Psychometric Credit Scoring with Bayesian Logistic Models

Summary

The document asks whether psychometric or behavioral assessments can help evaluate loan applicants who lack conventional credit histories, particularly in developing countries. It sketches an approach in which questionnaire responses or psychological factors would be reduced to useful predictors and modeled with Bayesian logistic regression to estimate creditworthiness. The motivation is to consider alternative applicant information where financial ratios or established credit records may be unavailable.

The text points to work on psychometrics and poverty reduction and to a technical note describing an Enterprise Finance Lab modeling methodology. It reports that the cited material claims a Bayesian hierarchical model improved prediction accuracy and out-of-sample stability in an active country, with similar findings elsewhere. However, this document is primarily a request for literature and explanation rather than a full study: it supplies no model specification, sample details, validation metrics, or independent evidence. The reported performance should therefore be understood as a claim in the referenced material, not a result demonstrated here.

Key ideas

• Psychometric and behavioral responses are proposed as predictors for applicants without conventional credit histories.
• The proposed method uses Bayesian logistic modeling to estimate loan creditworthiness.
• The author is interested in applications in developing countries, including Haiti and Peru.
• A referenced methodology describes a Bayesian hierarchical model and claims improved predictive stability.
• The document does not present enough study details to independently assess those claims.

Tags

# Bayesian logit model in Psychometric or Behavioural Testing for Credit Scoring in Developing Countries


# Bayesian logit model in Psychometric or Behavioural Testing for Credit Scoring in Developing Countries












A lot of parameters in one title, I know. So there's credit scoring but not using credit history. Then there's using a Bayesian logit model. Then there's doing so in a developing country such as Haiti or Peru.

Question: Does anyone know of any papers about psychometric or behavioural testing using Bayesian logit model to determine whether a loan applicant in a developing country would be creditworthy?

Background: How can psychometric or behavioural tests be used in banking to evaluate creditworthiness?

Something like they use psychological factors or questionnaires like in job interviews instead of financial factors such as financial ratios and then they have some methodology that reduces it to less factors then they make a regression model out of it or something but then they use Bayesian analysis or something there.

Edit: So far all I have found are Enterprising Psychometrics and Poverty Reduction

and Technical Note – EFL Modeling Methodology.

Last part:

> RESULTS Here we illustrate the improved performance from this modeling approach, using data from one active EFL country.

-

> Results from other countries show the same result: the EFL Bayesian Hierarchical Model produces more accurate predictions and more stability out-of-sample.

Reading this has already earned much appreciation.

Answering or commenting will merit much more.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.