15 /304
15
臺大管理論叢
第
28
卷第
1
期
value of equity in year t, scaled by book value of equity in year t; financial leverage
LEV
i,t
,
measured as long term debt for year t scaled by total assets in t-1;
ROA
i,t
the return on
assets for firm i, year t, measured as the ratio of pre-tax income for year t to beginning of
the year total assets;
NOL
i,t
indicator variable coded as one if net income from continuing
operations in year t is smaller than zero and zero otherwise; ∆
NOI
i,t
the change in net
income from continuing operations in year t, scaled by total assets in t-1);
STD_ROA
i,t
the
standard deviation of the previous five years’ return on assets; ∆
SALES
i,t
change in sales,
measured as the annual percentage change in net sales in year t;
CASH
i,t
the level of firm’s
cash and cash equivalent holdings;
INTAN
i,t
intangible assets for year t scaled by beginning
of the year total assets);
RD
i,t
is the total research and development expenditure in year t
divided by beginning of the year total assets;
PPE
i,t
the property, plant, and equipment in
year t divided by beginning of the year total assets;
ADV
i,t
advertising expense in year t
divided by net sale in year t; when missing, reset to 0;
SGA
i,t
is the selling, general, and
administrative expense in year t divided by net sales in year t; missing values of SGandA
are set to 0;
NUM_INVESTEE
i,t
the number of investees. The variable definitions are also
presented in Appendix 1.
To support our H1, we expect a negative coefficient,
β
1
on
LAYER
i,t
if pyramidal firms
with more number of layers engage in more tax avoidance than those with fewer number
of layers. In model (1), we control for firm characteristics that could be associated with
ETRs drawn from tax avoidance literature. We do not have a sign prediction for
SIZE
it
because of the lack of consensus. Dyreng et al. (2008) find that smaller firms have higher
tax rates, but Rego (2003) finds that larger firms are associated with higher worldwide
ETRs.
6
We expect growth firms have lower tax rate because they may make more
investments in assets that can generate temporary difference in the recognition of expenses
or permanent difference in tax credits (Chen et al., 2010). We also do not have sign
expectation for
LEV
because highly leveraged firms may have more ability to minimize
tax payment through interest expense deduction, leading to lower effective tax rate (Mills
et al., 1998; Stickney and McGee, 1982; Dyreng et al., 2008), however, they may have
less need for other non-debt tax deduction (Graham and Tucker, 2006).
6 This finding is consistent with the political cost theory, stating that the higher visibility of larger firms
causes them to face stricter regulatory scrutiny and wealth transfers (Watts and Zimmerman, 1986). In
contrast, it can be argued that larger firms have greater incentives and more resources to influence the
political process in favor for them and organize their activities to achieve optical tax savings.