2018/12/31 What Can Deter Catering? Information Environment, Governance, and Agency Problem of Overvalued Equity Shing-yang Hu Department of Finance, National Taiwan University and Yueh-hsiang Lin Department of Finance, National Taipei College of Business
2018/12/31 Motivation Catering: corporate decisions made to meet the demand of irrational investors who are willing to pay a price higher than fundamentals. When equity is overvalued, companies invest more (Baker, Stein, and Wurgler, 2003; Polk and Sapienza, 2009; Stein, 1996), initiate dividends (Baker and Wurgler, 2004), and use more accruals (Badertscher, 2011; Chi and Gupta, 2009). Jensen’s (2005): Overvalued equity gives rise to an agency problem as these market-driven behaviors can undermine core firm value. Researchers have yet to address fully whether any existing mechanism can deter managerial catering behaviors.
2018/12/31 Objective This study examines whether two monitoring mechanisms, a transparent information environment and a sound governance system, can reduce catering of two corporate decisions: capital expenditures and earnings accrual. Capital expenditure decides the cash flow of companies. Earnings accruals only affect reported earnings but not cash flows. Our empirical results: information environment is important when managers make investment and accrual decisions in the overvaluation and in reducing catering.
2018/12/31 Hypothesis Information hypothesis: A better information environment can deter catering by helping investors to see through those corporate actions with the sole purpose of catering. Under constant monitoring, managerial misbehavior such as inflated earnings through management or negative net present value investment is more likely to be detected, and investors are less likely to be fooled. If made less effective, managers are less likely to pursue catering behaviors. Implication: The relation between overvaluation and investment/accrual decisions are weaker when firms have a better information environment.
2018/12/31 Hypothesis Governance hypothesis: A good governance mechanism can deter catering by disciplining managers. For companies with good governance, managers may face employment termination if their misbehavior is revealed. Managers will resist catering even when faced with overvaluation if sufficient monitoring is in place. Implication: The relation between overvaluation and investment/accrual decisions are weaker when firms have a better governance system.
Information variables 2018/12/31 Information variables One way to improve the information environment is through monitoring from informational intermediaries such as rating agencies and security analysts. Inclusion in the S&P 1500 index. Analyst coverage (the number of analysts following the firm). Firm size: Large firms suffer less information asymmetry. The use of Big Eight auditors. Big audit firms have better technologies, a greater incentive to make an increased effort in auditing to protect their reputation. Given that the role of auditors is to improve the accuracy of accounting statements, we only use this proxy to examine the sensitivity of overvaluation and earnings management.
Governance variables Gompers et al.’s (2003) governance index: 2018/12/31 Governance variables Gompers et al.’s (2003) governance index: Measures both shareholder rights and the extent of takeover defense. Managers of firms with stronger shareholder rights or weak takeover defense are more highly monitored. Bergstresser and Philippon’s (2006) incentive ratio (equity-linked incentive compensation): Managers benefit more when they act to increase stock price and improve shareholders’ welfare. However, closely linking compensation to the stock price can also induce managers to seek short-term profits and increase personal wealth. Whether incentive compensation reduces or increases opportunistic managerial behavior is an empirical question.
Measure of overvaluation 2018/12/31 Measure of overvaluation Sample: Non-financial firms listed on the NYSE, Amex, and Nasdaq included in the COMPUSTAT and CRSP databases from 1987 through 2008. Estimate the following regression for each of the 12 Fama and French industries for each year t: The estimated fundamental value (V) is the natural exponent of the fitted value. M/V measures the discrepancy between market valuation and fundamental value and reflects the degree of overvaluation.
Classify sample observations into quartiles by M/V. 2018/12/31 Classify sample observations into quartiles by M/V. Firms in the highest (lowest) M/V quartile experience More (less) MA and SEO in the following year. High (low) lagged returns and low (high) future returns. High (low) forecast-based market-to-fundamental values (M/VF), where the fundamental values is estimated based on a discounted residual income approach and analysts’ forecast of future EPS from the I/B/E/S database.
