Analyst coverage, executive compensation and corporate risk‐taking: Evidence from property–casualty insurance firms

• Published date: 01 December 2023
• Author: Tao Chen,Shinichi Kamiya,Pingyi Lou,Andreas Milidonis
• Date: 01 December 2023
• DOI: http://doi.org/10.1111/jori.12437

Received: 3 February 2021

|

Revised: 25 May 2023

|

Accepted: 4 June 2023

DOI: 10.1111/jori.12437

ORIGINAL ARTICLE

Analyst coverage, executive compensation

and corporate risk‐taking: Evidence from

property–casualty insurance firms

Tao Chen

1

|Shinichi Kamiya

1

|Pingyi Lou

2

|

Andreas Milidonis

3

1

Nanyang Business School, Nanyang

Technological University, Singapore,

Singapore

2

School of Economics, Fudan University,

Shanghai, China

3

Department of Accounting and Finance,

School of Economics and Management,

University of Cyprus, Nicosia, Cyprus

Correspondence

Andreas Milidonis, University of Cyprus,

Nicosia, Cyprus.

Email: Andreas.Milidonis@ucy.ac.cy

Funding information

National Natural Science Foundation of

China, Grant/Award Number: 72173005

Abstract

Using an exogenous drop in analyst coverage intro-

duced by broker closures and mergers, we test for

the causal impact of analyst coverage on corporate

risk‐taking, in an opaque industry. We document an

increase in risk using several book‐based and market‐

based risk measures, including tail and default risk

measures. Results are driven by firms with stronger

managerial risk‐taking compensation incentives. The

increase in risk is stronger in more opaque firms, and

firms with weaker policyholder monitoring. Firm risk

increases through at least one risk‐taking action,

such as investing firm assets in higher‐risk bonds.

Our study highlights the importance of stock analysts

in affecting corporate risk‐taking, especially in the

presence of stronger managerial, compensation risk‐

taking incentives.

KEYWORDS

analyst coverage, compensation incentives, insurance,

risk‐taking

JEL CLASSIFICATION

G22, G32

Journal of Risk and Insurance. 2023;90:899–939. wileyonlinelibrary.com/journal/JORI

|

899

© 2023 American Risk and Insurance Association.

1|INTRODUCTION

Information intermediaries in financial markets transform the abundance of public and private

information about the firms they follow into an easily comprehensible signal for investors,

customers, and regulators. Popular examples of information intermediaries are stock analysts,

who serve shareholders and potential investors primarily, and credit rating agencies, who also

inform customers and regulators.

1

In addition to their information production role, which plays

an indirect monitoring role on the firm, information intermediaries also directly monitor firm

operations (e.g., Chen et al., 2015). While prior literature documents that various firm

stakeholders, such as banks and bondholders, can detect and impact corporate risk‐taking, the

role played by information intermediaries receives little attention.

2

This study fills this gap in the literature by focusing on stock analysts as our representative

information intermediary to examine their impact on firm risk. To address the potential

endogeneity issues that have led to limited research on the topic,

3

we use two natural

experiments that directly affect firms' analyst coverage but are not directly related to firms'

risk‐taking: brokerage closures (Kelly & Ljungqvist, 2012) and broker mergers (Hong &

Kacperczyk, 2010).

4

These events disrupt the information production process, while not being

directly related to firms' risk level. This identification strategy of using the exogenous drop in

analyst coverage, allows us to draw causal inferences about their impact on firm risk.

How do we expect analysts to impact firm risk? It is possible that stock analysts cater to their

customers which are existing shareholders and potential investors. According to the asset

substitution literature, shareholders prefer more risk than debtholders (e.g., Jensen &

Meckling, 1976). On the other hand, firm managers are considered risk‐averse and undiversified

with respect to their firm‐dependent wealth.

5

Moreover, managers face several different

mechanisms to suppress risk (such as credit rating agencies, regulators, and policyholder

monitoring). Recognizing this situation, shareholders offer managers compensation incentive

contracts that reward risk‐taking (Garen, 1994). If designed optimally, such contracts would

provide managerial incentives to increase firm risk at levels desired by shareholders.

Managers are typically shown to respond to such compensation risk‐taking incentives

(e.g., Cohen et al., 2000; Coles et al., 2006; Guay, 1999; Low, 2009). In the banking and

insurance industry, government guaranty mechanisms such as deposit insurance for banks and

1

However, information intermediaries are not problem free. See, for instance, the potential conflicts of interest present

in the ratings industry (e.g., Beaver et al., 2006; Berwart et al., 2019) and the long literature on bias present in analysts'

forecasts (Das et al., 1998).

