This paper investigates the factors of the banking stability in Haiti, over the period of 1996 to 2017, using macroeconomic, government and institutions, banking system, and economic freedom factors measured respectively by GDP growth and exchange rate, political stability index and regulatory quality index, bank lending-deposit interest rate spread, property rights index and investment freedom index. To carry out the analysis, the yearly data have been transformed into quarterly data, giving a sample of 88 observations. By means of OLS regressions, six statistical models have been specified. Banking stability which is the response variable is measured by the z-score. The results suggest that macroeconomic and economic freedom factors have positive effects on the banking stability, while the banking system factor impacts negatively the banking stability in Haiti. Conversely, government and institutions factor has no significant impact on the Haitian banking stability. When it comes to assess the impact of each explanatory variable (GDP growth, exchange rate, political stability index, regulatory quality index, bank lendingdeposit interest rate spread, property rights index and investment freedom index) on banking stability, the results show that they all have significant effects on the Haitian banking stability. However, when all of the independent variables are analyzed in one multiple regression together, the political stability index is not statistically significant. The findings of this study have important implications for decision makers. Governments and the Central Bank should intensity their efforts in creating a promising macroeconomic environment, adopting effective monetary policy, reducing restrictions in investment and reinforcing laws to protect property rights, in order to maintain or improve banking stability in Haiti.
Psychometric testing is claimed to be a powerful innovation in credit scoring. Pioneered by the Entrepreneurial Financial Lab (EFL), this technique would enhance credit decisions by screening out high-risk applicants. This paper aims to evaluate the predictive power of the EFL's psychometric credit scoring model in microfinance through evidence from Sogesol, a Haitian microfinance institution. This evaluation has been conducted at two different levels: 1) A sample of clients has been selected from Sogesol's database to carry out a back test of the EFL tool, using performance metrics such as the Kolmogorov-Smirnov (K-S) statistic, the area under the ROC curve (AUC) in comparison with the existing socio-demographic model in use at Sogesol; 2) We conduct an analysis of causality between the quality of the portfolio and the credit decisions made based on the EFL tool and/or the traditional credit scoring model through the estimation of a linear regression model. The results show that the psychometric credit scoring model would present low predictive power in terms of K-S and AUC. However, the EFL tool would outperform the socio-demographic credit scoring model in use at Sogesol. The study further indicates that there would not be any statistically significant relationship between the risk level and the decision of granting a loan or not. The paper concludes that psychometric testing in its original format would not be efficient in the context of Sogesol's microcredit operations. Thus, the paper develops a new credit scoring model along traditional socioeconomic and behavioral lines, using logistic regression. This new model presents a better discriminatory power than the EFL tool, regarding K-S and AUC. In addition, it is well-calibrated, considering the results of Hosmer-Lemeshow (HL) test and the Brier score. If properly maintained and integrated into the client selection process, this new model could significantly improve credit risk management practices at Sogesol.
In recent years, the expansion of Fintech has speeded the development of the online peer-to-peer lending market, offering a huge opportunity for investment by directly connecting borrowers to lenders, without traditional financial intermediaries. This innovative approach is though accompanied by increasing default risk since the information asymmetry tends to rise with online businesses. This paper aimed to predict the probability of default of the borrower, using data from the LendingClub, the leading American online peer-to-peer lending platform. For this purpose, three machine learning methods were employed: logistic regression, random forest and neural network. Prior to the scoring models building, the LendingClub model was assessed, using the grades attributed to the borrowers in the dataset. The results indicated that the LendingClub model showed low performance with an AUC of 0.67, whereas the logistic regression (0.9), the random forest (0.9) and the neural network (0.93) displayed better predictive power. It stands out that the neural network classifier outperformed the other models with the highest AUC. No difference was noted in their respective accuracy value which was 0.9. Besides, in order to enhance their investment decision, investors might take into consideration the relationship between some variables and the likelihood of default. For instance, the higher the loan amounts, the higher the likelihood of default. The higher the debt to income, the higher the likelihood of default. While the higher the annual income, the lower the probability of default. The probability of default has a tendency to decline as the number of total open accounts rises.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.