Accruals and investment 2018/12/31 Accruals and investment Measure of earnings management: performance-matched discretionary accrual, estimated through three steps. Total accruals (Ac) from items on the statement of cash flow: Conventional Modified Jones discretionary accrual models (regress Ac on change in sales minus change in AR and PPE) A matching-firm approach: For each sample firm in the highest and lowest M/V quartiles, we choose a matching firm with same two-digit SIC code as the sample firm and with the closest ROA in year t+1 from all observations in the second and third quartiles. Measure of investment: capital expenditures scaled by the lagged capital.
2018/12/31 Firms in the highest M/V quartile have larger average discretionary accrual (–0.0049) than those in the lowest quartile (–0.0090), although the mean difference is not significant at any conventional level. In untabulated tests, we use unmatched discretionary accruals and find that the accruals are significantly higher for overvalued companies. This result echoes Kothari et al.’s (2005) findings that controlling for operating performance is important for studies on accruals. There is a positive relation between M/V and capital expenditure. The mean of the capital expenditure is 0.71 (0.35) for the highest (lowest) quartile; the difference of 0.36 is significant at the 0.01 level.
2018/12/31 Overvalued firms do not (do) have higher accruals for firms with a better (poor) information environment [Panel A (Panel B)], which is consistent with the information hypothesis. Overvalued firms do not have higher accruals regardless of firms’ extent of governance mechanism, which is not consistent with the governance hypothesis. Because firms with a low governance index and firms with low incentive ratio are covered by ExecuComp and IRRC databases, they may face a better information environment and the governance mechanism become less important.
2018/12/31 A better information environment reduces the extent of investment of overvalued firms, which is consistent with the information hypothesis. Overvalued firms do not have higher investment regardless of firms’ extent of governance mechanism.
2018/12/31 Accrual–overvaluation sensitivity is weaker when firms have a better information environment. Accrual–overvaluation sensitivity is weak regardless of the extent of the governance index or incentive compensation.
2018/12/31 Capital expenditure–overvaluation sensitivity is weaker when firms have a better information environment. Capital expenditure–overvaluation sensitivity is weak regardless of the extent of the governance index or incentive compensation.
Economic significance 2018/12/31 Economic significance The sensitivity (the unit change of a quartile deviation of the accruals/investment with respect to per-unit change of a quartile deviation of M/V) drops in the earnings management/investment regressions when firms have a better information environment. Among the information variables, the S&P 1500 inclusion dummy variable have the strongest ability to reduce earnings management/investment of overvalued companies.
2018/12/31 Robustness Test The evidence related to information hypothesis is robust when using alternative overvaluation measures such as the duration of overvaluation, lagged stock returns, the forecast-based market-to-fundamental values, instrumental variable (market investor sentiment levels as well as firm characteristics), and accrual/investment premium.
2018/12/31 Robustness Test Prior literature suggests that managers with shorter shareholder horizons cater more. The information environment plays a role exactly when it is needed the most (i.e., firms with shorter managerial horizon have significantly negative interaction coefficients, Panel A).
2018/12/31 Robustness Test Catering is not necessarily to benefit fundamentals and may result in a worse stock performance in the future. If firms with a better information environment catering less, their future returns should be less negatively related to accruals and investment. For S&P index inclusion, we observe significantly negative coefficients on accruals and investment, and positive interaction coefficients.
2018/12/31 Conclusions Using U.S. data from 1987 to 2008, we find a weak sensitivity of overvaluation with different measures of accruals and capital investment when firms have a better information environment, such as S&P 1500 firms, firms covered by more analysts, and large firms. The impact of the governance index and incentive compensation on the earnings management and investment is empirically weaker. The effect of a strong information environment on managerial misbehavior is robust for alternative measures of overvaluation and a subsample of high turnover firms. For S&P 1500 firms, we find that future returns have a weaker negative relation to accruals and investment. For overvalued companies, information environment is able to deter the catering and resolve the agency problem of overvalued equity. The End