Printer Friendly

Corporate governance and cash holdings: evidence from the U.S. property-liability insurance industry.

ABSTRACT

This article examines the impact of board and finance committee characteristics on insurers' cash holdings using a sample of 1,454 U.S. stock property-liability insurer-year observations. We focus on the roles of independent board members and independent finance committee members. Our results suggest that independent board members allow managers to hold excess cash holdings to avoid underinvestment and play a monitoring role in managers' cash spending behavior in a regulated industry. The overall findings are consistent with the independent director responsibility hypothesis, which suggests that independent directors play a monitoring role in managers' cash spending behavior and avoiding underinvestment problems.

INTRODUCTION

This article examines the impact of corporate governance on cash holdings in the U.S. property-liability insurance industry. The literature suggests that firm managers need enough cash to fund all investment opportunities and to ensure a firm's overall liquidity at the same time. If they hold unjustified excess cash, however, managers can build their own empires, consume perquisites that benefit only themselves, and be less careful about efficient operation at the expense of shareholders. (1)

Effective internal and external corporate control mechanisms to mitigate agency conflicts over excess cash holdings have been of academic interest since Jensen's (1986) work on the agency theory of free cash flow. Studies on the association between cash holdings and shareholder rights for industrial firms provide mixed evidence. Dittmar, Mahrt-Smith, and Servaes (2003), Pinkowitz, Stulz, and Williamson (2004), and Kalcheva and Lins (2007) show that firms with greater shareholder rights have lower levels of cash holdings, suggesting that shareholders who have enough power will force managers to pay out cash to the shareholders and prevent managers from accumulating excess cash to serve their own interests.

Harford, Mansi, and Maxwell (2008), however, find that firms with weaker shareholder rights and low insider ownership have less in cash reserves than firms with stronger shareholder rights (low Governance Index [G-Index]) and high insider ownership. They present evidence that low cash reserves in poorly governed firms occur because managers spend cash on capital expenditures and acquisitions that reduce firm value. Similar results are found in Dittmar and Mahrt-Smith (2007). Dittmar and Mahrt-Smith also find that the negative impact of excess cash on operating performance will dissipate in a firm with stronger governance structures. It should be noted that these prior studies generally exclude financial firms and do not control for the impact of external control by regulators.

The purpose of this article is to examine the impact of corporate governance on cash holdings in property-liability insurers. This industry provides an exceptionally good setting for analyzing the effectiveness of insurers' internal corporate governance mechanisms in mitigating agency conflicts. First, property-liability insurers usually generate and hold substantial amounts of cash; virtually all their business transactions are completed in cash. (2) Second, huge losses (e.g., the recent earthquake and tsunami in Japan) may result in illiquidity or insolvency in the insurance companies. Previous studies (e.g., Grace, Klein, and Phillips, 2005) provide evidence that insurer insolvencies are, on average, three to five times more expensive than other financial institutions, and the cost of insolvency is inversely related to insurer liquidity.

Third, the insolvency of insurance companies may be worse than that of manufacturing firms because policyholders buy insurance policies and pay premiums for necessity (e.g., auto and homeowner insurance policies), while stockholders or creditors invest in manufacturing firms for profits (e.g., capital gains, dividends, and interests). Although policyholders are protected by guaranty funds, it takes time for policyholders to get reimbursed for the premiums paid and they may not be fully compensated. Finally, Demsetz and Lehn (1985) suggest that regulation "provides some subsidized monitoring and disciplining of the management of regulated firms" (p. 1161). The insurance industry is subject to regulations, which could represent a source of external corporate control. The insurance industry setting allows us to examine the efficacy of firms' internal control systems while controlling for the possible substitution effect of external corporate control mechanisms. In addition, our results can shed new light on whether corporate governance in a regulated industry such as the insurance industry is needed.

This article focuses on the effect of independence of boards and finance committees, among other corporate governance variables. The literature suggests independent directors do not have economic or psychological ties to management and thus should have more incentive to monitor managers. More important, the literature generally suggests that board of director independence is the most effective monitoring characteristic (Anderson, Mansi, and Reeb, 2004). Also, financial committees are responsible for financial affairs and investment activities. Based on the above arguments, we propose the independent director responsibility hypothesis that suggests that independent directors play a monitoring role in managers' cash spending behavior and avoiding underinvestment and insolvency problems.

We investigate the impact of boards and finance committees on insurers' cash holdings using a sample of 1,454 U.S. stock property-liability insurer-year observations over 1997-2002. (3) The empirical results are summarized below. We find insurers with a higher proportion of outsiders on their boards and finance committees hold larger cash holdings. Additional results show that the independence of boards and finance committees is positively related to the growth of an insurer when it has excess cash holdings. These two results imply that independent board and finance committee members allow managers to hold excess cash holdings to avoid underinvestment. We further find that excess cash spent on salary insignificantly declines with a higher proportion of outside directors on a board or a finance committee. These results suggest that independent board and finance committee members play a monitoring role in managers' cash spending behavior even in a regulated industry. The evidence overall is consistent with the independent director responsibility hypothesis that suggests independent board members and independent finance committee members not only aim to avoid underinvestment problems but also play monitoring roles.

Other important findings are summarized below. We find that large board size and duality are associated with more cash holdings. Further examination shows that these two variables are positively related to growth opportunities as proxied by Tobin's Q and negatively related to expenses, when insurers have positive cash holdings in the prior period. The evidence also shows that insurers with more diligent board and more financial expert on financial committee are positively related to growth opportunities when insurers have positive excess cash. Equally important, we also find excess cash spent on salary expenses does not increase with more financial expert on the financial committee. In summary, our overall results suggest that other than board or committee director independence, some corporate governance characteristics can further contribute to strengthening board-monitoring performance.

Our study provides several important contributions to the literature. First, this is the first study to examine the impact of the roles of boards and finance committees on cash holdings for property-liability insurers. Previous studies examine similar issues, but exclude financial firms. Second, this study examines the role of internal governance (such as director independence) while controlling for the role of external governance (such as regulation). We find corporate governance is important, even in a regulated industry such as the insurance industry. Finally, we find independent board members and independent finance committee members play ideal roles in corporate governance: avoiding underinvestment problems and monitoring agency costs of excess cash holdings. These empirical results should help regulators, managers, policyholders, and investors to determine which board characteristics can help to maximize stakeholders' wealth.

The article is organized as follows. The second section develops the hypotheses related to the association between insurers' cash holdings and a set of corporate governance variables. The third section describes the sample selection criteria, models, and variable definitions and provides descriptive statistics of the data. The fourth, fifth, and sixth sections report the empirical results. The seventh section offers conclusions.

HYPOTHESES DEVELOPMENT

This section discusses the association between corporate governance and cash holdings.

The Effectiveness of the Board and the Finance Committee

Opler et al. (1999) provide an excellent survey about the determinants and the impact of agency costs on cash holdings for nonfinancial companies. Inappropriately high levels of cash holdings lead to several agency costs, which are harmful to stockholders' wealth. (4) First, a high level of cash holdings will diminish pressures on managers to minimize costs or develop managers' discretionary expense-preference behaviors (Mester, 1989). Second, managers are usually replaced if their firms face financial distress or takeover. Self-interested managers generally prefer a higher level of cash holdings than the required level to reduce the likelihood of financial distress or defend takeover threats for their job security (Pinkowitz, 1998; Harford, 1999). Unfortunately, cash generates lower returns than do other liquid assets. Third, poorly controlled managers tend to retain cash to facilitate their perquisite consumption rather than to distribute to stockholders as dividends. Of all assets, cash reserves are the most easily accessible to managers with little scrutiny, and much of their use is discretionary. Harford, Mansi, and Maxwell (2008) suggest that firms with stronger governance structures tend to pay back cash to shareholders through increasing dividends, which usually signal a long-term commitment; firms with weaker governance structures tend to pay back cash through stock repurchases, which require no future commitment.

Performing the effective monitoring role of reducing the aforementioned agency costs of holding a higher level of cash, directors can attempt to deter self-interested managers from retaining excess cash reserves. Stulz (1990) and Myers and Majluf (1984) present models to demonstrate the underinvestment problem resulting from insufficient cash reserves. There is a possibility that boards may allow managers to hold large cash reserves to avoid underinvestment problems when insurers have positive net present value projects or future growth opportunities are expected.

Prior studies on the efficacy of corporate governance focus on characteristics of boards. We also examine the efficacy of characteristics of finance committees. It is important to include characteristics of finance committees because finance (or investment) committees are responsible for financial affairs and investment activities. (5)

Although boards of directors are responsible for general oversight of investment and financial affairs, the oversight is often delegated to a subcommittee of the full board that is the finance committee. The finance committee approves the investment guidelines that provide standards to ensure portfolio liquidity and safety, approves the hiring of investment managers and custodians for portfolio assets, and deals with other matters regarding the financial affairs of the company. Independent committee members have the responsibilities to review and evaluate information relating to the company's investable assets, investment policies, strategies, and objectives, and to select long-term investment and financing strategies.

After we examine the effectiveness of the board and the finance committee on the level of cash holdings, we need to test whether the excess cash holdings in the previous period are held to meet higher growth in the next period and whether the excess cash holdings in the previous period are not wasted by unconstrained managers in the next period. Excess (unexplained) cash is defined as the residual from a regression of cash holdings with a number of control variables used by Colquitt, Sommer, and Godwin (1999) and year dummy variables. We further decompose excess cash into positive excess cash and negative excess cash to test all main governance variables of interest.

Regulation and Cash Holdings

We next examine the impact of regulation on cash holdings and its implication on corporate governance in property-liability insurance companies. We first look for the objectives for insurance regulation. In general, there are four main objectives: maintain insurer solvency, compensate for inadequate consumer knowledge, ensure reasonable rates, and make insurance available. We believe the first objective, maintaining insurer solvency, is related to cash holdings and examine all Financial Analysis and Solvency Tracking (FAST) and Insurance Regulatory Information

System (IRIS) ratios that are related to cash holdings for property and liability companies. For example, ratio 8 (liabilities to liquid assets ratio) in IRIS ratios is adjusted liabilities divided by liquid assets. Liquid assets include bonds, stocks, cash, short-term investments, and receivable for securities plus accrued investment income; but exclude investments in affiliated companies. The acceptable range of this ratio in the period associated with our study is greater than 105 percent. In addition, the literature suggests that many insurers that later become insolvent reported an increasing liabilities to liquid assets ratio in the years approaching insolvency. It is clear from the above discussion that one objective of regulators related to cash holdings is to make sure insurance companies have enough liquidity. With this conclusion in mind we develop our hypotheses in the next section.

Corporate Governance and Cash Holdings

The literature examining the effect of corporate governance on cash holdings focuses mainly on one attribute of governance, shareholder rights (e.g., Dittmar, MahrtSmith, and Servaes, 2003; Dittmar and Mahrt-Smith, 2007; Harford, Mansi, and Maxwell, 2008), rather than on a broader set of corporate governance attributes on constraining managerial opportunism. Some governance attributes may complement one another in protecting stakeholders' interests, while other governance attributes may be substitutes. Therefore, inferences drawn from studying one attribute of governance may suffer from the omitted variables problem.

In 2002, Standard & Poor's developed a comprehensive framework for evaluating corporate governance based on four governance components, including ownership structure and influence, financial stakeholders' rights and relations, financial transparency and disclosure, and board structure and processes (Standard & Poor's Corporate Governance, 2002). We use this framework along with evidence in previous studies to identify the board and the finance committee attributes potentially associated with the effectiveness of boards and finance committees. These attributes examined in our study are also emphasized by the U.S. Securities and Exchange Commission (SEC, 2009) in its new rules to enhance disclosures relating to the functioning of board of directors.

