Testing the systemic risk differences in banks

This paper contains a testing framework for the reliability of systemic risk measurement of banks, using the three leading market-based measures of systemic risk. We test whether the difference within the same category and across dfferent categories of systemic risk of individual banks is signifcant. We find that in general the systemic risk categories defined by the Financial Stability Board are dfferent from those constructed in a full pairwise comparison approach based on the market measures. Moreover, these dfferences were more pronounced during episodes of high market turbulence.To account for model risk we introduce a more robust ranking method based on nonparametric confidence intervals. We show that there is a large number of banks with overlapping confidence intervals of their market-based systemic risk measures.Further, similarity measures indicate that the scoring based rankings are not perfectly aligned with rankings produced by market based systemic risk measures.