Insurers' Insolvency Prediction Using Random Forest Classification

Posted: 8 Dec 2013 Last revised: 22 Feb 2015

See all articles by Anastasia V. Kartasheva

Anastasia V. Kartasheva

University of St. Gallen - I.VW-HSG; Joshua J. Harris Alternative Investment Program; Swiss Finance Institute

Mikhail Traskin

Independent

Date Written: December 7, 2013

Abstract

This paper uses a modification of the Random Forest classification algorithm to predict insolvency of insurers. RF orders companies according to their propensity to default. We show that RF methodology delivers higher quality of prediction compared to other existing methods. In addition, RF classification can be used to gather further insights about the fragile companies. It ranks the explanatory variables in the order of their ability to predict insolvency. Also it is used to describe the relationship between the propensity to default and the individual characteristics of an insurer. We show that many of these relationships are highly non-linear.

Keywords: property-casualty insurance, insolvency prediction, Random Forest Classification

JEL Classification: C14, C44, G17, G22, G32

Suggested Citation

Kartasheva, Anastasia V. and Traskin, Mikhail, Insurers' Insolvency Prediction Using Random Forest Classification (December 7, 2013). Available at SSRN: https://ssrn.com/abstract=2364736

Anastasia V. Kartasheva (Contact Author)

University of St. Gallen - I.VW-HSG ( email )

Kirchlistrasse 2
St. Gallen, 9010
Switzerland

Joshua J. Harris Alternative Investment Program ( email )

The Wharton School
3620 Locust Walk
Philadelphia, PA 19104
United States

Swiss Finance Institute ( email )

c/o University of Geneva
40, Bd du Pont-d'Arve
CH-1211 Geneva 4
Switzerland

Mikhail Traskin

Independent ( email )

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Abstract Views
744
PlumX Metrics