Board and Finance Committee Independence

As mentioned above, the focus of this study is on the effect of independence of boards and finance committees. The literature generally suggests that board of director independence is the most effective aspect of governance (Anderson, Mansi, and Reeb, 2004).

The independence of a board is typically used to measure its effectiveness of monitoring excess cash holdings for several reasons. First and the most important, independent directors without economic, social, or psychological ties to the management have more motivation to challenge self-interested managers without fear of rocking the boat (Berton, 1995). The second reason is a reputation capital preservation/development motivation discussed by Fama and Jensen (1983). An independent director generally sees better monitoring as furthering developing his/ her reputation as a governance expert, especially when specifically assigned to monitor managers because of his/her independence. The above arguments suggest a negative association between the independence of directors and the level of cash holdings.

On the other hand, Harford, Mansi, and Maxwell (2008) propose the shareholder power hypothesis and suggest that insurers whose shareholders have better control of managers will allow managers to have more cash holdings to prevent the underinvestment problem. (6) The reason is that it is more costly to raise capital for new projects from the capital market because of information asymmetry between managers and investors in the capital market (Myers and Majluf, 1984; Stulz, 1990). We believe independent directors try to maximize shareholder wealth and have better control of managers. Therefore, the association between the independence of directors and the level of cash holdings is positive.

Given the above arguments, we propose the independent director responsibility hypothesis, which suggests responsible directors of boards and finance committees will maximize shareholder wealth using two approaches. First, independent directors will allow managers to hold large cash holdings to prevent underinvestment problems. These expectations lead to our first two testable hypotheses:

H1: There is a positive association between the level of insurers' cash holdings and the independence of board of directors (finance committees).

To provide additional evidence to support the independent director responsibility hypothesis, we need to test whether a higher fraction of independent directors would have higher growth in the next period when insurers have positive excess cash holdings in the previous period. We provide the next set of hypotheses below.

H2: There is a positive association between the growth of insurers and the independence of board of directors (finance committees) when insurers have excess cash holdings in the previous period.

Second, independent directors will constrain managers to reduce the agency costs related to excess cash holdings. To test this part of the independent director responsibility hypothesis, we propose the following arguments. When insurers have excess cash holdings in the previous year, insurers with a higher fraction of independent directors would not have higher expenses next year. A nonpositive relation between independent boards and finance committees and cash holdings is predicted.

H3: There is a nonpositive association between the expenses of insurers and the independence of board of directors (finance committees) when insurers have excess cash holdings in the previous period.

DATA DESCRIPTION

This section provides data sources and defines variables used in this article.

Sample and Data Sources

Data for this study come from three sources. Insurers' firm-specific information is from annual statements filed with the National Association of Insurance Commissioners (NAIC). Insurers' financial strength ratings are from A.M. Best Key Rating Guide. A G-Index that measures the power sharing relation between shareholders and managers is constructed using 24 provisions related to shareholder rights and takeover defenses found in insurers' proxy statements. Finally, we manually collect data from corporate proxy statements in the U.S. EDGAR database information on board structure, finance committee composition, institutional ownership, and corporate antitakeover provisions for the property and casualty insurers.

We examine insurers' cash holdings for the period from 1997 to 2002 and begin with all insurers in the 2002 property-casualty database of the NAIC. We eliminate insurers that report negative cash or negative invested assets. We also eliminate insurers that are not rated by A.M. Best and do not have complete data for calculating volatility of cash flow. (7) Sample insurers must be domiciled within the United States and be organized as stock companies. The final sample represents 1,454 insurer-years for years 1997 through 2002. (8)

Dependent Variable and Independent Variables

Dependent Variable. Our dependent variable, cash holdings, is measured as the ratio of cash and short-term investments to total invested assets. (9) We deflate cash holdings by total invested assets with the view that a firm's ability to generate future profits is a function of its assets in place. (10) We first examine the association between governance variables and the level of cash holding. The results will be reported in Table 4.

Excess (unexplained) cash is defined as the residual from a regression of cash holdings with a number of control variables used by Colquitt, Sommer, and Godwin (1999) and year dummy variables. We further decompose excess cash into positive excess cash and negative excess cash to test all main governance variables of interest. We then examine if a higher level of cash holdings is used to prevent underinvestment problems or is due to managers' discretionary expense-preference behaviors (Mester, 1989). The results will be reported in Tables 5 and 6.

Independent Variables

Main governance variables of interest. The major independent variables related to board characteristics are discussed next. Consistent with prior research, we first categorize directors as insiders, affiliated with the firm, or outsiders (see Weisbach, 1988; Byrd and Hickman, 1992; Brickley, Coles, and Terry, 1994). Insiders are current employees of the company. Affiliated directors could be past employees, relatives of the CEO, directors who have significant transactions and/or business relationships with the firm as defined by Items 404 (a) and (b) of Regulation S-X, or ones on interlocking boards as defined by Item 402 (j)(3)(ii) of Regulation S-X. Outsiders have no ties to the firm beyond position as a board member. We use the fraction of outside directors, BD Outside Director%, defined as the number of outside directors (excluding affiliated directors) divided by board size, to measure board independence. The fraction of outside directors, FC_Outside Director%, is defined as the number of outside directors (excluding affiliate directors) divided by finance committee size to measure finance committee independence.

Other governance control variables. This section discusses the association between cash holdings and other corporate governance variables.

Board size. Jensen (1993) suggests that the agency costs increase with the board size because large boards are less effective in monitoring and easier for the CEO to control. Research on group decision making has also shown that large groups may weaken motivation and participation (Herold, 1979) and present problems with coordination and organization (Hackman and Morris, 1975). Yermack (1996) finds that a smaller board is related to better firm performance. Based on the above literature, we argue insurers with large boards have more cash holdings because of ineffective monitoring.

On the other hand, the literature shows that board-monitoring effectiveness increases with board size instead because large board can commit more time and effort to oversee management, bring in a variety of expertise to handle complexity of the firm's operations (Chaganti, Mahajan, and Sharma, 1985; Lipton and Lorsch, 1992; Monks and Minow, 1995; Klein, 2002). Adams and Mehran (2002) find that some banking firms with larger boards do not underperform their peers. Anderson, Mansi, and Reeb (2004) document that larger boards of directors are more effective monitors of the financial accounting process. In addition, a recent study by Huang et al. (2011) finds that board size is positively related to efficiency in the property-liability insurance industry.

Finally, according to their shareholder power hypothesis, Harford, Mansi, and Maxwell (2008) suggest that firms whose shareholders are better represented by larger boards will allow managers to have more cash holdings to prevent the underinvestment problem. Based on the above arguments, we suggest that insurers with large boards have large cash holdings because large boards likely consist of members with financial expertise who can better understand the needs of holding large cash reserves to avoid underinvestment problems. Furthermore, managers of those insurers would not misuse large cash holdings because larger boards are effective in monitoring managers as mentioned above.

Although the above arguments regarding the association between board size and cash holdings are different, those different streams of arguments consistently predict that insurers with a larger board will have a higher level of cash holdings. Board Size, or the number of directors serving on the board, is used to measure board-monitoring power.

Additional board directorships. There are two competing views about the association between additional directorships held by directors and monitoring effectiveness. On the one hand, the greater the number of outside directorships held by board members, the more the board of directors invests in accumulating governance expertise; therefore, directors are more likely to engage in effective monitoring activities (Fama and Jensen, 1983). Thus, this argument predicts a negative association between the level of cash holdings and additional directorships.

On the other hand, too many outside directorships can spread the attention of directors over too many companies and make directors too busy to be effective in monitoring agency costs. Thus, it can be expected that there is a positive association between the level of cash holdings and additional directorship. Given the two competing arguments and absence of theories with respect to the relative strength of these two arguments, we believe that there is an association between additional directorship and the level of cash holdings, but do not predict the sign. BD_Avg. # of Directorships, or the average number of directorship positions board members held in other public companies, is used to measure monitoring expertise/ overextension.

Board of directors incentives. We use stock ownership of directors to proxy for directors' incentive. Morck, Shleifer, and Vishny (1988) find that the impact of board director ownership on firm value may be nonlinear due to a trade-off between agency costs related to the self-interests of director ownership and convergence of interest of director ownership. Following the argument of Morck, Shleifer, and Vishny (1988) and the agency cost perspective, we posit that when director ownership is small, agency cost effect dominates convergence-of-interest effect. Directors would allow managers to hold large cash holdings for directors' self-interests. The association between board of director ownership and cash holding is expected to be positive when director ownership is small.

On the other hand, when director ownership is high, the convergence-of-interest effect dominates because the interests of directors align with stockholders. To reduce agency costs, directors would not allow managers to hold larger cash holdings. We argue that the association between board of director ownership and cash holding is negative when director ownership is high.

Given the aforementioned arguments, a relation between board stock ownership and the level of cash holdings is nonlinear. We use the percentage of shares held by directors (i.e., Board Director Ownership) to proxy for the strength of outside directors' monitoring incentives.

Board of directors diligence. We next develop an association between the diligence of boards of directors and the effectiveness of the board. Vafeas (1999) suggests that board meeting frequency could be a proxy for the time and commitment to monitor management when boards meet more frequently after crisis. We expect boards that devote more time and energy to monitor management are more likely to do a better job in constraining managers from stockpile cash holdings to reduce agency costs. The number of board of directors' meetings, BD_# of Meetings, is used as a proxy for board diligence.

Financial experts. Finance committees are typically made up of members with financial expertise. Because of their understanding of more complex financial instruments and investment opportunities, finance committee members with finance expertise should be able to effectively advise the management about the optimal level of cash holdings. When the percentage of financial experts on the finance committee is small, cash holdings may increase as the percentage of financial experts increases to avoid underinvestment problems because financial experts are more likely to understand the underinvestment problem.

On the other hand, cash holdings decrease as the percentage of financial experts increases when the percentage of financial experts is large. The reason is that certain finance committee directors may tend to contribute less and free ride on other finance committee directors' monitoring. Given the aforementioned arguments, the relation between financial committee expertise and the level of cash holdings is nonlinear. We use the percentage of financial experts on the board as the variable for financial expertise.

Financial committee incentives. Next, we develop a hypothesis related to finance committee incentives. For large U.S. companies, large nonmanagement shareholders or their representatives with a seat on the board or the finance committee have greater incentive to take an active and interventionist role in the firm's financial affairs and investments. Shleifer and Vishny (1986) suggest that blockholders reduce the free rider problem, perform a monitoring function, and reduce the opportunism of managerial discretion. Jensen (1993) suggests that blockholders who have significant equity stakes in a firm are important to a well-functioning governance system because their financial interests give them the incentive to monitor management. While these arguments concern mainly blockholders on a board, we believe they can be applied to the finance committee. Thus, we propose that firms with at least one blockholder on their finance committees would hold small cash holdings to reduce agency costs.

On the other hand, blockholders can exercise undue influence over management to promote their self-interests. Shleifer and Vishny (1997) and Harford (1999) suggest that directors with equity would stockpile cash when firms generate excess cash flow so that the firm could defend itself against unwanted takeover attempts. In addition, blockholders on finance committees may want to hold high levels of cash to avoid underinvestment problems. Based on these arguments, it is expected that firms with at least one blockholder on their finance committees would have larger cash holdings.

As there are two streams of competing and relevant literature, an association between finance committees with at least one large blockholder and the level of cash holdings is expected, but the sign is not predicted. FC_5% Blockholder is a measure of finance committee incentives. It is defined as 1 if a firm has at least one outside blockholder on the finance committee and 0 otherwise, where blockholder is defined as 5 percent ownership.

