Corporate Taxation and Multinational Activity

the subsequent discussion of Table 2, we summarize our findings regarding the
impact of the tax variables on bilateral outbound FDI in the OECD.

In general, we would expect that the parent country tax rate exerts a posi-
tive impact on the country’s outbound FDI.²³ This is in line with the positive
point estimate in the pooled sample and in the exemption sub-sample. For
instance, the point estimate in the full sample indicates that a one percent-
age point increase in the parent country statutory tax rate (ti,t-1) is associ-
ated with an increase in outbound FDI by about 0.85 percent. For the credit
method sub-sample, the impact of the parent country corporate tax rate is the-
oretically ambiguous, which is in line with our findings. The interactive term
ti,t-1 × ∆SKij,t-1 × I(∆SKij,t-1 > 0) enters as expected in the sub-sample cov-
ering the exemption countries, but not so in the full sample and in the credit
sub-sample. Regarding the host country corporate tax rate, we expect a nega-
tive impact on a country’s outbound FDI.²⁴ This is confirmed by the empirical
finding of a significantly negative estimate of the corresponding parameter. Sim-
ilarly, Table 1 clearly suggests a negative relationship between the host country
withholding tax rate and outbound FDI, which is confirmed by the signifi-
cantly negative parameter estimates in all samples. Finally, the parent country
depreciation allowances (δi,t-1) exert a significantly negative impact on foreign
outbound FDI, which is not supported by the simulations in Table 1. A negative
but insignificant effect is estimated for its host country counterpart (δj,t-1), as
expected. Further, the point estimate of δi,t-1 × ∆SKij,t-1 × I(∆SKij,t-1 > 0)
is negative, and the one of δj,t-1 × ∆SKij,t-1 × I(∆SKij,t-1 > 0) is positive,
being in line with the theoretical hypotheses. Across the board, the support
for the tax-related hypotheses is remarkable, considering the large number of
parameters to be estimated.

Sensitivity analysis and discussion: We check the sensitivity of our find-
ings in various ways. Table 3 provides a summary, pointing to their robustness
in qualitative terms. For the sake of brevity, we refer to the full sample in Table
from that one in Carr, Markusen, and Maskus (2001) and Markusen and Maskus (2002), who
use U.S. foreign affiliate sales across a larger set of host countries over time. Finally, we
employ country-pair and time fixed effects. In particular, this wipes out all time-invariant
level information so that the parameters of the relatively time-invariant endowment variables
can hardly be estimated significantly.

²³The predicted impact across all methods (i.e., the pooled sample) is ambiguous (see Table
1). However, the number of cases in our sample, where deduction is applied or the excess credit
position is evident is rather small. Accordingly, we expect the positive impact to dominate in
the pooled sample.

²⁴Again, the prediction is ambiguous, but we expect the negative impact to dominate for
the same reasons as in the previous footnote.

24

<<   23   24   25   26   27   28   29   >>

More intriguing information