ABSTRACT
This chapter examines the application of multivariate models. The usual approach followed in developing multivariate - and, indeed, univariate - failure identification models is to pair bankrupt and surviving companies, matching by criteria such as year, industry membership and size. Various attempts have been made to try to capture information over a period of time prior to failure. One procedure is to develop a score based upon dependent variable characteristics over a number of years. 'Duration' models have long been used in engineering and medicine, when the focus of interest has been, for instance, how long a structure, engine or patient is going to survive. In the industrial economics literature there have been a number of studies which have examined the survival records of new businesses. In the UK the multiple regression analysis approach was used by D. A. J. Marais to develop a failure identification model for the Bank of England.
