Survey_BK_2006

Abstract

The study relies on the CFS Loan Data Set, collected under the Center for Financial Studies’ field research project on Credit Management (see Elsas et al. (1998) for further details). The data underlying our analysis include distressed corporate debtors of six major German banks, four of them are private listed companies, one is a public sector institution, and one is a cooperative bank. The unit of observation is a particular borrowing company or, more specifically, a particular bank-borrower relationship, using all the information regarding the borrowing company contained in the credit files6 of a bank. The data set contains in particular general characteristics of the company (e.g. legal form, industry),the company’s balance sheet data, an assessment of borrower risk, according to the bank’s internal risk rating, a complete account of the bank’s outstanding loans vis-à-vis the particular borrower, taken from its loan book, this includes information on loan terms, e.g. volume, maturity, collateral, spread, general information concerning other bank relationships of the company, including the existence of a bank pool, measures taken by the bank in order to reorganize or liquidate the company, or its assets. This information was collected directly from the banks’ credit files. Observations range from 1991 to 1999. The sample was drawn randomly from a population of all those corporate customers who fulfilled the following set of conditions at least once during 1992-1997. The resulting sample included 124 borrowers and a total number of 597 yearend observations8 . The incidence that a company entered distress in the last period of our observation window occurred for 23 companies. We dropped these observations since, for these companies, we have no information about distress behavior or reorganisation outcome. For an additional 6 companies the relevant balance sheet and loan information to be included in the regression analysis were missing. We therefore eliminated these 6 cases too, which left us with a sample of 95 companies and a total of 480 year-end observations.