Revue d'économie et de statistique appliquée
Volume 23, Numéro 1, Pages 24-36
2026-06-28
Modernizing Motor Insurance Pricing In Tunisia: A Generalized Linear Models Approach For Emerging Markets
Authors : Ben Nasr Mehrez .
Abstract
The Tunisian motor insurance market confronts critical actuarial challenges following seven years of frozen tariffs (2016-2023) despite persistent claims inflation and evolving risk landscapes. This empirical case study develops and validates a Generalized Linear Models (GLM) framework for risk-based motor insurance pricing within regulatory-constrained emerging market contexts. Using 148,505 motor insurance policies from a leading Tunisian insurer, we implement frequency-severity modeling: Poisson regression for claim frequency and Gamma regression for claim severity across 13 key risk variables. Key findings demonstrate substantial risk differentiation (87.2% coefficient of variation) between low-risk and high-risk profiles, enabling actuarially justified premium segmentation. Bonus-malus class exhibits 14.8-fold variation, driver age shows 3.64× elevation for young drivers, vehicle age follows U-shaped patterns, and fuel type generates 14% frequency differential. International comparison validates alignment with published literature while revealing Tunisian-specific nuances. Implementation projects 3-5 percentage point combined ratio improvements while maintaining regulatory compliance. This research contributes to African insurance modernization by demonstrating evidence-based GLM implementation in constrained emerging markets, providing a transferable framework for other African insurers seeking actuarial advancement.
Keywords
Generalized Linear Models, motor insurance, emerging markets, pricing, risk segmentation, frequency-severity, Africa, regulatory constraints, actuarial science
Les articles similaires
Khacef Djamel Eddine
.
pages 715-732.
Kalafate Nadia
.
Khiari Imen
.
pages 38-54.
Abdelhakim Necir
.
Brahimi Brahim
.
pages 03-15.
