台灣地區颱風洪水保險需求之研究The Demand for Typhoon and Flood Insurance in Taiwan

摘要
本文測試國內在1999-2005年對颱風洪水保險需求的影響因素，透過縱橫資料、固定與隨機效果模型進行實證分析。結果顯示：當以保險金額為指標時，政府補助金額與保費顯著負向影響颱風洪水保險的需求，但所得則顯著正向影響保險需求。當以保險件數為指標時，政府補助與所得分別對保險需求具有顯著負向與正向影響，價格對保險需求的影響則不顯著。損失經驗在上述兩者模型皆不顯著影響。對解釋變數間關係的檢定而言，以保險金額為指標時，採隨機效果模型較適切，但以保單件數為指標時則以固定效果模型較佳。
關鍵字：颱風洪水保險、保險需求、固定效果模型、隨機效果模型
Abstract
Using a panel data set from 1999 to 2005, we apply fixed and random effect models to examine the determinants of typhoon and flood insurance purchasing decisions in Taiwan. When the amount of insurance in force is considered as an index of insurance demand, we find that the government relief expenditures and premiums have significantly negative effects on the demand for insurance while disposable personal income is positively related. When, in terms of the number of insurance policies, the results show that the government relief expenditures have negative effects and disposable personal income has positive effects on the demand for insurance. But insurance premium and loss experience play no role at all. Moreover, our results indicate that a random effect model is more appropriate for explaining the amount of insurance in force, while a fixed effect model is better (one) when the number of insurance policies are concerned.
Keywords: typhoon and flood insurance, insurance demand, fixed effect, random effect