Despite a stream of information from recent research, as well as analytical interpretations, we still lack a general picture of pre-1914 multinational banking based on a unifying empirical approach comparable to that recently developed by financial economists for the wave of banks' multinational expansion in the late twentieth century. The main purpose of this article, based on a unique dataset covering the foreign branches of British, French and German banks from 1880 to 1913, is to test in a similar theoretically-driven fashion the determinants of multinational banking during the first globalisation. The empirical strategy is based on an augmented gravity model in which geography and institutions interact with economic factors in determining the patterns of multinational banking. Contrary to what has emerged from recent studies on present multinational banking, I find that the ‘fundamentals’ of pre-1914 foreign branching cannot easily be fitted into a gravity-like model, and that no unifying pattern of foreign expansion can be inferred from the data.