斡旋受贿精准量刑建议数字化的显示偏好理论面相——基于293例斡旋受贿犯罪裁判文书的数字化剖析及模型构建

Digitalization of Precise Sentencing Recommendations for Mediate Bribery in View of Revealed Preference Theory:Based on Digital Analysis and Modeling of 293 Mediate Bribery Cases

数字司法时代的来临推动了检察机关量刑建议的精准化,但司法活动受司法人员显示偏好影响并不适宜直接用技术算法拟合。对293例裁判文书进行结构化解析发现,斡旋受贿犯罪的量刑建议数字化面临困境:这一特殊的犯罪形式系多法条对应一罪名、数额特别巨大时缺少通行量刑细则、与人际关系密切。结合显示偏好理论,司法人员的裁判行为可被折叠到对司法机关量刑活动的分析评价中,实现对量刑建议数字化分析进行改良,并以此为基础构建对上述犯罪宣告刑与罚金刑的算法模型。模型初步揭示了司法人员在不同涉案金额下适用同一量刑情节会因群体性显示偏好,可能部分偏离并改变该情节适用的理论效果。

Digitalization of justice has advanced precise sentencing recommendations among China’s procuratorates.However,judicial activities might be affected by the revealed preferences of judicial officers,making them not direct-ly fit technical algorithms.The structured analysis of 293 cases of judgment exposed difficulties about the digitaliza-tion of sentencing recommendations for mediated bribery.Specifically,this special form of crime corresponds to mul-tiple legal provisions,lacks prevailing sentencing rules when the amount is particularly large,and is closely related to social interpersonal relationships.Based on the revealed preference theory,judicial behavior can be integrated into the analysis and evaluation of sentencing activities,which helps to improve the digital analysis of sentencing recommen-dations and then build an algorithm model of declared sentence and fine punishment for such crimes.The model pre-liminarily reveals that when judicial officers apply the same sentencing consequences for different amounts involved,they may partially deviate from or even change their theoretical effect due to revealed group preferences.