CEO/chairman duality. Boards may be headed by CEOs. It is likely that duality may hinder the monitoring efficacy. In particular, a CEO can control the information made available to other board members and thus may impede effective monitoring (Jensen, 1993). For example, one important function of the chairman is to compensate the CEO according to his/her value-creating performance to shareholders. Clearly, a chairman cannot perform this function apart from his/her personal interests such as job security and freedom from capital market discipline if the chairman is also the CEO. Furthermore, a CEO/chairman would presumably have great power and a broad control base (Hambrick and Finkelstein, 1987; Patton and Baker, 1987) and he/she may stockpile excess cash holdings to serve his/her personal interest at the expense of shareholders' interest. Therefore, for the board to be effective, Jensen (2010) argues that it is important to separate CEO and chairman positions. Alternatively, the "duality" of this leadership structure can effectively consolidate the management decisions and control processes. In other words, when the positions are held by the same person, a lower likelihood of power struggles can accelerate decision process (Harrison, Torres, and Kukalis, 1988). To avoid underinvestment by retaining large cash holdings, CEO/chairman may hold large cash holdings to avoid the costs of raising capital from the capital market. Given the two competing arguments and absence of theories with respect to the relative strength of these two arguments, we believe that there is an association between CEO/chairman duality and the level of cash holdings, but do not predict the sign. A CEO/chair Duality (BD_CEO/Chair Duality) dummy variable is defined as 1 if the CEO is also the chair of the board and 0 otherwise.

Institutional investors. Large stockholdings of institutional investors may motivate them to better perform monitoring activities as their voting power allows them to significantly influence the management (Shleifer and Vishny, 1986). On the other hand, institutional investors may influence managers just to secure institutional investors' private benefits at the expense of other shareholders. Shleifer and Vishny (1997) find that large shareholders actually act to promote their self-interests, and the firm suffers from the loss of managerial initiative (Burkart, Gromb, and Panunzi, 1997). We use the percentage of shares outstanding owned by institutional investors (Institutional Ownership) as a control variable but do not have an a priori prediction on the coefficient of this variable.

G-Index. Dittmar, Mahrt-Smith, and Servaes (2003), Pinkowitz, Stulz, and Williamson (2004), and Kalcheva and Lins (2007) show that firms with weaker shareholder rights have higher cash holdings. Conversely, Harford, Mansi, and Maxwell (2008) find that firms with weaker shareholder rights (high G-Index) have lower cash reserves because poorly controlled managers may elect to spend cash quickly on acquisitions and capital expenditures. We control the strength of shareholder rights by using a G-Index of 24 governance provisions (Gompers, Ishii, and Metrick, 2003), but do not predict the sign of G-Index.

Other Control Variables. Firm-specific variables motivated mainly by Colquitt, Sommer, and Godwin (1999) are also included to control for firm-specific characteristics that may affect the cross-sectional mean level of cash holdings. The discussions of the association between firm characteristics and cash holdings are brief to save space. Interested readers should refer to Colquitt, Sommer, and Godwin. The proxy for insurer size is the natural log of the insurer's total assets, and the expected sign on Size is negative. We use insurer ratings from A.M. Best Company as a proxy for financial strength, and the expected sign on Financial Strength is negative.

Volatility of Cash Flows, measured as the standard deviation of net cash flow from operation over the previous 5 years, is included in the model. The association between Volatility of Cash Flows and cash holdings is positive. Insurers need more cash reserves when their cash flows are more volatile. Duration of Liabilities is included as a proxy for the length of the claim. The timing of payouts of the lines of business an insurer underwrites clearly affects an insurer's need for cash. Insurers with short tail lines require a high liquidity level. Insurers with long tail lines of insurance can keep premiums for a longer period of time and pay out claims over many years. Our proxy for Duration of Liabilities is measured as estimated average duration of liabilities for the insurer, and the expected sign of the coefficient on Duration of Liabilities is negative. Our proxy for liability duration is based on the studies of Babbel and Klock (1994) and Cummins and Weiss (1991). Leverage is defined as total liabilities divided by total assets, and the expected sign on Leverage is not predicted. Market to book (M/B) ratio, a proxy for growth opportunities, is the ratio of the market value to book value of insurers' assets. A positive association between M/B Ratio and cash holdings is expected.

Non-Invested Assets is defined as the insurer's total noninvested assets divided by its total assets. A positive association between Non-Invested Assets and cash holdings is suggested by Colquitt, Sommer, and Godwin (1999). We measure the variable Common Stock as the ratio of insurer's common stock to its total invested assets and expect this variable to be negatively associated with the level of cash holdings. Firm Age is defined as natural logarithm of the difference between firms' ages and 5 to control for firm's ability to generate cash. In particular, we expect that older firms can generate more cash. We also include a dividend dummy (Dividend) in our models. Dividend is 1 if the insurer pays dividend and 0 otherwise. Insurers, which pay dividends, will have less cash; thus, the sign is expected to be negative. Harford, Mansi, and Maxwell (2008) suggest that firms select to pay back cash to shareholders through increases in dividends as a signal of good governance, holding low cash reserves. Separate yearly intercepts are included to control for time period conditions causing differences in cross-sectional mean level of cash holdings. Table 1 provides definitions and/or predictions for all the variables.

Descriptive Statistics

Table 2 reports descriptive statistics for the sample. Our dependent variable (Cash Holdings) has a mean of 8.9 percent and a median of 4.7 percent with a standard deviation of 11.8 percent, respectively. The median of 4.7 percent is quite close to the result (4.9 percent) in Harford, Mansi, and Maxwell (2008) for the nonfinancial industry. BD_Outside Director% has a mean of 60.4 percent and a median of 62.5 percent, suggesting that most insurers have boards with a majority of outside directors. FC_Outside Director% has a mean of 56.9 percent and a median of 63.2 percent, meaning that most insurers have a majority of outside directors on their finance committee. Board Size has a mean of 10.9 and a median of 9 with minimum (maximum) of 5 (19). (11)

BD_Avg. # of Directorships has a mean of 1.9 and a median of 2.2. These two values are slightly higher than those (1.44 and 1.41, respectively) found by Ahmed and Duellman (2007) for nonfinancial industries. BD_CEO/Chair Duality has a mean of 69.3 percent and a median of 1, meaning that for most insurers their CEOs are also board chairs.

Board Director Ownership has a mean of 15 percent and a median of 8.8 percent. Log (BD_# of Meeting) has a mean of 1.6 and a median of 1.6 with minimum (maximum) of 0 (2.8). FC_ _% Expert has a mean of 63 percent and a median of 60 percent. FC_5% Blockholder has a mean of 13.2 percent and a median of 0 percent. Our mean and median of G-Index are lower than other articles. The result may reflect the fact that the insurance industry is highly regulated.

Among firm-specific variables, the mean of M/B Ratio is 1.8. Dividend has a mean of 45.5 percent, suggesting that almost half of insurers pay a cash dividend during our sample period. Finally, the means of Leverage (57.4 percent) and Common Stock (12.4 percent) are similar to the findings of Colquitt, Sommer, and Godwin (1999).

Table 3 presents the results of Pearson correlation matrix for all variables of interest. One major finding is that BDOutside Director% is highly correlated with FC_Outside Director% and significant at the 1 percent level. Specifically, the correlation coefficient between two variables is 80.3 percent.

RESULTS

Our data are unbalanced panel data, and thus, Hausman tests are used to test whether random effects or fixed effects should be used. All the results of the Hausman tests suggest fixed effects models should be used. We estimate firm-level fixed effects to test the association between an insurer's cash holdings and corporate governance variables while controlling for their operational characteristics, Institutional Ownership, and the G-Index. The model is stated below.

Cash [Holdings.sub.i,t] = f(Cash [Holdings.sub.i,t-1], Board_Outside [Director%.sub.i,t]), FC_Outside [Director%.sub.i,t], Governance Control [variables.sub.i,t], Firm Characteristics [variables.sub.i,t], Year Dummies, Firm Fixed Effects). (1)

Our primary interest is the coefficient estimates on independent board and finance committee variables. The f-values are corrected for heteroskedasticity. We follow Harford, Mansi, and Maxwell (2008) and use the lagged value of the insurer's cash holdings to control for endogeneity. In Model 1 of Table 4, the coefficient on BD_Outside Director% is positive and significant at the 1 percent level, indicating that greater board independence is associated with higher levels of cash holdings. The result is consistent with the proposition suggested by Harford, Mansi, and Maxwell (2008). They propose the shareholder power hypothesis, which suggests that firms whose shareholders have better control of managers will allow managers to have more cash holdings to prevent underinvestment problems.

In Model 1, the coefficient on FC_Outside Director % is not significant. We suspect the insignificant result is due to the fact that FC_Outside Director% and BD_Outside Director% are highly correlated. Table 3 shows the correlation is 80.3 percent. To avoid the multicollinearity problem between BDOutside Director% and FC_Outside Director%, we drop FC_Outside Director% in Model 2. The results of Model 2 indicate that the coefficient on BD Outside Director% remains positive and is significant at the 1 percent level. We drop BD Outside Director% in Model 3. The coefficient on FC_Outside Director% in Model 3 becomes significantly positive, with a f-statistic of 3.15, indicating that greater finance committee independence is associated with higher level of cash holdings.

Among other corporate governance variables, the coefficient on Board Size is significantly positive in all three models and indicates that larger board size is associated with higher levels of cash holdings. The results support the argument that larger boards usually do not criticize the policies of top managers and/or are less efficient in restraining managers holding cash reserves (Lipton and Lorsch, 1992; Jensen, 1993). Oversized boards are less likely to function effectively and are not able to constrain managers from accumulating excess cash reserves. However, the result also supports the argument that a large board is more likely to have members who understand the needs of large cash holdings to avoid underinvestment problems. The coefficient BD Avg. # of Directorships is negative and significant at the 5 percent level, indicating that board members with more directorship experience tend to favor managers holding lower level of cash reserves.

The coefficients on Board Director Ownership and Board Director Ownership (2) are significant in all three models in Table 4. Specifically, the coefficients on Board Director Ownership and Board Director Ownership (2) are positive and negative, respectively, supporting our hypothesis that there is a nonlinear relation between the level of insurers' cash holdings and stock ownership by directors. It indicates that when director ownership is small, directors allow for large cash holdings because agency costs of directors dominate. When director ownership is high, directors constrain managers from stockpiling large cash reserves because the interests of directors align with stockholders. The coefficient on Log(BD_# of Meeting) is significantly negative in all three models, indicating that a diligent board leads to lower cash holdings.

We also find that the coefficients of FC_% Expert and FC_% Expert (2) are positive and negative, respectively. These results suggest that when the percentage of financial experts on the finance committee is small, the cash holdings increase with higher percentage of financial experts to avoid underinvestment. When the percentage of financial experts is high, the cash holdings decrease with higher percentage of financial experts because of a free rider problem. CEO/Chair Duality is also significantly and positively associated with cash holdings in all three models, indicating that a CEO/chairman tends to accumulate a higher level of cash holdings.

With respect to firm-specific control variables, the coefficients on Financial Strength, Duration of Liabilities, and Common Stock are negative and significant in all three models. The results imply insurers with better financial strength, longer duration of liability, and more common stocks have less cash holdings. These results are consistent with the predictions and findings of Colquitt, Sommer, and Godwin (1999). For other firm-specific control variables, signs are all expected but not significant.

We also run an additional regression using the specification of Colquitt, Sommer, and Godwin (1999) for robustness. Our empirical results show that the firm-specific variables, which are significant in our models, are consistent with those of Colquitt, Sommer, and Godwin (1999). Some of the variables statistically significant in Colquitt, Sommer, and Godwin are not statistically significant in ours. (12) One possible explanation is that we add corporate governance variables to our regression models.

