量刑自由裁量权的边界：集体经验、个体决策与偏差识别
吴雨豪量刑自由裁量权的边界：集体经验、个体决策与偏差识别
Boundary of Discretion in Sentencing: Collective Experience, Individual Decision and Deviation Identification
期刊名称：《法学研究》
期刊年份：
作者：吴雨豪
单位：
中文关键词：量刑规范化;自由裁量权;集体经验;偏差识别
英文关键词：sentencing normalization; judicial discretion; collective experience; deviation identification
中文摘要：
在量刑规范化改革中，如何在赋予法官自由裁量权的前提下，将“案件由谁审理”这一因素所导致的量刑偏差控制在合理限度内，是理论和实务要解决的问题。一种可行的方案是，将法官个体的刑罚裁量与法官量刑集体经验进行对比，对靠近集体经验量刑的法官的自由裁量权采取肯定和尊重的态度，而对量刑显著偏离集体经验的法官的量刑决策进行识别并纠正其偏差。基于北京地区五类案件近5万份刑事判决书，对盗窃、诈骗、抢劫、故意伤害和交通肇事案件的刑罚裁量展开实证研究后发现：“案件由谁审理”这一变量能够解释5% 11%的量刑差异；大多数法官的量刑结果在集体经验附近浮动，只有少部分法官的量刑结果显著偏离集体经验；更有少部分法官在多个罪名的刑罚裁量中均出现量刑偏差。量刑结果显著偏离集体经验的法官应成为自由裁量权规制中重点关注的对象，尤其是在多个罪名的刑罚裁量中均出现偏差的法官。量刑自由裁量权的规制方案，彰显了大数据与算法在量刑规范化改革中的应用前景，在引入“算法的程序控制”理念的前提下，具有推广价值。
英文摘要：
In the process of sentencing normalization reform, one issue that has drawn the attention of both academic and practical circles is how to acknowledge judicial discretion while keeping sentencing deviations within a legitimate limit. A feasible solution is to compare a judge's individual sentencing decision with the collective experience of judges and to adopt an affirmative and respectful attitude towards the discretion of judges whose decisions are close to the collective experience while identifying and correcting the decisions of judges that significantly deviate from the collective experience. This paper is an empirical study of nearly 50,000 criminal cases adjudicated in the People's Courts in Beijing, covering larceny, fraud, robbery, intentional injury, and traffic crimes. Firstly, the results of the multilevel model show that the variable "who hears and decides the case" can explain 5%-11% of the sentencing disparity. Secondly, most judges' sentencing decisions fluctuate around the collective experience, and only a small portion of judges' sentencing decisions deviate significantly from the collective experience and, therefore, need to be corrected. Finally, after comparing sentencing deviations across different types of criminal cases, we found that a small number of judges have sentencing deviations in multiple crimes. These judges should become the targets of discretionary governance. This discretionary governance scheme demonstrates the application of big data and algorithms in sentencing normalization reforms and has the prospect of being popularized under the premise of introducing the concept of "due process for algorithms".
全文阅读： 点击下载
