The WiseCredit project, led by Ioana Florina Coita at the Faculty of Faculty of Economics and Finance of the Bratislava University of Economics and Business, examines how behavioural information, financial data and artificial intelligence can contribute to more transparent and fair creditworthiness assessment. Rather than treating credit risk as a static characteristic, the project studies how people make repayment decisions when their financial circumstances change.
We spoke with the project’s principal investigator about why behavioural experiments matter for credit-risk research, what WiseCredit has achieved so far, and what comes next.
What is the main idea behind WiseCredit?
Traditional credit scoring relies mainly on financial information such as income, debt and repayment history. These variables are essential, but they often provide only a snapshot of a borrower’s financial position. Two people with similar income and debt can respond very differently when they face an unexpected expense, a reduction in income or another financial shock.
WiseCredit investigates whether these behavioural differences can provide useful information for understanding repayment risk. The project combines behavioural experiments, psychometric measures, financial and open-banking information, and data-driven modelling. At the same time, we are interested in an equally important issue: even when additional behavioural information improves prediction, this does not automatically mean that it should become part of an operational credit score.
The objective is therefore not simply to use more information, but to identify which information genuinely improves creditworthiness assessment while preserving transparency, fairness and human oversight.
Why use behavioural experiments to study credit repayment?
Real-world financial data tell us what happened, but they do not always explain why it happened. Controlled experiments allow us to observe these differences under comparable financial conditions. In WiseCredit, participants make repeated repayment decisions while facing changes in income, expenses, liquidity constraints, financial shocks, incentives and feedback. Instead of focusing only on the amount of overdraft, we can examine whether overdraft dependence persists after a shock, how repayment changes and whether normal financial behaviour subsequently returns.
Can behavioural information improve credit scoring without increasing bias?
This is one of the central challenges of the project. Personality and behavioural information may help explain why people respond differently to the same financial circumstances, but this does not mean that every such variable should be used in a credit score.
Take impulsivity as an example. A simplistic approach would translate higher impulsivity directly into higher credit risk. A more informative approach is to examine whether it helps explain how individuals respond to financial stress—for example, whether repayment changes after an unexpected expense or whether financial recovery takes longer.
A personality measure can therefore have scientific value without being appropriate for an operational lending decision. If psychometric information adds little predictive value once actual repayment behaviour and financial transactions are taken into account, there may be no strong justification for including it in a final scoring model. Personality could instead help researchers understand behavioural heterogeneity while observable financial behaviour provides the information needed for prediction.
Experiments also provide an opportunity to examine potential bias. When participants face similar financial conditions, we can test whether comparable behaviour produces comparable model outcomes. Prediction errors, false positives and model calibration can be compared across relevant groups. Experimental design cannot guarantee an unbiased model, but it can make potential sources of bias easier to identify and understand before practical deployment.
What comes next for WiseCredit?
The first empirical results will examine how participants respond to changing financial conditions and whether variables provide useful information beyond conventional financial indicators.
The next stage will compare alternative model specifications to determine the additional value provided by behavioural responses and psychometric measures. The objective is not simply to identify the algorithm with the highest predictive accuracy. Models will also be assessed in terms of calibration, robustness and fairness.
This is increasingly important as artificial intelligence becomes more widely used in financial decision-making. Before behavioural and open-banking information becomes a routine component of credit assessment, we need evidence about what these data genuinely contribute and how they should be interpreted.
For WiseCredit, the central question remains: how can richer financial and behavioural information improve our understanding of credit risk and make credit decisions more transparent and fair?
Selected publications
Coita, I. F., Machado, M. R., Gomez Teijeiro, L., Wenzlaff, K., Gregoriades, A., Themistocleous, C., Bernard, F. S., Rupeika-Apoga, R., Teng, H.-W., Dzurovski, A., Yilmaz, G. N., Stanca, L., van Heeswijk, W., Pop, M. V., Bolesta, K., Peliova, J., Filipovska, O., & Osterrieder, J. (2026). Taxonomy of fraud types in alternative finance using hybrid systematic review. International Journal of Information Management Data Insights, 6, 100435. https://doi.org/10.1016/j.jjimei.2026.100435
Coita, I. F., Filip, L., Peliova, J., & Popa, D. (2026). Evaluation of a low-code intelligent invoice processing prototype. The Engineering Economist. Advance online publication. https://doi.org/10.1080/0013791X.2026.2668105
About WiseCredit
WiseCredit – Integrating Personality Traits and Open Banking Data for Sustainable and Ethical Creditworthiness Assessments investigates how behavioural characteristics, financial
information and data-driven methods can contribute to more transparent, fair and sustainable approaches to creditworthiness.
Principal Investigator: Ioana Florina Coita
Host institution: Bratislava University of Economics and Business, Faculty of Economics and Finance Project No.: 09I03-03-V04-00502
Funded by the EU NextGenerationEU through the Recovery and Resilience Plan for Slovakia under the project No. 09I03-03-V04-00502.
