摘要: This paper examines the degree to which failure of one bank would cause subsequent collapse other banks. Using unique data on interbank payment flows, magnitude bilateral federal funds exposures is quanti- fied. These are used simulate impact various scenarios, and risk contagion found be economically small. quantifies present in U.S. banking system. Unlike previous studies that infer indirectly by identifying common characteristics banks affected some event (e.g., third-world debt crisis, large failure), this study estimates directly examining containing complete universe transactions across such allows for straightforward simulation exercises demonstrate might arise from these exposures. The cost direct approach measuring clear. analyzed only incorporate transactions. Because severe limitations, among cannot examined a basis. As result this, an obvious criticism results follow may actually much higher or distributed particularly contagion-enhancing way. While it will argued make up substantial fraction unsecured credit exposures, must realize conclusions reached conditional set being ex- amined. That is, reported here accurate, yet poten- tially conservative. Despite caveat, employed measure has at least three important advantages. First foremost, expo- sures bilaterally. each bank's exposure every known. views expressed those author do not necessarily reflect Federal Reserve Bank Chicago appreciates helpful comments Allen Berger (the editor) two anonymous referees. Craig H. Furfine Economic Advisor, Chicago. E-mail: