배당과 감사보수 사이의 관계

The Association between Dividends and Audit Fees

본 연구에서는 배당을 지급하는 기업과 배당을 지급하지 않는 기업들을 비교할 때 감사보수에 어떤 차이가 있는지를 연구하였다. 선행연구의 발견에 따르면 배당을 지급하는 기업들은 배당지급을 통해 정보불균형을 해소하고, 자신의 재무제표의 품질이 높고 숨겨진 감사위험이 낮다는 신호로 사용할 수 있다. 또한 배당을 지급하는 기업들은 대리인 비용도 적게 발생할 것이므로 고품질의 감사 서비스에 대한 수요가 낮을 수 있다. 따라서 고품질 감사에 대한 수요가 낮다면 엄밀한 감사를 요청하지 않을 것이므로 감사투입시간이 줄어들어 감사보수가 감소할 것이다. 그러나 그 반대로 감사인이 배당 이외의 보다 자세한 정보를 더 많이 가지고 있으므로 배당이 감사보수에 영향을 미치지 않을 수도 있다. 2002년부터 2014년까지의 19,131개의 기업-연도 표본을 이용하여 분석한 결과, 배당을 지급하는 기업들은 지급하지 않는 기업들과 비교할 때 감사보수가 유의적으로 낮았다. 또한 배당액이 증가할수록 감사보수가 더 낮았다. 그런데, 상대적으로 Big 4로부터 감사를 받는 피감기업들은 이러한 성향이 덜 나타났다. 이 결과는 Big 4가 non-Big 4에 비해 기업의 특성이 변하더라도 상대적으로 일정한 수준의 감사노력을 피감기업들에게 투입하기 때문에, 배당의 지급여부나 배당금의 수준이 변한다고 하더라도 감사노력 수준이 크게 변하지 않는다고 해석할 수 있다. 이런 발견은 배당의 역할과 감사인의 행동에 대해 여러 유익한 시사점을 제공한다.

This study examines whether dividend payments of client firms are associated with auditor’s audit efforts which is reflected in audit fees. Dividend payments is the process that return the profit, that the client firms generated through the operation, to shareholders. If there is no tax effect, theoretically, dividend payments should not influence the client firm value (Miller and Modigliani 1961). However, prior studies document that stock market significantly respond to the announcement or payment of dividends (or increase/decrease/initiation/omission of dividends). For example, Charest (1978), Aharony and Swary (1980) document that stock market interprets the news of dividend increase (decrease) as good (bad) news and stock prices, in contrast to the theoretical prediction, respond the news accordingly. There are two different views why stock market responds to the news on the dividends: signaling hypothesis and agency cost hypothesis. Under the signaling view, prior studies report that earnings quality is higher for firms that paying dividends (Skinner and Soltes 2011; Caskey and Hanlon 2013). As a result, dividend payment can be regarded as a signal of high quality earnings or financial reporting system. Specifically, Miller and Rock (1985) argue that dividend payment is costly, in the sense that it requires a cash outflow, thus credible, signal used in the capital market. Alternatively, from the view of agency cost, client firms pay cash to investors as a form of dividend, which reduce agency problem (Eastebrook 1984). For example, managers have incentive to try to build an empire by investing in unnecessary projects (i.e., empire building behavior, Jensen and Meckling 1976). Dividend payments can reduce such incentive by restricting the empire building behavior, thereby reducing agency problem. Hribar et al. (2014) argue that auditors’ unobservable private information on the financial quality of clients is reflected in audit planning process. Auditors, in their audit process, may respond to the client firms’ divided payment by adjusting their effort level. First, dividend payments signal the high-quality of the dividend-paying client firms. Thus, auditors may not feel that it is necessary to increase effort level for such high-quality client firms which are not likely to have any hidden risks. Second, dividend payments reduce the agency costs. Prior studies document that agency costs play a role in demanding for thorough audit service (Fan and Wong 2005). Thus, auditors do not have specific reasons for increasing effort level for the client firms who have less incentive to ask for high-quality audit to auditors. Summarizing, auditors may reduce the audit effort for dividend-paying client firms due to these abovementioned two reasons. Furthermore, it is possible that Big 4 auditors, who are likely to provide relatively stable quality of service in different situations (Choi et al. 2008; Khurana et al. 2011) than non-Big 4 auditors, are less likely reduce audit effort for dividend-paying client firms. However, it is also possible that auditors do not care for dividend payments or the amount of dividends. Auditors are supposed to collect many information, financial information as well as non-financial, qualitative information in the audit process. These informations include more detailed and less straight forward information which is not disclosed publicly, for example, in the annual report. Thus, as long as auditors already have more detailed information on the client firms, the dividend payment itself should not be related to auditor’s effort level. In short, how audit fee is related to dividend is an empirical question that need to be resolved. Using 19,131 firm-year observations collected over the period from 2002 to 2014 in Korean stock market, we empirically test these predictions. Our findings are summarized as follows.