The aim of this thesis is to shed light on German banks' foreign lending, which has grown substantially over the past two decades and forms an integral part of worldwide capital flows at present. For this purpose, a new micro data-set on German foreign bank lending from Deutsche Bundesbank is used. The thesis begins by providing background information on Germany's banking sector and the importance of foreign bank lending for different banking groups. It turns out that mainly Big Banks and Landesbanks are engaged in foreign lending. As a consequence, the data-set, which is described in detail, focuses on those banking groups. In the first empirical analysis, I study the major determinants of German foreign claims for the period 1996 to 2002. I find that market size, distance from Germany, German companies involvement abroad as well as debtor countries' risk rating altogether account for about 80 percent of the variation in German lending. Furthermore, I investigate the discrepancy in bank lending towards industrial and non-industrial countries. My results suggest that this discrepancy is largely explained by the groups' differences in fundamentals and only to a minor extend by banks' different treatment of these fundamentals. Hence, it appears that German banks do not discriminate between Industrial and Non-industrial countries making improvements in countries' fundamentals such as reducing political and economic uncertainties even more worthwhile. In the second empirical analysis, I study bank lending in crisis periods. In particular, I test for German banks' role in transmitting financial crises between developing countries. The findings suggest that German banks contributed to financial contagion in the Asian as well as the Russian crisis. This result supports earlier evidence that developing countries should diversify their sources of funding to be less prone to financial contagion. With regard to consequences for banks, the findings suggest that it does not suffice to assess the credit risks of each country individually. In particular, it is essential to assess the mix of investors exposed to a crisis country. Since individual banks do not have full access to individual data of other banks, the development of such early warning system models is also an important issue for bank supervisors.