Non-Parametric and Flexible Time Series Estimators

Abstract

The estimation of the probability of default based on information on the individual customer or the company is an important part of credit screening, i.e., judging the credit standing. It is essential for the establishment of a rating or for measuring credit risk to estimate the probability that a company will end in financial difficulties within a given period, for example, one year. Also, here nonparametric applications prove to be flexible tools in estimating the desired default probability without arbitrary assumptions. In this chapter we will give a brief overview of the various approaches for non- and semiparametric estimates of conditional probabilities.