2

A case in point about the important role of analysts, is Meredith Whitney, who was an analyst at the brokerage house

Oppenheimer in 2007. She issued a particularly pessimistic, but accurate, research report on Citigroup on October 31,

2007, arguing that Citibank might go bankrupt as a result of its subprime mortgage holdings. The industry took note,

and Fortune listed Whitney as one of the “50 Most Powerful Women in Business”https://money.cnn.com/magazines/

fortune/most-powerful-women/2011/snapshots/48.html.

3

The first concern is reverse causality. Risk‐taking might influence firms' information environment and, thus, analyst

coverage. For instance, McNichols and O'Brien (1997) reveal the presence of a self‐selection bias in analysts' ratings:

higher (lower) ratings assigned to newly covered (dropped) stocks are associated with fundamental information about

the stocks. The second potential concern relates to omitted variables. Unobservable corporate investment opportunities

could drive both analyst coverage and risk‐taking.

4

See, for example, Kelly and Ljungqvist (2012), who provide supporting evidence that the brokerage closures are not

motivated by negative information about individual stocks.

5

Risk‐averse managers may prefer a quiet life in some industries (e.g., Bertrand & Mullainathan, 2003), prefer a longer

career (e.g., Amihud & Lev, 1981), and choose a lower level of risk.

900

|

CHEN ET AL.

guaranty funds for insurance companies, may lead to excessive risk‐taking and other well‐

known moral hazard problems (e.g., Lee et al., 1997; Calomiris & Jaremski, 2019). In the case of

the insurance industry, such excessive risk levels can increase firm default risk (Milidonis &

Stathopoulos, 2011), thus increasing the probability of bankruptcy, inducing regulatory costs,

and harming investors.

There are also risk‐mitigating mechanisms that managers face. For example, monitoring from

regulators and customers, may discourage risk‐taking that might impact tail risk and default risk

measures (e.g., Epermanis & Harrington, 2006). Moreover, career concerns that managers might

face if their firms are close to default, might deter them from responding to compensation risk‐

taking incentives and even push them to decrease firm risk (Milidonis & Stathopoulos, 2014).

Consequently, if the convexity of managerial compensation and its anticipated increase on

managers' expected wealth (through an increase in firm risk), is not able to prevail over the

impact that managerial risk aversion has on managers' utility (through a decrease in firm risk),

then analysts might have a role to play. As a result, if the compensation risk‐taking incentives

are strong, we expect that analyst coverage will encourage a decrease in firm risk. On the other

hand, in cases where compensation risk‐taking incentives are weak due to shareholders' effort

to reduce the agency cost of debt (John & John, 1993) or perhaps because of weak corporate

governance (Chen et al., 2015), analyst coverage might encourage firm risk.

We focus our study on the U.S. publicly traded property and casualty (PC) insurance industry

for two reasons. First, the risk‐shifting problem in the insurance industry is more involved

because of the presence of policyholders in addition to debtholders. Moreover, the presence of

government guarantees in some lines of the insurance business directly impacts the strength of

policyholder monitoring. For example, Epermanis and Harrington (2006) find that companies

writing more commercial than personal insurance businesses, are more sensitive to firm risk,

because personal lines are covered to a larger extent by insurance guarantee funds than

commercial lines. Hence, we can test for the impact of the exogenous drop in analyst coverage on

firm risk, using the cross‐sectional variation in compensation risk‐taking incentives, and also in

the degree of policyholder monitoring within the PC insurance industry.

Second, PC insurers are more opaque than other (nonfinancial) firms (Chiang et al., 2022;

Epermanis & Harrington, 2006; Morgan, 2002). Insurers' opaqueness is largely a result of the

complexity of insurance assets (e.g., holdings of bonds and stocks) and liabilities (e.g., the

underwriting for homeowners, motor, and liability insurance, among others), and the risk

associated with these assets and liabilities. Since opaqueness also varies within the industry, we

expect that the effect of analyst coverage on firm risk through its informational role (Kim

et al., 2019) will be larger in more opaque firms.

Our sample comprises seven broker closures and mergers to the PC insurance industry that meet

our no overlapping requirement. Our base case scenario has 760 firm‐year observations with 114

unique insurance companies from 2001 to 2017. Using a set of market‐based and book‐based risk

measures, we examine the causal impact of the exogenous drop in analyst coverage on the same risk

measures, using a multiple‐shock, difference‐in‐differences (DID), and panel regression model.

Our primary findings are summarized as follows. First, we document an increase in firm risk as

perceived by investors and reflected in the measures of systematic and unsystematic risk, and

measures of excessive risk (such as tail and default risk) that we use. Second, we test the

mechanism provided by shareholders to managers to encourage risk‐taking by constructing CEO

compensation vega (Guay, 1999) and find that the overall increase in risk is driven by the

subsample of insurers with high vega, suggesting that stronger compensation risk‐taking incentives

cause excessive risk‐taking and analyst coverage suppresses risk. Third, the result is more apparent

CHEN ET AL.

|

901