FUTURE GROWTH, EXPENSES, AND INDEPENDENT GOVERNANCE

In this section, we examine whether a higher level of cash holdings is used to prevent underinvestment problems (Harford, Mansi, and Maxwell, 2008) or is due to managers' discretionary expense-preference behaviors (Mester, 1989).

We use two proxies for future growth: ratio of the market value to book value of insurer's assets, and the average growth rate of net premium written for the following 3 years. Market value to book value is commonly used to proxy Tobin's Q, which reflects growth opportunities. Also, average premium growth rate measures insurers' ex post actual growth. We use positive excess cash (Excess [Cash.sub.i,t-1] (positive)) interacting with main governance variables term to further test whether a higher level of cash holdings is used to prevent underinvestment problems. Excess (unexplained) cash is defined as the residual from a regression of cash holdings with a number of control variables used by Colquitt, Sommer, and Godwin (1999) and year dummy variables. We also add the independent variables used in Table 4, including lagged excess cash holdings (Excess [Cash.sub.t-1]) and growth-specific control variables. Main governance variables of interest are also lagged for one period.

Panel A of Table 5 reports the empirical results for Market-to-Book Ratio. Model 1 shows the coefficients of BD Outside [Director%.sub.i,t-1] x Excess [Cash.sub.i,t - 1(positive)] and FC_Outside [Director%.sub.i,t-1] x Excess [Cash.sub.i,t-1(positive)] are positive but insignificant. To avoid multicollinearity problem, we drop interaction terms of FC_Outside [Director%.sub.i,t-1] (BD_Outside [Director%.sub.i,t-1]) in Model 2 (Model 3) and rerun the regressions. Model 2 shows the interaction term BD Outside [Director%.sub.i,t-1] x Excess [Cash.sub.i,t-1 (positive)] is positive and significant. Model 3 shows the coefficient of FC_Outside [Director%.sub.i,t-1] x Excess [Cash.sub.i,t-1(positive)] is also positive and significant.

In addition, we use ex post premium growth rate as dependent variable and the empirical results are reported in Panel B of Table 5. The coefficients of BD Outside [Director%.sub.i,t-1] x Excess [Cash.sub.i,t-1(positive)] and FC_Outside [Director%.sub.i,t-1] x Excess [Cash.sub.i,t-1(positive)] are positive and significant in Models 2 and 3, respectively. The overall results suggest that positive excess cash holdings lead to actual future growth when the proportion of outside directors on the board or on financial committee is high. The evidence is consistent with the independent director responsibility hypothesis. (13)

We next examine the influence of independence governance variables on managers' discretion on expenditures. Mayers, Shivdasani, and Smith (1997) find that effective monitoring could reduce insurers' expenditures. We follow their study and let lagged positive excess cash (Excess [Cash.sub.i,t-1(positive)]) interact with the lagged proportion of outside directors on boards (BD_Outside [Director%.sub.t-1]) and the lagged proportion of outside directors on the finance committee (FC_Outside [Director%.sub.t-1]), respectively. If monitoring is effective, the expected sign of these interaction terms is not positive.

We use salary expense as a dependent variable (Salary Expense Ratio) because it is a proxy of agency costs (including perquisite consumption) of managers. Salary Expense Ratio is defined as the salary expenditure to net premiums written.

We also incorporate expense-specific control variables in the model. Commercial Line% is measured as the percentage of direct premiums written from commercial lines. Since commercial lines are more complex than personal lines, insurers underwriting more commercial lines need more skillful employees. Insurers are likely to compensate these skillful employees with higher salary (Mayers and Smith, 1992). Therefore, commercial lines may be positively associated with higher salary expense. On the other hand, Regan (1997) points out that to underwrite complex lines of insurance, insurers may receive information from outside experts. For example, insurers may rely on reports by agents or safety engineers to evaluate the adequacy of the insured's loss control programs. In such cases, insurers may spend lower salary expense, since they outsource some of underwriting and loss control tasks. Independent System, a proxy for the distribution systems used by insurers, is measured as 1 if the insurer uses independent agents and 0 otherwise. We expect the relation between Independent System and Salary Expense Ratio to be positive because the independent system involves a more complex line and costly claims settlement process. Kim, Mayers, and Smith (1996) argue that different distributors provide different levels of service, so costs should also be different.

Doing business in multiple lines of insurance and across a wider geographic area would require more expenses in setting rates and underwriting. Line_HHI, a proxy for operating complexities, is measured as the sum of the squared percentage of direct premiums written in each of the 26 lines of insurance. State_HHI, a proxy for geographic complexities, is measured as the sum of squared percentage of direct premiums written by state. Higher values of the indices indicate higher line-of-business concentration and geographic concentration. We expect the expenses and concentration are negatively related because highly concentrated insurers do not need to deal with a lot of operating and geographic complexities. Finally, we incorporate the natural log of direct premiums written (Size_DPW) in the model to control the size effect on expenses. The coefficient of Size_DPW is predicted to be negative because of economies of scale.

Table 6, Model 1 shows the coefficients of interaction variable (BD_Outside [Director%.sub.t-1] x Excess [Cash.sub.t-1(positive)]) is positive but insignificant, and interaction variable (FC_Outside [Director%.sub.t-1] x Excess [Cash.sub.t-1(positive)]) to be significantly negative. To avoid multicollinearity problem, we also rerun regression and report results in Models 2 and 3. The interaction variables are negative and insignificant. The evidence does not support that insurers with a higher proportion of outside directors on the board or in the financial committee 'spend positive excess cash on salary expenses. The evidence overall is consistent with the independent director responsibility hypothesis.

We next report some important findings related to control variables. The coefficients for Institutional Ownership are negative and significant in Models 1 and 3, indicating that insurers with higher institutional ownership are associated with lower salaries. State_HHI is negatively related to salary expense ratio, implying the insurers focusing on small number of states are able to lower salary expenses. Size is negatively significant, suggesting larger firms pay lower salary ratio, implying economies of scale.

ADDITIONAL TESTS

We also examine the relation between other corporate governance variables and growth opportunity and between other corporate governance variables and expenses. Our analyses of this section are similar to those of Tables 5 and 6. To save space, we briefly summarize the results and do not provide tabulated results. (14) The evidence shows that insurers with more diligent board and more financial expert on the financial committee are positively related to growth opportunities when insurers have positive excess cash. Equally important, we also find excess cash spent on salary expenses does not increase with more financial expert on the financial committee. Overall, the evidence suggests that diligent board and high percentage of financial experts in financial committees are associate with high growth opportunities and not high expenses, when insurers have positive cash holdings in the prior period. These results are consistent with the notion that some corporate governance is effective in preventing from underinvestment and monitoring expenses. We also find large board size and duality have similar relation with growth opportunities proxied by Tobin's Q and expenses. It should be noted that the relation between board size (duality) and growth opportunities proxied by past growth rates is not statistically significant. Likewise, insurers with higher institutional ownership are associated with higher premium growth when insurers have positive excess cash, and excess cash spent on salary expense does not significantly increase with higher institutional ownership.

CONCLUSIONS

We have examined the impact of the roles of the board and finance committee on insurers' cash holdings using a sample of 1,454 U.S. stock property-liability insurer-year observations from 1997 to 2002. We focus on the roles of independent board members and independent finance committee members. We find insurers with a higher proportion of outsiders on their boards and finance committees have more cash holdings. Additional results show that the independence of boards and finance committees is positively related to the growth of an insurer when it has excess cash in the previous period. These two results imply that independent board and finance committee members allow managers to hold excess cash to avoid underinvestment problem resulting from insufficient cash holdings. This finding also supports the argument that once future growth opportunities are anticipated, an independent board or finance committee would choose to hold excess cash to avoid the higher costs of raising funds externally. Our results are consistent with the shareholder power hypothesis, proposed by Harford, Mansi, and Maxwell (2008) in the context of industrial firms. The shareholder power hypothesis suggests shareholders with more effective control of managers would allow managers to hold excess cash to avoid underinvestment problems. However, Harford, Mansi, and Maxwell do not find evidence supporting the shareholder power hypothesis when nonfinancial firms are used as a sample.

We also find that excess cash spent on operating expenditures such as salaries does not significantly increases with a higher proportion of outside directors on a board or a finance committee. The overall evidence above supports our proposed independent responsibility hypothesis.

Other important findings are provided below. The evidence for other corporate governance variables suggests that large board size, high percentage of financial experts in financial committees, and duality are associated with more cash holdings. Further examinations show that these three variables are positively related growth opportunities as proxied by Tobin's Q and negatively related expenses, when insurers have positive cash holdings in the prior period. In other words, high cash holdings under these scenarios are to be used for future investment opportunities.

In summary, our results suggest that other than board or committee director independence, some corporate governance characteristics can further contribute to strengthening board-monitoring performance.

DOI: 10.1111/jori.12049

APPENDIX

Firm-years in the 1997 through 2002 NAIC property-casualty      16,241
  database
Less:
  Firm-years of insurers not organized as stock companies        3,202
  Firms must have operated continuously for the period from      4,119
    1992 to 2006 to calculate volatility of cash flow and
    future growth rate of insurers
  Unrated firms                                                  2,765
  Other insurers with insufficient data to compute               4,673
    variables corporate governance
  Insurers reporting negative cash                                  28
  Final sample                                                   1,454


REFERENCES

Adams, R., and H. Mehran, 2002, Board Structure and Banking Firm Performance, Working Paper, Federal Reserve Bank of New York.

Ahmed, A., and S. Duellman, 2007, Accounting Conservatism and Board of Director Characteristics: An Empirical Analysis, Journal of Accounting and Economics, 43:411-437.

A.M. Best Company, 1997-2002, Best's Key Rating Guide: Property/Casualty (Oldwick, NJ: A.M. Best Company).

Anderson, R., S. Mansi, and D. Reeb, 2004, Board Characteristics, Accounting Report Integrity, and the Cost of Debt, Journal of Accounting and Economics, 37: 315-342.

Babbel, D., and D. Klock, 1994, Measuring the Interest Rate Risk of Property-Casualty Insurer Liabilities, in: S. G. Gustavson and S. E. Harrington, eds., Insurance, Risk Management and Public Policy (Boston, MA: Kluwer Academic Publishers).

Berton, L., 1995, Corporate Woes Put Board Audit Panels in the Spotlight, Wall Street Journal, April 7.

Brickley, J., J. Coles, and R. Terry, 1994, Outside Directors and the Adoption of Poison Pills, Journal of Financial Economics, 35: 371-390.

Burkart, M., D. Gromb, and F. Panunzi, 1997, Large Shareholders, Monitoring, and the Value of the Firm, Quarterly Journal of Economics, 112: 693-728.

Byrd, J., and K. Hickman, 1992, Do Outside Directors Monitor Managers? Evidence From Tender Offer Bids, Journal of Financial Economics, 32: 195-222.

Chaganti, R., V. Mahajan, and S. Sharma, 1985, Corporate Board Size, Composition, and Corporate Failures in Retailing Industry, Journal of Management Studies, 22:400-417.

Colquitt, L., D. Sommer, and N. Godwin, 1999, Determinants of Cash Holdings by Property-Liability Insurers, Journal of Risk and Insurance, 66: 401-415.

Cummins, J. D., and M. Weiss, 1991, The Structure, Conduct, and Regulation of the Property-Liability Insurance Industry, in: R. W. Kopke and R. E. Randall, eds., The

Financial Condition and Regulation of Insurance Companies (Boston, MA: Federal Reserve Bank of Boston).

Demsetz, H., and K. Lehn, 1985, The Structure of Corporate Ownership: Causes and Consequences, Journal of Political Economy, 93: 1155-1177.

Dittmar, A., and J. Mahrt-Smith, 2007, Corporate Governance and the Value of Cash Holdings, Journal of Financial Economics, 83: 599-634.

