Abstract
We analyze a unique dataset to test an empirical model of retail bank fee determinants in five Central European countries. Due to the data structure we can cope with heterogeneity and cross-subsidization by employing a representative fee index instead of using variables associated with individual fees.