10 / 372
審計人員之產業專精與客戶租稅規避：中國實證研究
10
3.2 Measures of Tax Avoidance
Consistent with Dyreng et al. (2008, 2010), this study defines tax avoidance as a
strategy that reduces a firm’s tax liabilities. To proxy for firms’ tax avoidance activities, this
study first estimates firms’ book and cash effective tax rates:
BETR
= current income tax expense/pre-tax accounting income less special items;
CETR
= cash taxes paid/pre-tax accounting income less special items.
The book effective tax rate,
BETR
, is defined as total tax expense divided by pre-tax
book income less special items over a one-year period.
8
BETR
is a widely-used measure of a
firm’s tax burden and reflects tax avoidance activities that directly affect net income, but not
those activities that defer cash taxes paid to a later period (Hanlon and Heitzman, 2010). A
loIr value of
BETR
can reflect an increased level of tax avoidance (e.g., Rego, 2003). This
study measures
CETR
(Cash Effective Tax Rate) over a one-year period and defines it as
cash taxes paid divided by pre-tax book income less special items (Dyreng et al., 2008,
2010). Unlike
BETR
,
CETR
can reflect tax avoidance strategies that defer cash taxes paid to
later periods, but do not affect the tax expense on the financial statement. Like
BETR
, lower
values of
CETR
represent higher levels of tax avoidance.
This study then estimates book-tax differences (
BTD
) following the method developed
by Wilson (2009):
BTD
= pre-tax book income less special items-taxable income.
where taxable income is defined as current income tax expense divided by statutory tax
rate.
BTD
reflects tax avoidance activities that generate both permanent and temporary
differences between financial income and taxable income. Previous studies indicate that
BTD
is
positively associated with the probability of tax sheltering (Wilson, 2009). Accordingly, unlike
BETR
and
CETR
, larger values of
BTD
represent higher levels of tax avoidance.
3.3 Regression Models
Following prior tax avoidance model (Dyreng et al., 2008, 2010; Wilson, 2009), this
study examines the influence of auditor industry expertise on tax avoidance after controlling
for the effects of audit firm size, client size, client ownership type, client’s performance,
industry, and year. This study estimates the OLS regression model as follows:
8
BETRs
with negative denominators are deleted. The remaining non-missing
BETRs
are winsorized (reset)
at the 1st and 99th percentiles.