Integrating Quantile Regression and Extreme Value Theory for Robust Index-Based Agricultural Insurance

Index-based insurance is a critical tool for mitigating systemic climate risks in agriculture, yet its efficiency often suffers from significant basis risk. This study proposes a hybrid econometric framework combining Quantile Regression (QR) and Extreme Value Theory (EVT) to enhance risk assessment, using wheat yield data from Uzbekistan’s Surkhandarya region. Unlike conventional mean-based models, Quantile Regression reveals the asymmetric sensitivity of yields to environmental stressors, demonstrating that moisture deficits exert a significantly higher impact during extreme drought years (q=0.1) than under median conditions. To address catastrophic "tail risks," the Peak-Over-Threshold (POT) method is applied, providing a precise estimation of the Probable Maximum Loss (PML). The integration of these techniques allows for the calibration of more accurate insurance triggers and payout functions. This dual approach not only reduces basis risk but also provides the actuarial transparency required to transfer agricultural liabilities to international reinsurance markets, ensuring financial resilience against escalating climate volatility.