中文摘要：

犯罪形成的原因複雜而多元，根據過去的相關研究顯示，大抵跟社會或經濟因素有關。本文以 1998～2014 年期間臺灣地區共計 20 個縣市的追蹤資料，從三大犯罪類型－－全般刑案、竊盜犯罪、暴力犯罪的發生率統計，進行與所得分配不均程度的相關性實證研究。本研究針對不同犯罪類型的發生率，分別選取吉尼係數、相關社會經濟與警力遏阻等構面變數，運用 Pooled OLS 與固定效果模型分別進行實證。實證顯示，吉尼係數與「全般刑案發生率」呈正相關，與「竊盜犯罪發生率」及「暴力犯罪發生率」呈負相關；就業人口數、平均每人警政支出及全般刑案破獲率與三大犯罪類型均呈顯著性負相關；離婚率與三大犯罪類型均呈顯著性正相關；人口密度與全般刑案發生率在 Pooled OLS 模型呈顯著性正相關、固定效果呈負相關，與其他二大犯罪類型發生率均呈顯著性正相關；警民比與暴力犯罪發生率在 Pooled OLS 模型呈負相關、固定效果呈正相關，與其他二大犯罪類型發生率均呈正相關。綜上所述，所得分配不均及有關社會經濟因素與刑事案件具顯著相關性，跟國內外文獻普遍的研究結論大致符合，希冀本文實證結果可作為策訂維護治安作為及相關研究之參考。

英文摘要：

The causes of crimes are complex and diverse. According to previous studies, the occurrence of crimes is related to social and economic factors. This study investigates panel data of 20 cities in Taiwan from 1998 to 2014. In addition, we investigate relationship between three major types of crime rate, including grand total crime, larceny, and violent crime, and income inequality.In this study, we selected Gini coefficient, social and economic variables, and policedetered variables, and used pooled OLS model and fixed effect model to investigate the occurrence of three major types of crime. The empirical results show that the Gini coefficient has a positive correlation with the occurrence rate of grand total crime and has a negative correlation with the occurrence rate of larceny and violent crime. Number of employed persons, per capita policing expenditure and cracked rate of grand total crime have significant negative correlations with the occurrence rate of three major types of crimes. Divorce rate has a significant positive correlation with the occurrence rate of three major types of crimes. According to the pooled OLS model, population density has a significant positive correlation with the occurrence rate of grand total crime; according to the fixed effect model, population density has a negative correlation with the occurrence rate of grand total crime. For both the pooled OLS model and the fixed effect model, population density has significant positive correlations with the occurrence rate of larceny and violent crime. In the pooled OLS model, the police officer to population ratio has a negative correlation with the occurrence rate of violent crime; in the fixed effect model, the police officer to population ratio has a positive correlation with the occurrence rate of violent crime. In both the pooled OLS model and the fixed effect model, the police officer to population ratio has positive correlations with the occurrence rate of grand total crime and larceny.In summary, income inequality, social and economic factors have significant influence with the occurrence of crimes. The results accorded with previous studies. The empirical results can serve as a reference for policy plan of public security and for future research.

資料來源：

法源法律網

https://www.lawbank.com.tw/treatise/pl_article.aspx?AID=P000238850