Predicting Unlisted SMEs’ Default: Incorporating Market Information on Accounting-Based Models for Improved Accuracy

Posted: 17 Apr 2018 Last revised: 3 Jun 2019

See all articles by Panagiotis Andrikopoulos

Panagiotis Andrikopoulos

Centre for Financial and Corporate Integrity (CFCI), Coventry University

Amir Khorasgani

Coventry University

Date Written: March 3, 2018

Abstract

The risk associated with lending to small businesses has become more important since regulations started obliging banks to use separate procedures in assessing SMEs' credit worthiness. However, current accounting-based models for SMEs do not account for the impact of market information on default prediction. We fill this gap in the literature by introducing a hybrid default prediction model for unlisted SMEs that uses market information of listed SMEs (comparable approach) alongside existing accounting information of unlisted SMEs. Our results suggest that the accuracy of this default prediction modelling approach in the hold-out sample, during the period of the financial crisis 2007-09 and for the entire sample-period, improves considerably. We conclude that the proposed hybrid model is a good replacement for existing standard accounting-based methods on SMEs' default prediction.

Keywords: SMEs Finance, Merton-KMV Model, Default Prediction, Market-Based Factors, Accounting-Based Factors

JEL Classification: G33, G32, G17

Suggested Citation

Andrikopoulos, Panagiotis and Khorasgani, Amir, Predicting Unlisted SMEs’ Default: Incorporating Market Information on Accounting-Based Models for Improved Accuracy (March 3, 2018). British Accounting Review, Vol. 50, Issue 5, p. 559-573, DOI: 10.1016/j.bar.2018.02.003. Available at SSRN: https://ssrn.com/abstract=3153922

Panagiotis Andrikopoulos (Contact Author)

Centre for Financial and Corporate Integrity (CFCI), Coventry University ( email )

Priory Street
Coventry, CV1 5FB
United Kingdom
+44(0)247 765 7920 (Phone)

Amir Khorasgani

Coventry University ( email )

Priory Street
Coventry, CV1 5FB
United Kingdom

Register to save articles to
your library

Register

Paper statistics

Abstract Views
593
PlumX Metrics