Does informing plaintiffs about expected case outcomes lead them to hire better quality lawyers? We conduct a randomized field experiment with the Mexico City Labor Court to provide statistical predictions and meetings with court conciliators to potential plaintiffs. Almost all potential plaintiffs are first-time users of the court, who commonly contact a lawyer through intermediaries working at the entrance to the court building. On random days, we provide one of three treatments to dismissed workers approaching the court to for information. In one arm, we provide very basic information about the process of filing a lawsuit at the court; in a second, we add personalized predictions of case outcomes based on machine learning models drawing from historical case files; in the third arm we add a letter of appointment with the court conciliator. We find that providing information on case outcomes does not change the likelihood of filing a suit but increases the quality of the lawyer hired. Adding the letter of appointment with the conciliator reduces the likelihood the worker files a suit and increases the quality of the lawyer hire conditional on filing a suit. We measure the quality of lawyers through both subjective and objective measures. Using these measures, we show that lawyers contracted through intermediaries at the entrance of the court are low quality lawyers. We find that the effect of providing information works mainly through a reduction in the percentage of subjects who hire a lawyer through these intermediaries.
Joyce Sadka and Enrique Siera, Instituto Tecnológico Autónomo de México, Christopher Woodruff, University of Oxford