Table 3: Robustness
|
Tax components |
Model A |
Model B |
Model C |
Model D |
Model E |
Model F |
Model G |
Model H |
Model I |
|
Parent country corporate tax rate: ⅞,*-ι |
045 |
CÔ8Ï7 |
LÏ7Ô |
Гпз |
СЫ46 |
CË5Ï7 |
ЙЁ262 |
CË7Ô2 |
СШ6 |
|
(1.80) |
(0.87) |
(2.21) |
(2.41) |
(0.80) |
(1.15) |
(-0.36) |
(1.49) |
(1.71) | |
|
Host country corporate tax rate: ij,t-ι |
-1.498 |
-2.161 |
-1.716 |
-1.302 |
-0.892 |
-1.136 |
-1.402 |
-1.670 |
-1.626 |
|
(-2.57) |
(-1.91) |
(-2.57) |
(-2.21) |
(-1.27) |
(-2.10) |
(-1.66) |
(-2.76) |
(-2.74) | |
|
Withholding tax rate on repatriated profits: tj,t-1 |
-1.545 |
-2.439 |
-2.415 |
-1.359 |
-1.354 |
-1.606 |
-1.554 |
-1.591 |
-1.453 |
|
(-3.06) |
(-1.80) |
(-3.33) |
(-2.79) |
(-2.32) |
(-3.24) |
(-2.14) |
(-3.22) |
(2.93) | |
|
Parent country depreciation allowances: <5⅛,t-1 |
-4.530 |
-4.869 |
-4.452 |
-4.294 |
-3.360 |
-5.106 |
-4.296 |
-3.617 |
-3.848 |
|
(-4.74) |
(-2.28) |
(-3.77) |
(-4.45) |
(-2.93) |
(-4.91) |
(-2.04) |
(-3.68) |
(-3.99) | |
|
Host country depreciation allowances: |
0.201 |
-1.176 |
0.067 |
-0.229 |
0.562 |
-0.721 |
-1.620 |
-0.355 |
-0.324 |
|
(0.29) |
(-1.19) |
(0.10) |
(-0.34) |
(0.72) |
(-0.98) |
(-1.22) |
(-0.52) |
(-0.46) | |
|
i≈-,t-ι × ΔS7⅛,t-ι × I(ΔS7⅛.t-ι > 0) |
-0.001 |
0.164 |
1.077 |
0.079 |
0.063 |
0.034 |
0.044 |
-0.019 |
0.004 |
|
(-0.02) |
(2.48) |
(0.84) |
(2.22) |
(0.85) |
(0.85) |
(0.96) |
(-0.47) |
(0.09) | |
|
<5i,t-ι × ΔSΛy ,-ι × I(ΔSΛy ,~ι > 0) |
0.005 |
-0.073 |
0.698 |
0.018 |
-0.113 |
-0.063 |
-0.188 |
-0.109 |
-0.082 |
|
(0.09) |
(-0.56) |
(0.53) |
(0.54) |
(-0.98) |
(-0.79) |
(-1.42) |
(-1.41) |
(-1.07) | |
|
⅞,t-ι × ΔS7⅛.t-ι × I(ΔS7⅛.t-ι > 0) |
0.006 |
-0.007 |
-0.465 |
-0.093 |
-0.375 |
-0.000 |
-0.110 |
0.069 |
0.050 |
|
(°-15) |
(-0.06) |
(-0.59) |
(-2.33) |
(-1∙52) |
(-θ-θɪ) |
(-1∙49) |
_ (1∙9°) |
_ (1∙36) _ |
Notes : H AC-corrected t-values in parentheses (Newey and West, 1987).
Model A (2195 observations): Secondary school enrollment instead of tertiary school enrollment.
Model B (1011 observations): Skill measure as proposed by Markusen (2002: p. 228) rather than tertiary school enrollment.
Model C (1410 observations): Capital stock per worker instead of tertiary school enrollment.
Model D (2195 observations): Gravity model instead of Knowledge Capital model.
Model E (1792 observations): Dynamic model as proposed by Arellano and Bond (1991) for panel data; Lagged dependent variable amounts to 0.221 (s.e. = 0.110).
Model F (2011 observations): Excluding transition economies (Czech Republic, Hungary and Poland).
Model G (846 observations): Excluding transition economies and non-EU members. 12 remaining EU economies.
Model H: р-value of F-test on joint significance of all third-country effects: 0.000; р-value of F-test on joint significance of all third-country effects of profit taxation: 0.006.
Model I: p-value of F-test on joint significance of all third-country effects: 0.000; p-value of F-test on joint significance of all third-country effects of profit taxation: 0.008.
All sensitivity checks should be compared to the full sample results in Table 2.