Determinants of Long-term Economic Development: An Empirical Cross-country Study Involving Rough Sets Theory and Rule Induction

Empirical findings on determinants of long-term economic growth are numerous, sometimes inconsistent, highly exciting and still incomplete. The empirical analysis was almost exclusively carried out by standard econometrics. This study compares results gained by cross-country regressions as reported in the literature with those gained by the rough sets theory and rule induction. The main advantages of using rough sets are being able to classify classes and to discretize. Thus, we do not have to deal with distributional, independence, (log-)linearity, and many other assumptions, but can keep the data as they are. The main difference between regression results and rough sets is that most education and human capital indicators can be labeled as robust attributes. In addition, we find that political indicators enter in a non-linear fashion with respect to growth.