在本文中，我們提供懲罰對於犯罪行為嚇阻作用的台灣經驗。具體而言，本研究探討1997年採行提高門檻的假釋政策是否與犯罪率的降低有關。研究結果顯示，台灣實施新的假釋政策後，整體犯罪率顯著下降，包括輕罪與重罪，結果也發現新的假釋政策實施後，累犯率和假釋犯再犯率大幅下降。這些結果在控制與犯罪率相闋的變數後仍獲得支持，包括失業率、教育程度、定罪率、和警政支出。另外，在通過新的假釋政策後，新入監人數也顯著下降。總體而言，本研究結果顯示:較嚴格的假釋政策對於犯罪行為的確存在嚇阻作用，符合嚇阻假說。最後，本研究發現在1997年新的假釋政策宣告後，矯正機關成本有顯著增加的現象。

In this paper, we provide evidence on the deterrence effects of severity of punishment on crime rates based on the experience of Taiwan. Specifically, we investigate whether a tougher parole policy, which was implemented in Taiwan during 1997, is associated with a decline in crime rates. Our results indicate that the overall crime rate in Taiwan declined significantly after the implementation of the new parole policy. This is reflected in both misdemeanor and felony crimes. We also find that both the overall recidivism rate and the parolee recidivism rate declined significantly after the implementation of the new policy. These results hold after controlling for other variables that are likely to be associated with crime rates such as unemployment rate, educational level, crime conviction rate, and police outlays. We also find that the number of new prisoners admitted to correctional institutions declined significantly after the passage of the new parole policy. Overall, our results suggest that the tougher parole policy has a deterrent effect on criminal behavior, which is consistent with the deterrence hypothesis. Finally, we find that correctional institution costs increased after the 1997 parole policy change.