Financial Studies 22(1), 435–480.
Rajan, R. G., and L. Zingales, 1998, Financial dependence and growth, American
Economic Review 88(3), 559–586.
Ramalingegowda, S., and Y. Yu, 2012, Institutional ownership and conservatism,
Journal of Accounting and Economics 53(1/2), 98–114.
Rauh, J. D., 2006, Investment and financing constraints: evidence from the funding of
corporate pension plans, The Journal of Finance 61(1), 33–71.
Rego, S. O., and R. Wilson, 2012, Equity risk incentives and corporate tax
aggressiveness, Journal of Accounting Research 50(3), 775–810.
Richardson, S., 2006, Over-investment of free cash flow, Review of Accounting Studies 11
(2/3), 159–189.
Roll, R., 1986, The hubris hypothesis of corporate takeovers, The Journal of Business 59
(2), 197–216.
Sanders, W. G., and S. Boivie, 2004, Sorting things out: valuation of new firms in
uncertain markets, Strategic Management Journal 25(2), 167–186.
Schreyer, P.2000, High-growth firms and employment, OECD Science, Technology and
Industry working papers, 2000/03.
Shleifer, A., and R. W. Vishny, 1997, A survey of corporate governance, The Journal of
Finance 52(2), 737–783.
Stein, J. C., 1988, Takeover threats and managerial myopia, Journal of Political
Economy 96(1), 61–80.
Stiglitz, J. E., and A. Weiss, 1981, Credit rationing in markets with imperfect
information, American Economic Review 71(3), 393–410.
Stock, J. H., J. H. Wright, and M. Yogo, 2002, A survey of weak instruments and weak
identification in generalized method of moments, Journal of Business and Economic
Statistics 20(4), 518–529.
Subrahmanyam, A., and S. Titman, 2001, Feedback from stock prices to cash flows, The
Velury, U., and D. S. Jenkins, 2006, Institutional ownership and the quality of earnings,
Journal of Business Research 59(9), 1043–1051.
Verdi, R. S.2006, Financial reporting quality and investment efficiency, Working paper,
Massachusetts Institute of Technology. Available at: https://ssrn.com/abstract=
930922.
Williamson, O., 1985, The Economic Institutions of Capitalism: Firms, Markets,
Relational Contracting (Free Press, New York, NY).
Wurgler, J., 2000, Financial markets and the allocation of capital, Journal of Financial
Economics 58(1/2), 187–214.
Xiong, J., and D. Su, 2014, Stock liquidity and capital allocation efficiency: evidence
from Chinese listed companies, China Journal of Accounting Studies 2(3), 228–252.
Zhang, X. F., 2006, Information uncertainty and stock returns, The Journal of Finance
61(1), 105–136.

Appendix
Notations and definitions of variables
Variables Definition Source
INV The sum of research and development (R&D), capital Compustat

expenditure and acquisition expenditure less cash receipts
from sale of property, plant and equipment multiplied by 100
and scaled by lagged total assets.
Capex Capital expenditure, computed as the capital expenditure in Compustat
year t + 1 scaled by lagged total assets.
Non_Capex Non-capital expenditure, calculated as the sum of R&D Compustat
expenditures and acquisitions in year t + 1 scaled by lagged
total assets.
LIQFHT The cost of trading as a percentage of stock price, measured as a CRSP
function of return volatility and proportion of zero returns
(Fong et al., 2017). This measure is multiplied by minus 100.
The higher (lower) values of this measure indicate higher
(lower) stock liquidity.
LIQBAS The closing percent quoted spread (LIQBAS) developed by CRSP
Chung and Zhang (2014), measured as the yearly average of
daily closing bid-ask spread scaled by the mean of daily
closing bid and closing ask prices. This measure is multiplied
by minus 100 and is excluded from analyses if: (1) daily
negative bid-ask spreads (crossed quotes) for which the ask
price is smaller than the bid price (Balakrishnan et al., 2014);
or (2) the daily bid-ask spread that is greater than 50% of the
quote midpoint for which the bid-ask spread is unreasonably
larger than the mean of ask and bid price (Chung and Zhang,
2014); or (3) stocks that have less than 200 days of bid and ask
price data available in the CRSP daily file during fiscal year t
(this allows more reliable estimation of the parameters). The
higher (lower) values of this measure indicate higher (lower)
stock liquidity.
LIQindex Composite score of stock liquidity, a continuous variable
computed as the standardised average of LIQFHT and
LIQBAS. The higher (lower) values of this measure indicate
higher (lower) stock liquidity.
LIQAM −1 × (the natural logarithm of one plus firm i’s Amihud (2002) CRSP
illiquidity ratio), where the Amihud illiquidity ratio is
calculated as the daily ratio of absolute value of stock returns
to dollar volume, averaged over firm i’s fiscal year t. Stock
included in the computation meets three criteria: (1) it has at
least 200 days of return and volume data, (2) stock price > $5
at the end of the fiscal year, and (3) it is required to be listed at
the end of its fiscal year.
LIQHL
(continued)

