Table 5.
Support of Hypothesis 1: Robustness tests 
Panel least squares regression results on GDP growth
Panel least squares regression results on GDP growth| Omit Japan & Sweden 20 countries | Omit Singapore & Canada 20 countries | Omit four outliers 18 countries | |
|---|---|---|---|
| Intercept | 0.2038 | 0.1901 | 0.2244 |
| (4.0200)*** | (3.8373)*** | (4.1036)*** | |
| LTIREM(−1) | 0.0753 | 0.0744 | 0.0873 |
| (2.9510)*** | (3.4190)*** | (3.1862)*** | |
| INVRATIO | 0.1361 | 0.1275 | 0.1365 |
| (2.3276)** | (1.9903)** | (2.1679)** | |
| LEGAL | −0.0020 | 0.0014 | 0.0002 |
| (−0.6279) | (0.4090) | (0.0688) | |
| LABOR | 0.0018 | 0.0006 | 0.0009 |
| (0.7077) | (0.2440) | (0.3356) | |
| LN(EDUCATION) | −0.0958 | −0.0896 | −0.1079 |
| (−3.4665)*** | (−3.2931)*** | (−3.6089)*** | |
| INDIVCOLLECT | −0.0007 | −0.0009 | −0.0009 |
| (−3.1741)*** | (−3.2590)*** | (−3.4002)*** | |
| No. of panel observations | 100 | 100 | 90 |
| Adjusted R2 | 0.176 | 0.147 | 0.165 |
| F-statistic | 4.534*** | 3.844*** | 3.932*** |
| Durbin Watson statistic | 1.375*** | 1.472** | 1.417** |
| Omit Japan & Sweden 20 countries | Omit Singapore & Canada 20 countries | Omit four outliers 18 countries | |
|---|---|---|---|
| Intercept | 0.2038 | 0.1901 | 0.2244 |
| (4.0200)*** | (3.8373)*** | (4.1036)*** | |
| LTIREM(−1) | 0.0753 | 0.0744 | 0.0873 |
| (2.9510)*** | (3.4190)*** | (3.1862)*** | |
| INVRATIO | 0.1361 | 0.1275 | 0.1365 |
| (2.3276)** | (1.9903)** | (2.1679)** | |
| LEGAL | −0.0020 | 0.0014 | 0.0002 |
| (−0.6279) | (0.4090) | (0.0688) | |
| LABOR | 0.0018 | 0.0006 | 0.0009 |
| (0.7077) | (0.2440) | (0.3356) | |
| LN(EDUCATION) | −0.0958 | −0.0896 | −0.1079 |
| (−3.4665)*** | (−3.2931)*** | (−3.6089)*** | |
| INDIVCOLLECT | −0.0007 | −0.0009 | −0.0009 |
| (−3.1741)*** | (−3.2590)*** | (−3.4002)*** | |
| No. of panel observations | 100 | 100 | 90 |
| Adjusted R2 | 0.176 | 0.147 | 0.165 |
| F-statistic | 4.534*** | 3.844*** | 3.932*** |
| Durbin Watson statistic | 1.375*** | 1.472** | 1.417** |
Notes:This table presents the panel least squares regression results using random effects for 18 or 20 of the original 22 countries as indicated. All regressions are run with five years of data: 2001–05 if lagged, and 2002–06 if not lagged. Numbers in parentheses are t-statistics.
**Statistically significant at α = 5 percent level; ***statistically significant at α = 1 percent level. The cross-section random effects are not reported as they sum to zero and are not relevant to the discussions.