Simulation based stress tests of banks' regulatory capital adequacy

Abstract

Banks' holding of reasonable capital buffers in excess of minimum requirements could alleviate the procyclicality problem potentially exacerbated by the rating-sensitive capital charges of Basel II. Determining the sufficient buffer size is an important risk management task for banks, which the Basel Committee suggests should be approached via stress testing. We present here a simulation-based approach to stress testing of regulatory capital adequacy where rating transitions are conditioned on business-cycle phase, and which takes into account business-cycle dynamics. Our approach is an extension of a typical credit portfolio analysis in that we simulate actual bank capital and minimum capital requirements simultaneously. Actual bank capital (absent mark-to-market accounting) is driven by bank income and default losses, whereas capital requirements within Basel II are driven by rating transitions. The joint dynamics of these determine the necessary capital buffers, given a confidence level for regulatory capital adequacy chosen by bank management. We provide a tentative calibration of this confidence level to data on actual bank capital ratios, which enables a ceteris-paribus extrapolation of bank capital under the current regime to bank capital under Basel II.