(continued)
Appendix (continued)
The price-based spread proxy developed by Corwin and Schultz Professor Shane
(2012), multiplied by minus one. The higher (lower) values of A. Corwin’s
this measure indicate higher (lower) stock liquidity. The data website
can be downloaded from http://www3.nd.edu/~scorwin/.
LIQindex4 A composite stock liquidity index (LIQindex4) compiled by
standardising the two additional measures of liquidity
(LIQAM, LIQHL) and the two liquidity measures from the
main analyses (LIQFHT, LIQBAS) and then calculating the
unweighted mean of these four standardised measures.
Firm_Age The difference between the first year when the firm appears in CRSP
CRSP and the current year.
IncVol Operating income volatility (IncVol), measured as the standard Compustat
deviation of annual income before extraordinary items divided
by total assets for the 5-year window from t−4 to t.
OverFirm A decile ranked variable used to distinguish between settings Compustat
where over- or under-investment is more likely; OverFirm is
increasing in the likelihood of over-investment (it is the
average of ranked values of two partitions variables, i.e., cash
balance (total cash and short-term investments scaled by total
assets) and leverage (multiplied by minus one) which are used
as ex-ante firm-specific characteristics that are likely to affect
the likelihood that a firm will over- or underinvest).
AQ Standard deviation of the firm-level residuals from the model of Compustat
Francis et al. (2005) (which is adapted from Dechow and
Dichev, 2002) during the years t−5 to t−1 and multiplied by
minus one. The model is a regression model of total current
accruals on lagged, current and future cash flows, plus the
change in revenues, PPE for firm i at year t. All variables are
scaled by average total assets. The model is estimated cross-
sectionally for each industry with at least 20 observations in a
given year based on the Fama and French (1997) 48-industry
classification.
IO The percentage of firm shares held by institutional investors. Thomson Reuters
13F
Analyst The number of analysts following the firm. IBES
G_Score The measure of investor protection rights created by Gompers Andrew
et al. (2003), multiplied by minus one. http://faculty.som.ya Metrick’s
le.edu/andrewmetrick/data.html website
G_Dummy An indicator variable that takes the value of one if G_Score is
missing, and zero otherwise.
LogAsset The log of total assets. Compustat
MB The ratio of the market value to the book value of total assets. Compustat
σ(CFO) Standard deviation of the cash flow from operations deflated by Compustat
average total assets from year t−5 to t−1.
σ(Sales) Standard deviation of the sales deflated by average total assets Compustat
from year t−5 to t−1.
σ(INV) Compustat
(continued)

(continued)
Appendix (continued)
Standard deviation of investment deflated by average total

assets from year t−5 to t−1.
Z_Score A measure of distress computed following the methodology in Compustat
Altman (1968).
Tangibility Tangibility is the ratio of PPE to total assets. Compustat
Ind_Kstruct The mean market leverage for firms in the same SIC 3-digit Compustat
industry. Market leverage computed as the ratio of long-term
debt to the sum of long-term debt to the market value of
equity.
CFOsale The ratio of CFO to sales. Compustat
DIV Dividend is an indicator variable that takes the value of one if Compustat
the firm paid a dividend, and zero otherwise.
OCycle Operating cycle is the log of receivables to sales plus inventory Compustat
to COGS multiplied by 360.
Loss An indicator variable that takes the value of one if net income Compustat
before extraordinary item is negative, and zero otherwise.
EF External financing dependence is measured as the industry- Compustat
median value of the difference between capital expenditure and
cash flow from operations scaled by capital expenditure.
IdioRisk Idiosyncratic risk is the standard deviation of the residuals from CRSP
the firm-specific market model regression.
Industry FE Industry fixed effects based on the Fama-French (1997) 48- Kenneth
industry classifications. French’s
website

