Table 2
Instrumental-variables models of countries’ research impact and innovation
| Model (1): Research impact | Model (2): Innovation | |
| Contribution of international collaborators | 68.272 (158.432) | 8.109 (11.384) |
| No of articles per organisation | 33.668 (48.883) | −2.363 (4.174) |
| Efficiency | 211.320 (63.625) | 18.819 (6.108) |
| Geographical entropy | 244.527 (55.760) | 30.850 (4.641) |
| Constant | 74.600 (106.865) | 7.116 (10.900) |
| No observations | 56 | 56 |
| Centred R2 (uncentred R2) | 0.572 (0.614) | 0.519 (0.587) |
Root MSE ( ) | 184.4 | 17.63 |
| F(10,45) | 3.26 | 7.17 |
The models are efficient GMM instrumental-variables estimators with SEs robust to heteroskedasticity. Robust SEs of parameter estimates are within parentheses. Reported values of the square root of the MSE refer to the estimated standard deviation
of the idiosyncratic disturbance. Estimated parameters in boldface are statistically significant at the 1% level (p<0.01).GMM, generalised methods-of-moments; MSE, mean squared error.
