Data Science for Justice: Scale-up of a nationwide randomized intervention

Together with the Kenyan judiciary and the World Bank, we want to scale an innovation that was implemented in a nationwide randomized experiment including all the courts in Kenya, and that increased the speed of courts and had downstream economic effects on contract enforcement, firm productivity and the livelihood of all.

Following a $120 million judicial reform supported by the World Bank, we use the newly collected administrative data on cases going through all the courts in Kenya to develop an algorithm that can measure for every court the reasons for delays and predict the improvement in performance if those delays were addressed. We display this information on visually appealing “one-pagers”.

We sent the one-pagers to judges in one third of the courts throughout Kenya. In another third of the courts, we sent the one-pagers to judges and a court user committee comprising representatives from the civil society, lawyers, and police; to boost accountability.

In our study, we found strong effects of the second intervention on the speed of courts and on downstream economic effects. The intervention has a high potential to scale to the rest of the country. It is simple and cheap: it only consists in analyzing already existing data and sending one-pagers to the courts in Kenya. The Kenyan judiciary is already actively involved in designing and organizing the intervention. We have already sent the one-pagers to two-thirds of the courts in Kenya.