A Simulation Study of Risk Measure and Dynamic Financial Analysis for Flood Insurance: An Example of Taiwan Keelung River District

洪水保險風險衡量與動態財務模擬分析：以基隆河流域住宅洪災為例

台灣近年廣受洪災損失威脅，本研究探討洪水保險保費擬定對保險公司經營洪水保險之影響。因洪災損失具巨災特性，本文目的在於考量尾端風險衡量指標做為洪水保險保費評估之依據，並以動態財務分析模擬比較尾端風險衡量指標和傳統期望值定價方法對公司財務的影響以及分析商品設計採自負額或再保險之影響。由於洪災損失具有地域性，本文以糠瑞林 (2005) 所發展之台灣洪災損失模型進行基隆河流域住宅洪災損失之模擬分析；動態財務分析之資產模型是依據台灣產險公司資產分佈來假設資產配置及資產動態過程。研究結果發現因洪災損失具厚尾分配性質，以期望值(Expectation) 做為保費標準易使保險公司產生清償問題，以尾端風險衡量指標（如CTE、TSD）衡量保費較能降低經營風險，且再保險機制可以顯著降低保險公司發行洪水保險之風險。本研究提供洪水保險保費評估之動態財務分析架構及盈餘評估結果，研究之架構及結果可提供政府與保險公司發行洪水保險之參考。

In recent years, Taiwan has been threatened by flood risk. This study attempts to evaluate the flood insurance premium in Taiwan. Particularly, we focus on the catastrophic feature of flood losses and apply three tail risk measures to evaluate the premium. Dynamic financial analysis is carried out to examine the effect of risk measure on the financial soundness of the insurer. The effects of the deductible and reinsurance are also investigated. We employ the domestic flood model proposed by Kang (2005) to simulate flood loss using the Keelung River District as an example. The asset model for dynamic financial analysis is built based on the asset allocation of a domestic property liability insurance company. We find that the probability of insolvency of insurers is high when they apply the expected value to price flood insurance. In order to reduce the risk, insurers have to apply CTE or TSD to measure the premium for flood insurance. Moreover, we find that the reinsurance mechanism could significantly reduce the insolvency risk. The evaluation framework and the simulated results in this paper can benefit the government or insurance companies who deal with flood insurance.