Dittmar, A., J. Mahrt-Smith, and H. Servaes, 2003, International Corporate Governance and Corporate Cash Holdings, Journal of Financial and Quantitative Analysis, 38: 111-133.

Fama, E., and M. Jensen, 1983, Agency Problems and Residual Claims, Journal of Law and Economics, 26: 327-349.

Gompers, P., J. Ishii, and A. Metrick, 2003, Corporate Governance and Equity Prices, Quarterly Journal of Economics, 118: 107-155.

Grace, M., R. Klein, and R. Phillips, 2005, Insurance Company Failures: Why Do They Cost So Much? Working Paper, Georgia State University.

Hackman, J., and C. Morris, 1975, Group Tasks, Group Interaction Process, and Group Performance Effectiveness: A Review and Proposed Integration, in: L. Berkowitz, ed., Advances in Experimental Social Psychology, Volume 8, (New York: Academic Press), pp. 45-99.

Hambrick, D., and S. Finkelstein, 1987, Managerial Discretion: A Bridge Between Polar Views of Organizational Outcomes, Research in Organizational Behavior, 9:369406.

Harford, J., 1999, Corporate Cash Reserves and Acquisitions, Journal of Finance, 54: 1969-1997.

Harford, J., S. Mansi, and W. Maxwell, 2008, Corporate Governance and Firm Cash Holdings in the US, Journal of Financial Economics, 87: 535-555.

Harrison, J., D. Torres, and S. Kukalis, 1988, The Changing of the Guard: Turnover and Structural Change in the Top-Management Positions, Administrative Science Quarterly, 33: 211-232.

Herold, D., 1979, The Effectiveness of Work Groups, in: S. Kerr, ed., Organizational Behavior (Columbus, OH: Grid Publishing), pp. 95-118.

Huang, L., G. Lai, M. McNamara, and J. Wang, 2011, Corporate Governance and Efficiency: Evidence From U.S. Property-Liability Insurance Industry, Journal and Risk and Insurance, 78: 519-550.

Jensen, M., 1986, Agency Costs of Free Cash Flow, Corporate Finance and Takeovers, American Economic Review, 76: 323-339.

Jensen, M., 1993, The Modern Industrial Revolution, Exit, and Failure of Internal Control Systems, Journal of Finance, 48: 831-880.

Jensen, M., 2010, The Modern Industrial Revolution, Exit, and Failure of Internal Control Systems, Journal of Applied Corporate Finance, 22: 43-58.

Kalcheva, I., and K. Lins, 2007, International Evidence on Cash Holdings and Expected Managerial Agency Problems, Review of Financial Studies, 20: 1087-1112.

Kim, W., D. Mayers, and C. Smith, 1996, On the Choice of Insurance Distribution Systems, Journal of Risk and Insurance, 63: 207-227.

Klein, A., 2002, Economic Determinants of Audit Committee Independence, The Accounting Review, 77: 435-452.

Lipton, M., and J. Lorsch, 1992, A Modest Proposal Improved Corporate Governance, Business Lawyer, 48: 59-77.

Mayers, D., A. Shivdasani, and C. Smith, 1997, Board Composition and Corporate Control: Evidence From the Insurance Industry, Journal of Business, 70: 33-62.

Mayers, D., and C. Smith Jr., 1992, Executive Compensation in the Life Insurance Industry, Journal of Business, 65(1): 51-74.

Mester, L., 1989, Testing for Expense Preference Behavior: Mutual versus Stock Savings and Loans, RAND Journal of Economics, 20: 483-498.

Monks, R., and N. Minow, 1995, Corporate Governance (New York, NY: Blackwell Publishing).

Morck, R., A. Shleifer, and R. Vishny, 1988, Management Ownership and Market Valuation: An Empirical Analysis, Journal of Financial Economics, 22: 293-315.

Myers, S., and N. Majluf, 1984, Corporate Financing and Investment Decisions When Firms Have Information That Investors Do Not Have, Journal of Financial Economics, 13: 187-221.

Opler, T., L. Pinkowitz, R. Stulz, and S. Williamson, 1999, The Determinants and Implications of Corporate Cash Holdings, Journal of Financial Economics, 52: 3-46.

Patton, A., and J. Baker, 1987, Why Won't Directors Rock the Boat? Harvard Business Review, 65: 10-12,16,18.

Pinkowitz, L., 1998, The Market for Corporate Control and Corporate Cash Holdings, Working Paper, The Ohio State University.

Pinkowitz, L., R. Stulz, and S. Williamson, 2004, Do Firms With Poor Protection Investor Rights Hold More Cash? Working Paper, Georgetown University.

Regan, L., 1997, Vertical Integration in the Property-Liability Insurance Industry: A Transaction Cost Approach, Journal of Risk and Insurance, 64: 41-62.

Securities and Exchange Commission (SEC), 2009, Final Rule: Proxy Disclosure Enhancements, Release Nos. 33-9089, December 23.

Shleifer, A., and R. Vishny, 1986, Large Shareholders and Corporate Control, Journal of Political Economy, 94: 461-488.

Shleifer, A., and R. Vishny, 1997, A Survey of Corporate Governance, Journal of Finance, 52: 737-783.

Standard & Poor's, 2002, Standard & Poor's Corporate Governance Scores: Criteria, Methodology and Definitions, (New York, NY: McGraw-Hill Companies, Inc.).

Stulz, R., 1990, Managerial Discretion and Optimal Financing Policies, Journal of Financial Economics, 26: 3-27.

Vafeas, N., 1999, Board Meeting Frequency and Firm Performance, Journal of Financial Economics, 53: 113-142.

Weisbach, M., 1988, Outside Directors and CEO Turnover, Journal of Financial Economics, 20: 431-460.

Wooldridge, J. M., 2006, Introductory Econometrics: A Modern Approach (Mason, OH: Thomson Higher Education).

Yermack, D., 1996, Higher Market Valuation of Companies With a Small Board of Directors, Journal of Financial Economics, 40: 185-211.

(1) Detailed reviews of the relevant literature are provided throughout the article.

(2) We compare cash to net premium written ratio for insurance companies with cash to sales ratio for noninsurance companies. Our results show that cash to net premium ratio is about 50 percent for insurance companies while cash to sales ratio for noninsurance companies is only about 18 percent (Harford, Mansi, and Maxwell, 2008).

(3) The actual sample period used in this study is from 1992 to 2006 because we need to calculate standard deviation of cash flows and future actual growth rate of direct premium written.

(4) The reasons provided are based on Opler et al. (1999).

(5) Please see AIG proxy statement.

(6) The investment opportunities can be to expand geographically or into a new product line. For example, Arnica Mutual Insurance expanded their business from mainly the East Coast to the state of Oregon and the state of Washington. Excess cash holdings can also be used to acquire another insurance company. For example, Liberty Mutual bought Safeco Insurance for $6.5 billion in 2008.

(7) We use firms that have been in existence for many years because we need to calculate standard deviations of cash flows to estimate firm risk. This restriction is likely to result in larger and more successful insurers in our sample and introduce a survivorship bias.

(8) In examining the association between excess cash and board and finance committee independence, we use the lagged variables. Therefore, in Tables 5 and 6, the sample period is from 1998 to 2003.

(9) This definition has been used in the insurance literature. Even though bonds and stocks insurance companies are highly liquid and can be considered marketable securities, insurance companies, in general, do not hold bonds and stocks for the liquidity purpose. Our study does not treat bonds and stocks as part of liquidity purposes.

(10) We deleted firms with negative cash holdings. The results are similar when insurers with negative cash holdings are included in the model.

(11) Mayers, Shivdasani, and Smith (1997) use stock life insurance companies in 1985 as their sample. Their sample mean and median of outside directors are 44 and 44 percent, respectively. Our sample consists of stock property-liability insurance companies; thus, our sample mean is different from that of Mayers, Shivdasani, and Smith.

(12) For comparison, the adjusted R2 using only the nongovernance variables is 0.1173, compared to 0.2224 in Colquitt, Sommer, and Godwin (1999).

(13) The noninteraction term FC_Outside Director% can be tricky to interpret in the presence of interaction terms in the regression. An interesting and meaningful way to measure its effect on dependent variable is to measure at its mean. Following Wooldridge (2006), we estimate the following equation.

Y = [a.sub.0] + [a.sub.1] Excess [Cash.sub.t-1] + [a.sub.2] FC-Outside [Director%.sub.t-1] + [a.sub.3] (Excess [Cash.sub.t-1-[mu]1]) (FC_Outside [Director%.sub.t-1-[mu]2]) + other variables + error.

Other variables are similar to those in Table 5 and [micro]s are their respective means. We are interested in [a.sub.2]. The results are not tabulated and show [a.sub.2] is negatively and significantly associated with the growth of direct premium written. A possible reason is that excessive growth of direct premium written may lead to an increase of insolvency. For example, insurers with excessive growth of premiums may be more likely to fail IRIS test. A cautious Financial Committee may not in favor of excessive growth premiums.

(14) The authors would be happy to provide the results upon request.

Wen-Yen Hsu is an Associate Professor in the Department of Risk Management and Insurance, Feng Chia University, Taiwan. Hsu can be contacted via e-mail: wyhsu@fcu.edu.tw. Yenyu (Rebecca) Huang is an Associate Professor in the Department of Tourism and Leisure Management, St. John's University Taiwan. Huang can be contacted via e-mail: vision@mail. sju.edu.tw. Gene Lai (corresponding author) is a Safeco Distinguished Professor of Insurance in the Department of Finance and Management Science, Washington State University. Lai can be contacted via e-mail: genelai@wsu.edu. Both Gene Lai and Wen-Yen Hsu are research fellows in Risk and Insurance Research Center, College of Commerce, National Chengchi University, Taiwan. The authors would like to thank the two anonymous referees for their helpful suggestions and comments on earlier drafts of the manuscript.

Table 1
Variable Definitions

                           Predicted
Variables                  Sign        H #     Definitions/Descriptions

Cash Holdings                                  Ratio of insurer's cash
                                               plus short-term
                                               investments to its total
                                               invested assets

BD_Outside Director%       +/+/-       1/2/3   % of outside directors
                                               (excluding affiliate
                                               directors) on the board

FC_Outside Director%       +/+/-       1/2/3   % of outside directors
                                               on finance committee

Board Size                 +                   Number of the board
                                               directors

BD_Avg. # of               ?                   Average number of
  Directorships                                directorship positions
                                               that board directors
                                               hold in other firms'
                                               boards

Board Director Ownership   +,-                 % of shares held by
                                               directors

Log(BD_# of Meetings)      -                   Natural log of number of
                                               board meetings

FC_% Expert                +,-                 % of directors with
                                               investment and financial
                                               matters oversight
                                               responsibility on the
                                               finance committee,
                                               including chief
                                               executive officer; chief
                                               financial officer; vice
                                               president of finance;
                                               executives in the
                                               investment, investment
                                               banking, or commercial
                                               banking industry; or
                                               financial consultants,
                                               all categorized as
                                               having corporate
                                               financial management
                                               experience

FC_5% Blockholder          ?                   1 if the firm has at
                                               least one outside
                                               blockholder on the
                                               finance committee, where
                                               block is defined at the
                                               5% ownership level, 0
                                               otherwise

BD_CEO/Chair Duality       ?                   1 if the CEO is also the
                                               chair of the board, 0
                                               otherwise

Institutional Ownership    ?                   % of shares held by
                                               institutions

G-Index                    ?                   An index of shareholder
                                               rights developed by
                                               Gompers, Ishii, and
                                               Metrick (2003)

Size                       -                   Natural log of total
                                               assets

Financial Strength         -                   Best's rating (0 = below
                                               B-; 1 = B, B-, 2 = B++,
                                               B+; 3 = A, A-; 4 = A++,
                                               A+)

Volatility of Cash Flows   +                   The standard deviation
                                               of net cash flow from
                                               operation over the
                                               previous 5 years

Duration of Liabilities    -                   Estimated average
                                               duration of liabilities
                                               for the insurer

Leverage                   ?                   Ratio of the insurer's
                                               total liabilities to
                                               total assets

M/B Ratio                  +                   Ratio of the market
                                               value to book value of
                                               insurers' assets

Non-In vested Assets       +                   Ratio of the insurer's
                                               total noninvested assets
                                               to its total assets

Common Stock               -                   Ratio of the insurer's
                                               common stock holdings to
                                               its total invested
                                               assets

Firm Age                   +                   Natural log of the
                                               difference between
                                               firms' ages and 5

Dividend                   -                   1 if the insurer pays a
                                               cash dividend to
                                               stockholders in the
                                               year, 0 otherwise

Table 2
Summary Statistics

Variables                      Mean      Median    Std. Dev.