10 1111@acfi 12656 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

10 1111@acfi 12656 PDF

Uploaded by

Copyright:

Available Formats

Accounting & Finance

The eﬀect of stock liquidity on investment eﬃciency

Heidi Quah, Janto Haman , Dharmendra Naidu

We examine whether the association between stock liquidity and investment

Key words: Stock liquidity; Investment eﬃciency; Financial constraint;

JEL classiﬁcation: M40, M41, M48, G12, G32, G38

Firm investments are a major factor through which ﬁrms increase

Please address correspondence to Janto Haman via email: janto.haman@monash.edu

© 2020 Accounting and Finance Association of Australia and New Zealand

© 2020 Accounting and Finance Association of Australia and New Zealand

© 2020 Accounting and Finance Association of Australia and New Zealand

lower institutional shareholdings, higher external ﬁnancing dependence and

2. Literature review and hypotheses development

2.1. Literature review

© 2020 Accounting and Finance Association of Australia and New Zealand

Malmendier and Tate, 2008). As such, it is important to identify factors that

© 2020 Accounting and Finance Association of Australia and New Zealand

feedback eﬀect by making prices more informative for market participants to

2.2. Hypotheses development

We speciﬁcally examine whether the association between stock liquidity and

© 2020 Accounting and Finance Association of Australia and New Zealand

2.2.1. Hypotheses 1a and 1b

H1a: The eﬀect of higher stock liquidity on lowering under-investment is more

H1b: The eﬀect of higher stock liquidity on lowering over-investment is more

2.2.2. Hypotheses 2a and 2b

© 2020 Accounting and Finance Association of Australia and New Zealand

managers to act opportunistically by investing suboptimally. Further, due to

H2a: The eﬀect of higher stock liquidity on lowering under-investment is more

H2b: The eﬀect of higher stock liquidity on lowering over-investment is more

© 2020 Accounting and Finance Association of Australia and New Zealand

3. Sample selection and research method

3.1. Sample selection

3.2. Research method

Before moving on to testing our hypotheses, we ﬁrst examine whether higher

© 2020 Accounting and Finance Association of Australia and New Zealand

Panel A: Sample selection Observations

Compustat data for years between 1992 and 2014 143,610

Panel B: Industry distribution of sample firms

Fama and French industry name Number of ﬁrms % of sample

Business Services 7,004 13.58

Year Number of ﬁrms % of sample

1992 1,855 3.60

© 2020 Accounting and Finance Association of Australia and New Zealand

Panel C: Year distribution of sample firms

Year Number of ﬁrms % of sample

2005 2,266 4.39

classiﬁcation to control for industry-speciﬁc shocks to investment. Our

INVi,tþ1 ¼ α þ β1 LIQi,t þ β2 LIQi,t OverFirmi,t þ β3 OverFirmi,t

INV is the sum of capital expenditures, R&D expenditures and acquisitions,

© 2020 Accounting and Finance Association of Australia and New Zealand

In order to assess whether stock liquidity can alleviate under (over)-

© 2020 Accounting and Finance Association of Australia and New Zealand

4.1. Summary statistics

© 2020 Accounting and Finance Association of Australia and New Zealand

4.2. Stock liquidity and investment eﬃciency

Table 3 reports the regression results of investment on stock liquidity

n Mean Std. dev. Min. Median Max.

INVt+1 (%) 51,576 15.323 16.714 −2.271 10.132 146.860

© 2020 Accounting and Finance Association of Australia and New Zealand

n Mean Std. dev. Min. Median Max.

IO 51,576 0.497 0.286 0.000 0.515 1.000

(LIQBAS) [LIQindex] is associated with an increase of under-investment by 1.2

© 2020 Accounting and Finance Association of Australia and New Zealand

LIQ (1) Under-invest + 1.283* 0.524* 1.790***

CFOsale −0.800* −0.775* −0.789***

CFOsale −0.663 −1.936

Tangibility 0.043 9.392* 7.562* 7.925* 10.277*

LIQ (1) Under-invest 2.556* 1.060* 1.004 3.541* 2.063***

LIQ (1) Under-invest 2.028* 0.996* 1.341 3.975* 2.962***

LIQ (1) Under-invest 4.078 1.718 −1.779 0.400

LIQ (1) Under-invest 3.583 1.242 −2.611 0.682