Behavioral technology credit scoring model with time-dependent covariates for stress test

Yonghan Ju, Song Yi Jeon, So Young Sohn

Research output: Contribution to journalArticlepeer-review

9 Citations (Scopus)


Technology based loan default is related not only to technology-oriented attributes (management, technology, profitability and marketability), and firm-specific characteristics but also to the economic situation after the loan. However, the default phenomenon for technology based loan has not reflected the change of economic situation. We propose a framework of utilizing a time varying Cox hazard proportional model in the context of technology based credit scoring. The proposed model is used for stress test with various scenarios of lending portfolio and economic situations. The results indicate that the firms with higher management score than average have the lower loan default rates than the firms with higher profitability or marketability score than average due to the effect of manager's knowledge and experience and fund supply ability when they are exposed under the same economic condition. In scenario test, we found the highest default rate under stable exchange rate with high consumer price index. Moreover, firms with a high level of marketability factors turn out to be significantly affected by economic conditions in terms of technology credit risk. We expect the result of this study can provide valuable feedback for the management of technology credit fund for SMEs.

Original languageEnglish
Pages (from-to)910-919
Number of pages10
JournalEuropean Journal of Operational Research
Issue number3
Publication statusPublished - 2015 May 1

Bibliographical note

Publisher Copyright:
© 2014 Elsevier B.V. All rights reserved.

All Science Journal Classification (ASJC) codes

  • Computer Science(all)
  • Modelling and Simulation
  • Management Science and Operations Research
  • Information Systems and Management


Dive into the research topics of 'Behavioral technology credit scoring model with time-dependent covariates for stress test'. Together they form a unique fingerprint.

Cite this