Cash Holdings                  0.089      0.047        0.118
BD Outside Director%           0.604      0.625        0.218
FC_Outside Director%           0.569      0.632        0.249
Board Size                    10.861      9.000        3.632
BD Avg. # of Directorships     1.893      2.188        1.037
BDCEO/Chair Duality            0.693      1.000        0.462
Board Director Ownership       0.150      0.088        0.158
Log(BD_# of Meetings)          1.554      1.609        0.436
FC_% Expert                    0.630      0.600        0.199
FC_5% Blockholder              0.132      0.000        0.339
Institutional Ownership        0.337      0.250        0.271
G-Index                        2.868      3.000        1.470
Size                          18.984     18.912        1.681
Financial Strength             3.400      3.000        0.640
Volatility of Cash Flows       0.106      0.052        0.188
Duration of Liabilities        2.281      2.294        0.618
Leverage                       0.574      0.641        0.202
M/B Ratio                      1.789      1.501        1.128
Non-Invested Assets            0.159      0.118        0.138
Common Stock                   0.124      0.037        0.178
Firm Age                       3.105      3.135        1.035
Dividend                       0.455      0.000        0.498

Variables                    Minimum    Maximum

Cash Holdings                  0.000      0.907
BD Outside Director%           0.000      0.938
FC_Outside Director%           0.000      1.000
Board Size                     5.000     19.000
BD Avg. # of Directorships     0.000      3.750
BDCEO/Chair Duality            0.000      1.000
Board Director Ownership       0.000      0.673
Log(BD_# of Meetings)          0.000      2.773
FC_% Expert                    0.000      1.000
FC_5% Blockholder              0.000      1.000
Institutional Ownership        0.000      0.926
G-Index                        0.000      8.000
Size                          15.361     24.389
Financial Strength             0.000      4.000
Volatility of Cash Flows       0.003      2.124
Duration of Liabilities        0.665      4.957
Leverage                       0.005      0.972
M/B Ratio                      0.346      6.241
Non-Invested Assets            0.000      0.926
Common Stock                   0.000      0.940
Firm Age                       0.000      5.333
Dividend                       0.000      1.000

Note: Sample of 1,454 U.S. stock property-liability insurers firm-
year observations during 1997-2002. See Table 1 for variable
definitions.

Table 3
Pearson Correlation Matrix

                    1           2           3           4

1. Cash Holdings    1.000
2. BD_Outside       -0.126      1.000
Director%           (<0.001)
3. FC_Outside       -0.071      0.803       1.000
Director%           (0.007)     (<0.001)
4. Board Size       -0.085      0.198       0.126       1.000
                    (0.001)     (<0.001)    (<0.001)
5. BD_Avg. # of     -0.034      -0.008      0.059       0.134
Directorships       (0.198)     (0.751)     (0.025)     (<0.001)
6. Board Director   0.181       -0.495      -0.482      -0.373
Ownership           (<0.001)    (<0.001)    (<0.001)    (<0.001)
7. Log(BD_# of      -0.179      0.399       0.250       0.119
Meeting)            (<0.001)    (<0.001)    (<0.001)    (<0.001)
8. FC_% Expert      0.071       0.021       0.043       0.043
                    (0.007)     (0.422)     (0.104)     (0.104)
9. FC_5%            0.087       -0.129      -0.098      -0.175
Blockholder         (0.001)     (<0.001)    (0.000)     (<0.001)
10. CEO/Chair       0.052       -0.143      0.025       0.305
Duality             (0.046)     (<0.001)    (0.341)     (<0.001)
11. Institutional   -0.066      0.202       0.214       -0.068
Ownership           (0.012)     (<0.001)    (<0.001)    (0.009)
12. G-Index         -0.047      -0.139      -0.216      0.188
                    (0.072)     (<0.001)    (<0.001)    (<0.001)
13. Size            -0.260      -0.078      -0.053      0.186
                    (<0.001)    (0.003)     (0.043)     (<0.001)
14. Financial       -0.127      -0.171      -0.124      0.050
Strength            (<0.001)    (<0.001)    (<0.001)    (0.055)
15. Volatility of   0.080       0.170       0.158       0.108
Cash Flows          (0.002)     (<0.001)    (<0.001)    (<0.001)
16. Duration of     -0.126      -0.094      -0.046      0.083
Liabilities         (<0.001)    (0.000)     (0.079)     (0.002)
17. Leverage        -0.217      0.199       0.124       0.079
                    (<0.001)    (<0.001)    (<0.001)    (0.003)
18. M/B Ratio       -0.129      -0.015      -0.069      0.460
                    (CO.001)    (0.565)     (0.008)     (<0.001)
19. Non-Invested    0.030       0.194       0.122       0.221
Assets              (0.248)     (<0.001)    (<0.001)    (<0.001)
20. Common          -0.100      -0.296      -0.272      -0.097
Stock               (0.000)     (<0.001)    (<0.001)    (0.000)
21. Firm Age        -0.126      -0.035      -0.006      0.138
                    (<0.001)    (0.177)     (0.818)     (<0.001)
22. Dividend        -0.204      0.007       -0.050      -0.061
                    (<0.001)    (0.782)     (0.059)     (0.020)

                    5           6           7           8

1. Cash Holdings
2. BD_Outside
Director%
3. FC_Outside
Director%
4. Board Size
5. BD_Avg. # of     1.000
Directorships
6. Board Director   -0.156      1.000
Ownership           (<0.001)
7. Log(BD_# of      -0.073      -0.319      1.000
Meeting)            (0.006)     (<0.001)
8. FC_% Expert      -0.155      -0.094      -0.185      1.000
                    (<0.001)    (0.000)     (<0.001)
9. FC_5%            -0.328      0.411       0.027       -0.132
Blockholder         (<0.001)    (<0.001)    (0.312)     (<0.001)
10. CEO/Chair       0.164       0.025       -0.299      0.122
Duality             (<0.001)    (0.331)     (<0.001)    (<0.001)
11. Institutional   0.079       -0.528      0.254       0.222
Ownership           (0.003)     (<0.001)    (<0.001)    (<0.001)
12. G-Index         0.015       -0.137      0.138       -0.339
                    (0.563)     (<0.001)    (<0.001)    (<0.001)
13. Size            0.199       -0.106      -0.015      -0.084
                    (<0.001)    (<0.001)    (0.574)     (0.001)
14. Financial       0.248       0.146       -0.133      -0.298
Strength            (<0.001)    (<0.001)    (<0.001)    (<0.001)
15. Volatility of   -0.005      -0.099      0.103       0.112
Cash Flows          (0.841)     (0.000)     (<0.001)    (<0.001)
16. Duration of     0.099       -0.147      -0.156      0.013
Liabilities         (0.000)     (<0.001)    (<0.001)    (0.630)
17. Leverage        0.052       -0.309      0.210       -0.051
                    (0.047)     (<0.001)    (<0.001)    (0.050)
18. M/B Ratio       0.274       -0.090      0.051       -0.207
                    (<0.001)    (0.001)     (0.051)     (<0.001)
19. Non-Invested    0.113       -0.098      0.093       -0.062
Assets              (<0.001)    (0.000)     (0.000)     (0.017)
20. Common          -0.005      0.280       -0.203      -0.029
Stock               (0.862)     (<0.001)    (<0.001)    (0.262)
21. Firm Age        -0.026      -0.137      -0.002      0.111
                    (0.329)     (<0.001)    (0.933)     (<0.001)
22. Dividend        -0.002      -0.019      0.133       -0.152
                    (0.944)     (0.474)     (<0.001)    (<0.001)

                    9           10          11          12

1. Cash Holdings
2. BD_Outside
Director%
3. FC_Outside
Director%
4. Board Size
5. BD_Avg. # of
Directorships
6. Board Director
Ownership
7. Log(BD_# of
Meeting)
8. FC_% Expert
9. FC_5%            1.000
Blockholder
10. CEO/Chair       -0.075      1.000
Duality             (0.004)
11. Institutional   -0.236      -0.214      1.000
Ownership           (<0.001)    (<0.001)
12. G-Index         -0.157      -0.182      -0.022      1.000
                    (<0.001)    (<0.001)    (0.408)
13. Size            -0.070      0.081       -0.031      0.152
                    (0.007)     (0.002)     (0.230)     (<0.001)
14. Financial       0.068       0.167       -0.258      0.115
Strength            (0.010)     (<0.001)    (<0.001)    (<0.001)
15. Volatility of   0.036       0.085       -0.016        -0.057
Cash Flows          (0.168)     (0.001)     (0.530)      (0.029)
16. Duration of     -0.060      0.078       0.004          0.119
Liabilities         (0.022)     (0.003)     (0.880)     (<0.001)
17. Leverage        -0.091      -0.095      0.257          0.079
                    (0.001)     (0.000)     (<0.001)     (0.003)
18. M/B Ratio       -0.050      0.295       -0.212         0.166
                    (0.055)     (<0.001)    (<0.001)    (<0.001)
19. Non-Invested    -0.010      0.028       -0.086         0.044
Assets              (0.693)     (0.293)     (0.001)      (0.095)
20. Common          0.084       0.076       -0.265         0.012
Stock               (0.001)     (0.004)     (<0.001)     (0.648)
21. Firm Age        -0.061      -0.011      0.086          0.050
                    (0.020)     (0.686)     (0.001)      (0.055)
22. Dividend        -0.010      -0.175      -0.002         0.106
                    (0.707)     (<0.001)    (0.926)     (<0.001)

                    13          14          15          16

1. Cash Holdings
2. BD_Outside
Director%
3. FC_Outside
Director%
4. Board Size
5. BD_Avg. # of
Directorships
6. Board Director
Ownership
7. Log(BD_# of
Meeting)
8. FC_% Expert
9. FC_5%
Blockholder
10. CEO/Chair
Duality
11. Institutional
Ownership
12. G-Index

13. Size            1.000

14. Financial       0.208       1.000
Strength            (<0.001)
15. Volatility of     -0.184      -0.094       1.000
Cash Flows          (<0.001)     (0.000)
16. Duration of        0.199       0.027      -0.020       1.000
Liabilities         (<0.001)     (0.297)     (0.437)
17. Leverage           0.379      -0.063      -0.107       0.131
                    (<0.001)     (0.017)    (<0.001)    (<0.001)
18. M/B Ratio          0.135       0.301      -0.004      -0.064
                    (<0.001)    (<0.001)     (0.871)     (0.014)
19. Non-Invested       0.021       0.042       0.110      -0.104
Assets               (0.423)     (0.108)    (<0.001)    (<0.001)
20. Common             0.452       0.187      -0.121       0.081
Stock               (<0.001)    (<0.001)    (<0.001)     (0.002)
21. Firm Age           0.442       0.020      -0.033       0.152
                    (<0.001)     (0.446)     (0.207)    (<0.001)
22. Dividend           0.380       0.113      -0.068       0.004
                    (<0.001)    (<0.001)     (0.010)     (0.875)

