Original
Statistical footprints in corruption measurement: general approach, with application to automobile license plates in Russia
Abstract
We offer a novel big data approach to corruption measurement by using statistical anomalies in publicly observable allocations affected by corruption. While individual corruption incidents remain undetectable, systemic corruption leaves measurable statistical footprints by affecting distributions of observable outcomes. We apply our approach to measuring corruption in the Russian traffic police, which issues automobile licence plates. Plates that signal status are found to be excessively concentrated among upmarket and luxurious classes and brands. We use these discrepancies to measure traffic police corruption, and discuss the potential of such approach in assessing the scale of abuse of power for private gains.
中文
腐败测量中的统计足迹：通用方法及在俄罗斯汽车牌照上的应用
摘要
我们提出一种新颖的大数据腐败测量方法，利用受腐败影响的公开可观测分配中的统计异常。尽管单个腐败事件仍无法被识别，但系统性腐败通过影响可观测结果的分布而留下可测量的统计足迹。我们将该方法应用于测量俄罗斯交通警察的腐败，该部门负责发放汽车牌照。我们发现，具有身份象征意义的牌照在高端和豪华车型及品牌中过度集中。我们利用这些差异来测量交通警察腐败，并讨论这种方法在评估以权谋私规模方面的潜力。
关键词
腐败测量、大数据方法、统计异常、汽车牌照、俄罗斯、交通警察、以权谋私