The flow of homicide cases in Brazil can be described as a funnel, as numerous events are reported to the police but only a few are convicted. This article aims to identify the variables influencing time to clearance and time to conviction for murder cases in Belo Horizonte (Minas Gerais) by examining a sample of police investigations (uncleared cases) and criminal proceedings (cleared cases) closed between 2003 and 2013. Using Cox proportional hazard model, we found that time to clearance is affected by crime attributes and police procedures although time to conviction is affected only by arrest in flagrante delicto.

中文翻译：

---

谋杀案在巴西城市中通过刑事司法系统流动

巴西发生的凶杀案可谓是一个漏斗，因为向警方报告了许多事件，但只有少数被定罪。本文旨在通过调查2003年至2013年间警方调查（未结案）和刑事诉讼（已结案）的样本，确定影响贝洛奥里藏特（米纳斯吉拉斯州）谋杀案件的通关时间和定罪时间的变量。在Cox比例风险模型中，我们发现通关时间受犯罪属性和警察程序的影响，尽管定罪时间仅受在公然逮捕中的影响。