                    17          18         19         20

1. Cash Holdings
2. BD_Outside
Director%
3. FC_Outside
Director%
4. Board Size
5. BD_Avg. # of
Directorships
6. Board Director
Ownership
7. Log(BD_# of
Meeting)
8. FC_% Expert
9. FC_5%
Blockholder
10. CEO/Chair
Duality
11. Institutional
Ownership
12. G-Index
13. Size
14. Financial
Strength
15. Volatility of
Cash Flows
16. Duration of
Liabilities
17. Leverage           1.000
18. M/B Ratio          0.038       1.000
                     (0.150)
19. Non-Invested       0.326       0.221      1.000
Assets              (<0.001)    (<0.001)
20. Common            -0.144      -0.024     -0.127      1.000
Stock               (<0.001)     (0.364)   (<0.001)
21. Firm Age           0.136      -0.038     -0.032      0.226
                    (<0.001)     (0.143)    (0.222)   (<0.001)
22. Dividend           0.233      -0.044     -0.094      0.178
                    (<0.001)     (0.091)    (0.000)   (<0.001)

                    21         22

1. Cash Holdings
2. BD_Outside
Director%
3. FC_Outside
Director%
4. Board Size
5. BD_Avg. # of
Directorships
6. Board Director
Ownership
7. Log(BD_# of
Meeting)
8. FC_% Expert
9. FC_5%
Blockholder
10. CEO/Chair
Duality
11. Institutional
Ownership
12. G-Index
13. Size
14. Financial
Strength
15. Volatility of
Cash Flows
16. Duration of
Liabilities
17. Leverage
18. M/B Ratio
19. Non-Invested
Assets
20. Common
Stock
21. Firm Age           1.000
22. Dividend           0.214      1.000
                    (<0.001)

Note: Sample of 1,454 U.S. stock property-liability insurers firm-
year observations during 1997-2002. See Table 1 for variable
definitions.

Table 4
Regression Analysis of the Level of Cash Holdings and Board and
Finance Committee Structure

Cash [Holdings.sub.i,t] = f(Cash [Holdings.sub.i,t-1],
Board_Outside [Director%.sub.i,t], FC_Outside [Director%.sub.i,t],
Governance Control [variables.sub.i,t], Firm Characteristics
[variables.sub.i,t], Year Dummies, Firm Fixed Effects) (1)

                                             Model 1
                                             w Board     Model 2
                                Predicted    and FC      w/o FC
Variables                         Sign      Outsider    Outsider

Intercept                                     0.351       0.386
                                             (0.90)      (0.99)
Cash [Holdings.sub.t-1]                       0.030       0.031
                                             (0.67)      (0.68)
BD_Outside Director%                +         0.134 ***   0.167 ***
                                             (2.60)      (3.63)
FC_Outside Director%                +         0.053
                                             (1.56)
Board Size                          +         0.006 *     0.007 **
                                             (1.79)      (2.19)
BD_Avg. # of Directorships          ?        -0.015 **   -0.012 **
                                            (-2.52)     (-2.11)
Board Director Ownership            +         0.596 ***   0.570 **'
                                             (3.43)      (3.28)
Board Director Ownership (2)        -        -0.733 ***  -0.680 ***
                                            (-2.76)     (-2.57)
Log(BD_# of Meeting)                -        -0.023 *    -0.023 *
                                            (-1.92)     (-1.91)
FC_% Expert                         +         0.259 *     0.201 *
                                             (1.89)      (1.61)
FC % Expert (2)                     -        -0.235 *    -0.195 *
                                            (-1.89)     (-1.67)
FC_5% Blockholder                   ?        -0.021      -0.025 *
                                            (-1.56)     (-1.93)
BD_CEO/Chair Duality                ?         0.015 **    0.015 **
                                             (2.08)      (2.10)
Institutional Ownership             ?        -0.013      -0.025
                                            (-0.41)     (-0.76)
G-Index                             ?        -0.006      -0.005
                                            (-0.73)     (-0.69)
Size                                -         0.016       0.015
                                             (1.28)      (1.25)
Financial Strength                  -        -0.030 **   -0.030 **
                                            (-2.08)     (-2.10)
Volatility of Cash Flows            +        -0.003      -0.004
                                            (-0.15)     (-0.16)
Duration of Liabilities             -        -0.107 ***  -0.108 ***
                                            (-3.35)     (-3.39)
Leverage                            ?        -0.015      -0.018
                                            (-0.28)     (-0.35)
M/B Ratio                           +         0.004       0.004
                                             (1.10)      (1.16)
Non-In vested Assets                +        -0.042      -0.042
                                            (-0.71)     (-0.71)
Common Stock                        -        -0.263 ***  -0.259 ***
                                            (-4.59)     (-4.58)
Firm Age                            +         0.054 ***   0.054 ***
                                             (2.80)      (2.82)
Dividend                            -        -0.009      -0.009
                                            (-1.22)     (-1.32)
Y1997                                         0.001       0.000
                                             (0.11)      (0.03)
Y1998                                         0.013       0.014
                                             (1.20)      (1.27)
Y1999                                         0.004       0.005
                                             (0.41)      (0.48)
Y2000                                         0.023 **    0.022 **
                                             (2.24)      (2.16)
Y2001                                         0.004       0.004
                                             (0.46)      (0.46)
Firm-year observations                       1,454       1,454
Hausman m-statistic                          78.41       72.74
(p-value)                                   (<0.0001)   (<0.0001)
[R.sup.2]                                     0.595       0.594

                                              Colquitt,
                                 Model 3       Sommer,
                                w/o Board    and Godwin
Variables                       Outsider       (1999)

Intercept                         0.369
                                 (0.95)
Cash [Holdings.sub.t-1]           0.027
                                 (0.61)
BD_Outside Director%

FC Outside Director%              0.098 ***
                                 (3.15)
Board Size                        0.006 **
                                 (2.02)
BD_Avg. # of Directorships       -0.015 **
                                (-2.44)
Board Director Ownership          0.530 ***
                                 (3.04)
Board Director Ownership (2)     -0.626 **
                                (-2.34)
Log(BD_# of Meeting)             -0.020 *
                                (-1.75)
FC_% Expert                       0.284 **
                                 (2.08)
FC % Expert (2)                  -0.252 *'
                                (-2.04)
FC_5% Blockholder                -0.021
                                (-1.55)
BD_CEO/Chair Duality              0.011 *
                                 (1.62)
Institutional Ownership          -0.003
                                (-0.10)
G-Index                          -0.005
                                (-0.63)
Size                              0.016       -0.0189 ***
                                 (1.33)       (0.0001)
Financial Strength               -0.030 **    -0.0180 ***
                                (-2.11)       (0.0001)
Volatility of Cash Flows          0.002        0.1333 ***
                                 (0.11)       (0.0351)
Duration of Liabilities          -0.109 ***   -0.0213 ***
                                (-3.36)       (0.0001)
Leverage                         -0.009       -0.0356 *
                                (-0.16)       (0.0888)
M/B Ratio                         0.002        0.0054
                                 (0.76)       (0.5126)
Non-In vested Assets             -0.046        0.0229
                                (-0.79)       (0.4459)
Common Stock                     -0.265 ***   -0.1753 ***
                                (-4.56)       (0.0001)
Firm Age                          0.055 ***
                                 (2.79)
Dividend                         -0.009
                                (-1.23)
Y1997                             0.002
                                 (0.21)
Y1998                             0.015
                                 (1.45)
Y1999                             0.004
                                 (0.40)
Y2000                             0.026 **
                                 (2.47)
Y2001                             0.006
                                 (0.70)
Firm-year observations           1,454
Hausman m-statistic              79.840
(p-value)                       (<0.0001)
[R.sup.2]                         0.592        0.117

Note: Sample of 1,454 U.S. stock property-liability insurers firm-
year observations during 1997-2002. The t-statistics are in
parentheses. See Table 1 for variable definitions. * Significant at
the 10% level. ** Significant at the 5% level. *** Significant at
the 1% level.

Table 5
Regression Analysis of Future Growth and Governance Variables of
Interest

Panel A: Using Market-to-Book Ratio as Dependent Variable

Variables                        Model 1      Model 2       Model 3

Intercept                        -3.758 ***   -4.084 ***    -3.777 ***
                                 (-2.64)      (-2.86)       (-2.65)
Excess [Cash.sub.t-1]            -0.321       -0.309        0.005
                                 (-0.97)      (-0.91)       (0.02)
BD Outside [Director%.sub.t-1]   1.985 ***    1.281 ***
                                 (5.71)       (4.51)
BD_Outside [Director%.sub.t-1]   0.609        1.710 **
  x Excess [Cash.sub.t-1         (0.22)       (1.97)
  (positive)]
BD_Outside [Director%.sub.t-1]   5.906        -0.512
  x Excess [Cash.sub.t-1         (1.62)       (-0.47)
  (negative)]
FC_Outside [Director%.sub.t-1]   -1.081 ***                 -0.353 *
                                 (-4.11)                    (-1.67)
FC_Outside [Director%.sub.t-1]   1.244                      1.403 *
  x Excess [Cash.sub.t-1         (0.52)                     (1.85)
  (positive)]
FC_Outside [Director%.sub.t-1]   -6.688 *                   -1.132
  x Excess [Cash.sub.t-1         (-1.79)                    (-1.06)
  (negative)]
Board [Size.sub.t-1]             0.006        0.012         0.019
                                 (0.22)       (0.43)        (0.67)
BD_Avg. # of                     0.048        0.010         0.066 **
  [Directorship.sub.t-1]         (1.50)       (0.32)        (2.04)
Board Director                   6.368 ***    7.128 ***     5.568 ***
  [Ownership.sub.t-1]            (5.67)       (6.32)        (4.97)
Board Director                   -10.991 ***  -12.498 ***   -9.725 ***
  [Ownership.sup.2.sub.t-1]      (-6.09)      (-6.84)       (-5.30)
Log(BD_# of Meeting)             -0.123 *     -0.111 *      -0.051
                                 (-1.92)      (-1.70)       (-0.86)
FC_% [Expert.sub.t-1]            0.449 *      0.750" *      0.520 *
                                 (1.66)       (2.85)        (1.84)
FC_5% [Blockholder.sub.t-1]      -0.119 *     -0.072        -0.131 *
                                 (-1.65)      (-1.02)       (-1.84)
BD_CEO/Chair [Duality.sub.t-1]   -0.125 **    -0.141 ***    -0.179 ***
                                 (-2.18)      (-2.56)       (-3.29)
Institutional                    0.366        0.577 **      0.526 **
  [Ownership.sub.t-1]            (1.47)       (2.18)        (1.99)
[G-Index.sub.t-1]                0.195 ***    0.187 ***     0.197 ***
                                 (4.54)       (4.31)        (4.16)
Independent System               0.197        0.201         0.234
                                 (1.36)       (1.38)        (1.56)
Line_HHI                         0.216        0.210         0.268 *
                                 (1.41)       (1.35)        (1.71)
State_HHI                        -0.125       -0.120        -0.109
                                 (-1.07)      (-1.01)       (-0.92)
Size                             0.158 **     0.176 **      0.177 ***
                                 (2.22)       (2.47)        (2.49)

