臺大管理論叢 NTU Management Review VOL.29 NO.2

223 NTU Management Review Vol. 29 No. 2 Aug. 2019 According to Chen et al. (2010), CETR is defined as cash taxes paid divided by pretax income. This study considers tax avoidance as a measure of a firm’s ability to pay a low amount of cash income taxes relative to its earnings. Therefore, a lower value of CETR reflects an increased level of tax avoidance. Our primary explanatory variables are LTTIE_XXX , including a firm’s number of board ties to low-tax firms, scaled by the firm’s total number of board members ( LTTIE_ BOA ); a firm’s number of executive director ties to low-tax firms, scaled by the firm’s total number of executive directors ( LTTIE_EXE ); whether a firm has board ties to low-tax members of its family ( LTTIE_FAM ); and whether a firm has network ties to low-tax firms engaging the same individual auditor ( LTTIE_CPA ). Low-tax firms are identified as those firms whose CETR ranked in the lowest quintile in their industry. Following prior studies, this study includes various firm characteristics to control for other determinants of tax avoidance. First, Khurana and Moser (2013) indicate that firms with higher levels of fixed assets tend to have higher non-debt tax shields in the form of higher depreciation deductions for tax purposes; therefore, this study includes capital intensity ( CAPINT ) as a control variable. Equity income from subsidiaries ( EQINC ) is included to account for the treatment difference between financial and tax reporting (Frank, Lynch, and Rego, 2009). Firm size ( SIZE ) and profitability ( PTROA ) are included as control variables because larger firms have access to tax-planning strategies that might exhibit economies of scale (Rego, 2003), and profitable firms are more likely to be engaged in tax avoidance behavior (Wilson, 2009). This study also controls for the debt ratio ( DEBT ) to account for debt tax shields (Graham and Tucker, 2006; Chen et al., 2010). Following Lisowsky (2010), this study uses the level of research and development expenditure scaled by total assets ( RD ) as a proxy for intangibles to accurately reflect the level of a firm’s self-created intangibles. Finally, two dummy variables ( IND and YEAR ) are added to control for industry effects and year effects. 3. Findings This study hypothesizes that tax-planning strategies and experiences are shared among firms through board interlocks. The main empirical results of this study support the view that a firm’s tax avoidance behavior is related to its directorate ties to low-tax firms. In addition, this study indicates that ties formed by executive directors are more