본 연구는 상장기업을 대상으로 세무위험 측정치들이 감사인의 감사위험으로 평가되어 감사보수 및 감사시간을 증가시키는지 실증적으로 분석하였다. 특히, 본 연구는 세무위험 측정치로 네 가지를 살펴보았다. 첫 번째와 두 번째 측정치는 Dhaliwal et al.(2017)에서 제안된 과거 5년간 과세소득의 변동성(이하 TI_VAR)과 회계이익과 과세소득 간의 5년간 상관성 정도(BT_COV)이고, 세 번째 측정치는 과거 5년간 CASH 및 GAAP 유효세율의 변동성(CASH 또는 GAAP ETR_VAR)이며, 네 번째 측정치는 과거 5년간 회계이익과 과세소득의 차이 변동성(BTD_VAR)이다. 이들 변수 중 TI_VAR, BT_COV 및 CASH ETR_VAR은 선행연구들에서 투자자 측면에서 기업위험과 관련성이 나타난 측정치이고(Hutchens and Rego 2015; Dhaliwal et al. 2017; Guenther et al. 2017), BTD_VAR은 본 연구에서 새롭게 제안한 측정치이다. 이들 모두는 세무관련 불확실성을 나타내는 측정치들이다. 아울러 추가분석에서는 앞서의 세무위험 측정치와 감사보수 및 감사시간과의 양(+)의 관계가 발생액의 질에 따라 차별적인지에 대해서도 살펴보았다. 분석을 위해 본 연구는 분석기간 2003년부터 2016년까지 자료를 이용하여 금융업을 제외한 12월 결산법인을 대상으로 이용가능 했던 감사보수(감사시간)의 최종표본 6,582개(6,584개) 기업/연 자료가 분석에 이용되었다. 실증결과는 다음과 같다. 첫째, 감사보수와 감사시간에 영향을 미치는 일정 변수를 통제한 후에도 과세소득의 변동성이 높거나, 회계이익과 과세소득 간의 상관성이 낮을수록 감사보수 및 감사시간은 통계적으로 높게 나타났다. 또한 이들 세무관련 변수들 간에는 추가적인 정보력이 있는 것으로 나타났다. 둘째, 앞서의 세무위험 측정치에 추가하여 ETR의 변동성을 포함하면 GAAP ETR_VAR의 경우 감사인의 감사보수 및 감사시간에 대해 유의한 양(+)의 관계로 나타났다. 또한 TI_VAR과 BT_COV는 여전히 유의한 양(+)의 관계로 나타났다. 하지만 CASH ETR_VAR은 감사인의 감사보수 및 감사시간과 유의한 결과로 나타나지 않았다. 셋째, 앞서 BT_COV와 ETR_VAR를 포함한 후에도 회계이익과 과세소득의 차이 변동성이 높은 기업일수록 감사보수 및 감사시간은 유의하게 높게 나타났고, 또한 GAAP ETR_VAR은 여전히 양(+)의 결과를, BT_COV는 더 이상 유의한 결과를 보이지 않았다. 이상의 결과는 과세소득의 변동성, ETR 중 GAAP ETR의 변동성 및 회계이익과 과세소득의 차이 변동성이 클 때 감사인은 주로 감사위험으로 평가하여 감사보수가 증가할 뿐만 아니라 추가적인 감사노력이 투입된다는 발견이다. 마지막으로, 추가분석에서 전체표본을 다시 발생액의 질(AQ)을 기준으로 중위수에 따라 나누어 분석하면 발생액의 질이 높은 집단보다는 낮은 집단일 때 앞서의 관계는 더 뚜렷한 양(+)의 결과로 나타났다. 이 결과는 낮은 발생액의 질이 주로 세무위험 측정치와 감사보수 및 감사시간 간의 양(+)의 관계에 영향을 준다는 것을 시사한다. 이상을 종합하면, 본 연구는 과세소득의 변동성, GAAP ETR의 변동성 및 BTD의 변동성이 높을 때 감사인은 이를 감사위험으로 평가하여 감사보수뿐 아니라 추가적인 감사노력이 투입된다는 것을 보여주었다는데 의의가 있다. 또한 세무위험을 다룬 선행연구들이 주로 투자자 측면의 기업위험과의 관계를 분석한 것과 달리, 본 연구는 이를 감사인의 감사위험 측면에서 분석한 처음의 연구라는 점에서 의미가 있다. 아울러 본 연구는 앞서의 결과가 재무보고의 질이 낮을 때 더 뚜렷한 반응이 있음을 보여주었다. 따라서 본 연구결과는 감사보수 및 감사시간의 결정모형을 다룬 연구 및 세무위험과 기업위험 간의 관계를 다룬 관련연구에 추가적인 새로운 실증적 증거를 제공할 것으로 기대된다. 아울러 본 연구의 발견은 국내 상장기업들에서의 세무위험이 감사위험으로 평가됨을 보여주고 있기 때문에 학계뿐 아니라 과세당국, 규제당국이나 정책당국에게도 유용한 정보를 더불어 제공할 것으로 예상된다.

This study investigates whether the relation between the tax risk and audit fees as well as audit hours. Specifically, we utilize tax risk measures that are designed to capture greater tax-related uncertainty, including each of four measures of firms risk related to the volatility of taxable income (hereafter TI_VAR), the covariance between taxable income and book income (hereafter BT_COV), the volatility of cash or GAAP effective tax rates (hereafter CASH ETR_VAR, GAAP ETR_VAR), and the volatility in book-tax differences (hereafter BTD_VAR) over five years (the current year and the previous four years). Each of these measures captures tax decisions that potentially impose significant costs on firms and thus, should influence the market’s assessment of future after-tax cash flows. Recent research in the tax literature suggests that tax risk is an important construct, and numerous studies have introduced potential academic tax risk measures (e.g., Hutchens and Rego 2015; Dhaliwal et al. 2017; Guenther et al. 2017 etc.). The effects of tax risk have been discussed in almost every decision context in the extant tax literature, but the relationship between tax risk and auditing has amazingly merited little attention. We explore the relationship between tax risk and auditor pricing as well as auditor effort measured as audit hours. In this study, we investigate whether greater tax risk is associated with increased audit risk, thus tax risk that affect audit pricing and auditors’ additional audit effort. Therefore, we test our prediction that auditors assess higher audit fees and audit hours on client with higher levels of tax risk. Thus, we test four hypotheses about the relation between tax risk and audit fees as well as audit hours. In additional analysis, we next examine whether the effect of tax risk according to accruals quality proposed by Francis et al. (2005) have differential impact on the audit risk by auditors. For analysis, we consider four additional proxies for tax risk in our empirical tests. First, we follow Dhaliwal et al. (2017) use the volatility of taxable income (TI_VAR), which is the standard deviation of taxable income deflated by lagged total assets calculated for years t-4 through t. Second, we follow Dhaliwal et al. (2017) use the covariance between book income and taxable income (BT_COV), which is the correlation between pretax book income and taxable income deflated by lagged total assets calculated for years t-4 through t. Then, we measure BT_COV is (-1) times, thus higher value of BT_COV indicate increased tax risk. Third, we use the volatility of annual both cash effective tax rates or GAAP effective tax rates over the five prior years (CASH ETR_VAR, GAAP ETR_VAR), which is the standard deviation of cash taxes paid or tax expense divided by pretax income calculated over years t-4 through t as a proxy for tax risk from Hutchens and Rego (2015) and Guenther et al. (2017). Lastly, we use the volatility in book-tax differences (BTD_VAR), which is the standard deviation of book-tax difference scaled lagged total assets calculated over years t-4 through t as our fourth proxy of tax risk proposed by this study. This study sample covers KOSPI and KOSDAQ listed firms in non-financial industries with fiscal year-end in December from 2003 to 2016 period. Our main sample consists of 6,582 (6,584) firm-year observations in audit fees (audit hours) data. We document several findings. First, after controlling for several factors that affect audit fees and audit hours, we find evidence consistent with our hypothesis that there is a positive and significant association between TI_VAR and audit fees as well as audit hours. In addition, we find that there is a positive and significant association between BT_COV and audit fees as well as audit hours. This result suggests that auditors perceive the volatility of taxable income and will be decreasing in the covariance between book income and taxable income as a risk factor that they incorporate into audit fees and hours. Also, these results provide evidence that both TI_VAR and BT_COV provide incremental information about audit risk beyond the volatility of book income as a additional control variable in our test. Second, we also include the GAAP ETR_VAR in the model, we find evidence consistent with our hypothesis that there is a positive and significant association between GAAP ETR_VAR and audit fees as well as audit hours. Nevertheless, it is still positive and significant coefficient TI_VAR as well as BT_COV measures. In contrast, we include CASH ETR_VAR in the model, we find no significant association between CASH ETR_VAR and audit fees as well as audit hours. Third, we also include BTD_VAR, BT_COV, and ETR_VAR in th model, we find evidence consistent with our hypothesis that there is a positive and significant association between BTD_VAR and audit fees as well as audit hours. Nevertheless, it is still positive and significant coefficient GAAP ETR_VAR measures. In contrast, we find no significant association between BT_COV and audit fees as well as audit hours. This is, higher BT_COV do not appear to be associated with increased audit risk even when controlling for BTD_VAR and ETR_VAR, as a additional test variable in our test. Therefore, compared to BT_COV proxy, BTD_VAR is recognized as indicators of audit risk by auditors. Lastly, when we also divided the full samples into high accruals quality and low accruals quality subsamples, according to the median level of accruals quality, we find that a positive and significant relation between tax risk measures and audit fees as well as audit hours is more pronounced among low accruals quality samples than high accruals quality samples. Therefore, this results suggest that low accruals quality drives the association between tax risk and audit risk by auditors. Overall, this study increases our understanding of which tax risk metrics capture tax-related uncertainties that lead to assessments by auditors of higher audit risk. In particular, our results suggest that the volatility of taxable income, the volatility of GAAP effective tax rates, and the volatility in book-tax differences over five years are superior proxies for a firm’s exposure to tax risk that leads to increased uncertainty about a firm’s future cash flows. Consequently, we shows that auditors, on average, perceive higher values for each of these tax risk metrics to be related to an increase in audit risk. More importantly, we find that the positive association between tax risk and audit fees as well as audit hours is more pronounced for firms with low-quality accruals. Thus, our study makes several contributions to the tax research as well as auditing literature. To the best of our knowledge this is the first study to examine the relation between tax risk and audit fees as well as audit hours. Together, our research increases our understanding of which tax risk metrics capture tax-related uncertainties that lead to assessments by auditors of higher audit risk, which serves as an important empirical evidence to academics as well as investors, tax authorities, regulators, and policymakers may provide useful information.