LFIN Seminar Eva Lütkebohmert-Holtz

Eva Lütkebohmert-Holtz (U. Friburg)

will give a presentation on

Deep Learning Name Concentration Risk for Portfolios of Multilateral Development Banks

Abstract:

As institutions that predominantly lend to sovereign borrowers to achieve their development goals, Multilateral Development Banks’ (MDBs) loan portfolios typically consist of a small number of borrowers. The approaches currently applied by rating agencies to account for this exposure to single name concentration risk in the assessment of MDBs’ capital adequacy have been criticized for being overly conservative and thereby restricting MDBs’ lending headroom. In this paper, we suggest a new deep learning approach for the quantification of single name concentrations in MDB portfolios. Based on simulated as well as stylized portfolios that adequately represent MDB portfolios, we demonstrate the accuracy of our new approach and its superior performance compared to existing analytical methods for assessing name concentration risk in small and concentrated portfolios.