odeling the Cumulative Impact of Information Asymmetry on Algerian Insurers' Turnover: A Panel ARDL Approach (2010-2024)

Résumé fourni par la source

This study aims to evaluate and analyze the impact of information asymmetry manifested through adverse selection and moral hazard on the premium turnover and operational efficiency of the Algerian insurance market during the period from 2010 to 2024. Utilizing a panel data approach, the study examines a sample of 14 domestic insurance companies, yielding 196 dynamic observations. Methodologically, after validating slope heterogeneity and cross-sectional dependence, the Panel ARDL approach (Pooled Mean Group estimator) was deployed to capture both short- and long-run equilibrium relationships. The econometric results reveal significant structural heterogeneity among insurers in their vulnerability to informational imbalances. Notably, the error correction term (COINTEQ) was estimated at -0.18, being both negative and statistically significant. This confirms a long-run equilibrium adjustment mechanism whereby approximately 18% of short-run disequilibrium is corrected in each time period. Long-run estimations demonstrate a powerful, statistically significant negative impact of the loss ratio (LR) on corporate turnover, whereby a 1% increase in the loss ratio induces a 1.54% structural contraction in turnover. This outcome underscores the long-term accumulation of claims triggered by fraudulent tendencies and unrationalized customer behaviors. Furthermore, the reinsurance ratio (RI) exerts a significant negative long-run impact of 0.28%, highlighting that while risk-transfer mitigates uncertainty, it simultaneously dilutes domestic premium retention. Based on these insights, the study recommends empowering regulatory bodies and the "Alfa Agency" to institute specialized behavioral laboratories to detect the 50% of uninvestigated fraudulent claims, alongside adopting risk-bundling mechanisms and dynamic pricing models to stabilize the market.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.