風險值的計算近年來已成為金融機構風險管理的重要工具。而過去產險業於此的應用，在關鍵的承保風險上經常受限於資料統計的頻率，而大幅影響其估計的精確度。本研究利用佔產險業務最大宗的車險月損失資料來計算個公司的風險損失率，相對於過去文獻僅能使用年資料，本文發現實際核保風險損失率可能被低估。此外，再由進一步的橫斷面迴歸發現，各產險公司之核保風險損失率也可能受經營風險和自由現金流量等因素的影響。最後，本文也依此重新估算承保市場貝它值，結果發現核保之系統風險不同於由年資料之估計而呈現顯著負值。本文的發現可作為產險公司與監理單位在費率制定及清償能力管理上的重要參考。

Using Value at Risk (VaR) as a tool of risk measurement has gained increasingly attention not only in the bank industry but also in the insurance industry, in which underwriting risk is also known as the most critical risk in the property and casualty insurance industry. However, one of the shortcomings in previous studies is that the loss data frequency appears to be low, thereby, significant bias may produce when measuring actual risk. In this paper, we use monthly loss data for the automobile insurance in Taiwan to calculate underwriting loss ratio-at-risk. Since it is generally believed to be a more adequate measurement in risk management, we find that the loss ratio-at-risk based on annual data are significantly underestimated. In the additional cross-sectional analysis, we find that some firm’s characteristics may be closely related to the underwriting loss ratio-at-risk. Finally, we also discover that the underwriting beta based on monthly data is significant different from zero which is inconsistent with zero beta result from annual data. Our findings can give important sights for regulators as well as firm managers when they are making decisions about pricing and risk management.