Research and teaching are complements in terms of quality

Are good teachers also good researchers? Does spending more time on research improve one's teaching? Or does research get in the way of good teaching performance? It seems everyone has his own theory about this, some thinking that teaching and research are substitutes (you hear that mostly in teaching colleges) or complements (research universities think that). What about some empirical evidence on the matter?

Aurora García-Gallego, Nikolaso Georgantzís, Joan Martín-Montaner and Teodosio Pérez-Amaral provide this for a Spanish university. It turns out research and teaching are complements, at least for that university. What I find striking is that the professors who perform the most research also teach the most, and do it better. Those doing no research are among the worst teachers. That seems like the way a social planner would do it: Get the best people to work their ass off, while the worst should be kept away from anything productive. That is the way to increase productivity. That does not seem very fair, though, unless pay is tied to performance, which is not likely in this case. And how this study can generalize to other universities is not clear, especially as we do not know much about the university in this case.