Panel A: Using Market-to-Book Ratio as Dependent Variable

Variables                        Model 1      Model 2       Model 3

Firm Age                         0.410 ***    0.399 ***     0.453 ***
                                 (2.77)       (2.65)        (3.06)
Y1998                            0.542 ***    0.542 ***     0.556 ***
                                 (7.15)       (7.02)        (7.25)
Y1999                            0.216 ***    0.186 ***     0.254 ***
                                 (2.71)       (2.36)        (3.16)
Y2000                            0.613 ***    0.590 ***     0.620 ***
                                 (7.27)       (7.12)        (7.35)
Y2001                            0.384 **'    0.397 ***     0.411 ***
                                 (7.22)       (7.21)        (7.51)
Y2002                            -0.065       -0.079        -0.052
                                 (-1.20)      (-1.47)       (-0.97)
Firm-year observations           1,142        1,142         1,142
Hausman m-statistic              30.93        92.88         69.21
(p-value)                        (0.020)      (CO.0001)     (<0.0001)
[R.sup.2]                        0.823        0.820         0.817

Panel B: Using 2 Years Direct Premium Growth Rate as Dependent
Variable

Variables                        Model 1      Model 2       Model 3

Intercept                        3.114 ***    2.976 ***     3.021 ***
                                 (3.56)       (3.44)        (3.45)
Excess [Cash.sub.t-1]            -0.186       -0.150        -0.147
                                 (-1.14)      (-0.94)       (-0.91)
BD_Outside [Director%.sub.t-1]   0.019        -0.229
                                 (0.11)       (-1.59)
BD_Outside [Director%.sub.t-1]   -0.819       0.978 *
  x Excess                       (-0.54)      (1.66)
  [Cash.sub.t-1(positive)]
BD_Outside [Director%.sub.t-1]   2.585        -0.691
  x Excess                       (1.25)       (-1.10)
  [Cash.sub.t-1(negative)]
FC_Outside [Director%.sub.t-1]   -0.333 **                  -0.281 ***
                                 (-2.39)                    (-2.59)
FC_Outside [Director%.sub.t-1]   1.911                      1.055 *
  x Excess                       (1.33)                     (1.89)
  [Cash.sub.t-1(positive)]
FC_Outside [Director%.sub.t-1]   -3.359                     -0.874
  x Excess                       (-1.53)                    (-1.3)
  [Cash.sub.t-1(negative)]
Board [Size.sub.t-1]             0.018 **     0.013         0.019 **
                                 (2.04)       (1.44)        (2.14)
BD_Avg. # of                     0.047 **     0.040 *       0.048 **
  [Directorship.sub.t-1]         (2.25)       (1.88)        (2.31)
Board Director                   -0.726       -0.564        -0.664
  [Ownership.sub.t-1]            (-1.20)      (-0.95)       (-1.09)
Board Director                   0.676        0.301         0.563
  [Ownership.sup.2.sub.t-1]      (0.69)       (0.31)        (0.57)

Panel B: Using 2 Years Direct Premium Growth Rate as Dependent
Variable

Variables                        Model 1      Model 2       Model 3

Log(BD_# of Meeting)             -0.023       -0.018        -0.023
                                 (-0.74)      (-0.60)       (-0.75)
FC_% [Expert.sub.t-1]            -0.276 ***   -0.201 **     -0.274 ***
                                 (-2.73)      (-2.15)       (-2.73)
FC_5% [Blockholder.sub.t-1]      0.063        0.075 *       0.062
                                 (1.43)       (1.71)        (1.41)
BD_CEO/Chair [Duality.sub.t-1]   -0.039       -0.044 *      -0.037
                                 (-1.52)      (-1.75)       (-1.51)
Institutional                    -0.251 **    -0.201        -0.255 *
  [Ownership.sub.t-1]            (-1.79)      (-1.49)       (-1.85)
[G-Index.sub.t-1]                0.009        0.009         0.008
                                 (0.38)       (0.36)        (0.34)
Independent System               0.045        0.051         0.046
                                 (0.81)       (0.94)        (0.83)
Line_HHI                         0.198 *      0.195 *       0.198 *
                                 (1.75)       (1.72)        (1.75)
State_HHI                        -0.104       -0.105        -0.105
                                 (-1.33)      (-1.31)       (-1.34)
Size                             -0.176 ***   -0.168 ***    -0.172 ***
                                 (-3.94)      (-3.80)       (-3.85)
Firm Age                         -0.020       -0.023        -0.025
                                 (-0.19)      (-0.22)       (-0.25)
Y1998                            0.049        0.054         0.051
                                 (1.14)       (1.28)        (1.22)
Y1999                            0.095 **     0.092 **      0.096 **
                                 (2.30)       (2.24)        (2.41)
Y2000                            0.102 ***    0.102 ***     0.105 ***
                                 (2.77)       (2.72)        (2.91)
Y2001                            0.082 **     0.087 ***     0.081 **
                                 (2.29)       (2.49)        (2.31)
Y2002                            0.020        0.019         0.022
                                 (0.73)       (0.69)        (0.81)
Firm-year observations           1,111        1,111         1,111
Hausman m-statistic              27.91        41.13         62.12
(n-value)                        (0.032)      (0.001)       (<0.0001)
[R.sup.2]                        0.547        0.544         0.546

Note: Sample period is from 1998 to 2003. The f-statistics are in
parentheses. The dependent variable, future growth rate, is the
ratio of the market value to book value of insurers' assets in Panel
A, and 2 years direct premium growth rate in Panel B. Excess
[Cash.sub.t-1] is measured as the lagged excess cash holdings. Board
and finance committee variables are also lagged for one period and
can be referred in Table 1. Independent System is measured as 1 if
the insurer uses independent agent and 0 otherwise. Line_HHI
(State_HHI) is measured as the sum of squared percentage of direct
premiums written by lines of insurance (by state). Size is the
natural log of assets. Firm Age is the natural log of the difference
between firms' ages and 5. Finally, Y1998 to Y2002 are year dummy
variables. * Significant at the 10% level. ** Significant at the 5%
level. *** Significant at the 1% level.

Table 6

Regression Analysis of the Impact of Main Governance Variables of
Interest on the Association Between the Excess Cash Holdings and
Salary Expense Divided by Net Premium Written

Variables                          Model 1      Model 2     Model 3

Intercept                          0.231 *      0.267 **    0.267 **
                                   (1.89)       (2.10)      (2.15)
Excess [Cash.sub.t-1]              0.077 *      0.067       0.62 *
                                   (1.70)       (1.48)      (1.45)
BD_Outside [Director%.sub.t-1]     0.036        0.009
                                   (1.00)       (0.29)
BD_Outside [Director%.sub.t-1] x   0.450        -0.126
  Excess                           (1.46)       (-1.12)
  [Cash.sub.t-1(positive)]
BD_Outside [Director%.sub.t-1] x   -1.496 *     -0.424 **
  Excess                           (-1.68)      (-2.17)
  [Cash.sub.t-1(negative)]
FC_Outside [Director%.sub.t-1]     -0.060 ***               -0.074 ***
                                   (-2.95)                  (-3.11)
FC_Outside [Director%.sub.t-1] x   -0.597 *                 -0.133
  Excess                           (-1.87)                  (-1.19)
  [Cash.sub.t-1(positive)]
FC_Outside [Director%.sub.t-1] x   1.116                    -0.347 **
  Excess                           (1.34)                   (-2.09)
  [Cash.sub.t-1(negative)]
Board [Size.sub.t-1]               0.004 **     0.001       0.003 *
                                   (2.02)       (0.66)      (1.86)
BD_Avg. # of                       0.000        0.005       0.000
  [Directorship.sub.t-1]           (0.08)       (-1.01)     (0.06)
Board Director                     -0.055       -0.020      -0.126
  [Ownership.sub.t-1]              (-0.54)      (-0.18)     (-1.20)
Board Director                     0.274        0.206       0.398 **
  [Ownership.sup.2.sub.t-1]        (1.61)       (1.14)      (2.26)
Log(BD_# of Meeting)               0.009        0.008       0.010
                                   (1.28)       (1.18)      (1.45)
FC_% [Expert.sub.t-1]              -0.011       0.015       -0.012
                                   (-0.56)      (0.81)      (-0.57)
FC_5% [Blockholder.sub.t-1]        -0.011       -0.006      -0.011
                                   (-1.35)      (-0.76)     (-1.44)
BD_CEO/Chair [Duality.sub.t-1]     -0.005       -0.006      -0.007
                                   (-1.04)      (-1.23)     (-1.43)
Institutional                      -0.050 *     -0.028      -0.044 *
  [Ownership.sub.t-1]              (-2.18)      (-1.23)     (-1.89)
[G-Index.sub.t-1]                  -0.007 *     -0.008 **   -0.007 *
                                   (-1.75)      (-2.00)     (-1.74)
M/B Ratio                          0.004        0.005 *     0.004
                                   (1.29)       (1.64)      (1.58)
Commercial Line%                   -0.000       0.000       -0.005
                                   (-0.01)      (0.00)      (-0.24)
Independent System                 -0.001       -0.003      -0.000
                                   (-0.17)      (-0.05)     (-0.06)
Line_HHI                           0.007        0.004       0.007
                                   (0.32)       (0.17)      (0.32)
State_HHI                          -0.031 **    -0.027 *    -0.029 **
                                   (-2.21)      (-1.90)     (-2.07)
Size_DPW                           -0.010 *     -0.010"     -0.010 *
                                   (-1.86)      (-1.94)     (-1.82)
Firm Age                           0.012        0.011       0.015
                                   (0.70)       (0.66)      (0.91)
Y1998                              -0.002       -0.006      -0.004
                                   (-0.25)      (-0.63)     (-0.41)
Y1999                              0.019 ***    0.013       0.020 ***
                                   (2.48)       (1.61)      (2.48)
Y2000                              -0.002       -0.008      -0.004
                                   (-0.29)      (-0.97)     (-0.48)
Y2001                              0.003        0.003       0.004
                                   (0.50)       (0.41)      (0.61)
Y2002                              0.002        -0.001      0.002
                                   (0.46)       (0.10)      (0.34)
Firm-year observations             1,082        1,082       1,082
Hausman m-statistic                53.80        22.09       56.47
(p-value)                          (<0.0001)    (0.003)     (<0.0001)
[R.sup.2]                          0.659        0.652       0.655

Note: Sample period is from 1998 to 2003. In examining the
association between excess cash and board and finance committee
independence, we use the lagged variables. Therefore, in Tables 5
and 6, the sample period is starting from 1998 to 2003. The
f-statistics are in parentheses. The dependent variable, Salary
Expense Ratio, is measured as the salary expense divided by new
premium written. Excess [Cash.sub.t-1] is measured as the lagged
excess cash holdings. Board and finance committee variables are
also lagged for one period and can be referred in Table 1. M/B
Ratio is the ratio of the market value to book value of insurers'
assets. Commercial Line% is measured as the percentage of direct
premium from commercial lines. Independent System is measured as 1
if the insurer uses independent agent and 0 otherwise. Line_HHI
(State_HHI) is measured as the sum of squared percentage of direct
premiums written by lines of insurance (by state). Size_DPW is
measured as the natural log of direct premiums written. Firm Age is
the natural log of the difference between firms' ages and 5.
Finally, Y1998 to Y2002 are year dummy variables. * Significant at
the 10% level. ** Significant at the 5% level. *** Significant at
the 1% level.
COPYRIGHT 2015 American Risk and Insurance Association, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2015 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Hsu, Wen-Yen; Huang, Yenyu "Rebecca"; Lai, Gene
Publication:Journal of Risk and Insurance
Article Type:Abstract
Geographic Code:1USA
Date:Sep 1, 2015
Words:15317
Previous Article:A portfolio optimization approach using combinatorics with a genetic algorithm for developing a reinsurance model.
Next Article:Insurance and the Law of Obligations.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters