외환위기 전후의 한계세율을 이용한 부채사용의 순세금효과 추정과 자본구조분석

Abstract

본 연구는 추정한계세율을 이용하여 부채의 한계효용곡선을 추정하고, van Binsbergen et al.(2010)의 방법론을 이용하여 부채의 한계비용을 추정함으로써 개별기업의 부채조달에 따른 순세금효과를 계산하고 균형부채조달 여부를 분석하였다. 외환위기 이후 부채사용이 감소하고 채무불이행 위험이 증가한다는 연구결과를 기초로 하여 부채의 한계효용과 한계비용 분석은 외환위기 이전과 이후로 구분하여 분석을 진행하였다. 유가증권 상장기업의 1991년 이후의 재무자료를 이용하여 부채수준 IOB(이자비용/총자산)의 변화에 따른 한계효용과 한계비용을 분석한 결과, 외환위기 이전과 이후 기간의 실제부채수준은 각각 0.047과 0.017로 외환위기 이후 부채수준이 약 36% 수준으로 감소한 것으로 나타났다. 외환위기 이전에는 실제부채수준 100%에서의 한계효용과 한계비용은 각각 0.232와 0.265로 한계비용이 높은 반면, 실제부채수준 80%에서의 한계효용과 한계비용은 각각 0.259와 0.232로 한계효용이 높게 나타나 한계효용과 한계비용이 일치하는 균형부채수준은 실제부채수준의 80%∼100% 수준인 것으로 나타났다. 따라서 외환위기 이전에는 개별기업들이 평균적으로 균형수준보다 과대하게 부채를 사용한 것으로 나타났다. 반면 외환위기 이후 기간에는 실제부채수준 100%의 한계효용과 한계비용은 각각 0.201과 0.205로 유사하게 나타나 평균적으로 균형수준에서 부채를 사용한 것으로 나타났다. 외환위기 이전에는 균형부채수준에서의 총효용과 총비용은 각각 기업가치의 8.4%와 5.0%로 순세금효과는 3.4% 수준인 반면 실제부채수준의 총효용과 총비용은 각각 기업가치의 8.9%와 6.3%로 순세금효과는 2.6% 수준으로 나타났다. 그리고 외환위기 이후 기간에는 균형부채수준의 총효용과 총비용은 각각 기업가치의 6.1%와 4.2%로 순세금효과는 1.9% 수준인 반면 실제부채수준의 총효용과 총비용은 각각 기업가치의 5.4%와 4.5%로 순세금효과는 0.9% 수준으로 나타났다. 따라서 균형 및 실제부채수준의 순세금효과는 외환위기 이전에 비해 이후기간에 감소하는 것으로 나타났다. 그 이유는 법인세율 인하에 따라 부채사용에 따른 한계효용이 감소한 반면 한계비용은 증가하였을 뿐만 아니라, 전반적으로 부채를 보수적으로 사용함에 따라 부채수준이 감소하면서 총효용이 총비용에 비해 큰 폭으로 감소한데 따른 결과이다. 분석기간 중 부채를 과소 사용하는 경우의 후생손실(dead weight loss)은 평균적으로 총자산의 0.8%인 반면 과대 사용하는 경우의 후생손실은 총자산의 1.3%로 나타났다. 그리고 외환위기 이전에 비해 이후에는 한계효용의 감소와 한계비용의 증가에도 불구하고 부채수준이 큰 폭으로 감소하여 부채의 과소 또는 과대 사용에 따른 후생손실은 크게 변화하지 않는 것으로 나타났다. 본 연구는 한계효용곡선과 한계비용함수를 이용하여 개별 기업 차원의 균형부채수준을 분석하는 틀을 제시하였다는 점에 연구의 의의가 있다. 한계효용곡선과 한계비용함수를 이용하면 실제부채수준 뿐만 아니라 균형부채수준에서 부채를 조달함에 따른 세금효과를 개별 기업 차원에서 직접 추정할 수 있으므로 자본구조에서 오랫동안 논의되어 왔던 균형(적정) 자본구조의 연구에도 추가적인 공헌점을 제공한 것으로 판단된다. 외환위기 이후 국내 기업들의 부채정책 변화와 이와 같은 현상을 설명하는 요인이라는 점을 밝히고 있는 것도 본 연구의 공헌점이라고 하겠다.The main purpose of this paper is to investigate firm-specific equilibrium debt level by estimating firm-specific benefit and cost functions of debt. The equilibrium debt level occurs at the intersection of these benefit and cost functions. The benefit functions of debt are estimated using simulated marginal tax rates and are downward sloping reflecting that the marginal benefit of debt declines as more debt is used. To estimate cost function of debt, we observe a firm`s actual debt choice which is represented by a single point on its benefit function, and assume for our estimation sample that this point represents the equilibrium intersection of the marginal benefit and cost functions of debt. As the benefit functions move, the shift in the intersection points allows us to identify the location of the cost functions of debt. That is, the cost functions are estimated using exogenous variation in benefit functions and are upward sloping reflecting the marginal costs that occur as a firm increases its use of debt. To interpret firms` actual debt level as representing the intersection of marginal benefit and cost curves, we use financially unconstrained and financially undistressed firms as our estimation sample. In order to identify the marginal cost function, we need to assume that all of the variation in debt level occurs due to benefit function shifts while the cost function remains fixed in one place. By using instrumental variable that proxy for benefit shifts and that are uncorrelated with cost shifts, we can identify the cost curve. In this paper, we use annual corporate top tax rates as our identifying instrument. We use 2 stage least square method and instrument variable to estimate cost functions of debt. The location of the cost functions varies with firm characteristics such as asset collateral, firm size, book-to-market ratio, profitability, dividend dummy, financial distress, profitability, and earnings volatility. We use the estimated coefficient from cost functions to compute a cost of debt for any firm. These estimated marginal cost functions should capture all possible ex ante costs of debt that managers trade off against tax benefit of debt as they choose their equilibrium capital structure. These costs includes financial distress costs, agency costs, and so on. We use these benefit and cost functions of debt to find out a firm-specific equilibrium level of debt that a firm should use. Firm-specific equilibrium capital structure occurs where the marginal benefit equals the marginal cost of debt. We also integrate the area between benefit and cost functions of debt to estimate the net tax benefit of debt financing, and similarly estimate the cost of deviating from the equilibrium. Prior research suggests that Korean firms seem to do not fully utilize the tax benefits of debt, particularly after the 1997 Asian financial crisis, however, considering that expected default costs offset the majority of the potential tax savings, Korean firms` conservative use of debt after the financial crisis period could indeed be a rational response to a new business environment. Based on this prior research, we analyse equilibrium level of debt and net tax benefit of debt by dividing the sample period before and after the financial crisis. We use two taxable income forecasting method, the well known random walk approach(Graham 2000) and nonparametric approach(Blouin et al. 2010), to estimate simulated marginal tax rates. Because the results are qualitatively similar, we use random walk approach as a main analysis. In our analysis, we use Korean listed firms with available Korea Investors Services (KIS)-Vaule data from 1991 to 2011. We restrict the sample to non-financial, non-utility firms and have a December year-end. Additionally, firm-years with missing financial data and firm-years which can not estimate marginal tax rates are excluded. We delete observations in which any continuous variable lies beyond the highest and lowest 1% of the distribution for that variable to control for outliers. These data requirements result in a sample of 8,728 firm-year observations which consist of 2,778 observations before the financial crisis(1991-1998) and 5,950 observations after the financial crisis(1999-2011). The result shows that observed debt level(we normalize interest expense over book asset) is 0.047 before the financial crisis and 0.017 after the financial crisis, resulting in only 36% after the financial crisis. Before the financial crisis, marginal benefit and marginal cost of debt on the observed debt level are estimated as 0.232 and 0.265 respectively which means Korean firms use more debt than equilibrium level of debt on average. Equilibrium debt level before the financial crisis is estimated to be 80%-100% of observed debt level. After the financial crisis, marginal benefit and marginal cost of debt on the observed debt level are estimated as 0.201 and 0.205 respectively which means on average Korean firms use debt in equilibrium level. By integrating the area between the benefit and cost functions, we estimate the observed and equilibrium net benefit of debt. Before the financial crisis, the equilibrium total benefit, total cost and net benefit are 8.4%, 5.0%, and 3.4% of total assets respectively, and the observed level of total benefit, total cost and net benefit are 8.9%, 6.3%, and 2.6% of total assets respectively. After the financial crisis, the equilibrium total benefit, total cost and net benefit are 6.1%, 4.2%, and 1.9% of total assets respectively, and the observed level of total benefit, total cost and net benefit are 5.4%, 4.5%, and 0.9% of total assets respectively. As a result, the observed and equilibrium net benefit of debt are reduced after the financial crisis. The reduced net tax benefit is mainly due to the decrease in marginal benefit resulting from corporate tax rates reduction and marginal cost increase in debt usage after the financial crisis. Also, due to conservative debt usage, decrease in total benefit of debt is more than offset the increase in total cost of debt after the financial crisis. Regardless of the period, the cost of being overlevered is asymmetrically higher than the cost of being underlevered. This paper contributes to the corporate capital structure study by providing an additional evidence of implementing firm-specific marginal cost function and thereby providing framework to recommend equilibrium level of debt. The framework to estimate benefit and cost functions of debt and to find out firm-specific equilibrium level of debt also can be helpful to explore and rebalance firm-specific optimal capital structure. The findings that Korean firms` conservatism in debt usage after the financial crisis period are mainly due to the decrease in marginal debt benefit resulting from corporate tax rates reduction and increase in marginal debt cost which represents various firm characteristics are helpful to understand debt policy in Korean firms.