Randomized experiments by government institutions and American political development

Are the methods of causal inference and, in particular, randomized controlled trials, compatible with the study of political history? While many important questions regarding political institutions and American political development cannot be answered with randomized controlled trials, scholars can and should be using the many instances of randomized experiments conducted by and within government institutions to further our understanding of institutions and political behavior. We argue that a surprising abundance of opportunities are available for scholars to utilize methods of random audits as natural experiments. Public and administrative officials have engaged in randomized interventions or audits to test for policy effects, to encourage compliance with the law, or to distribute government resources or personal risk to citizens fairly. With rare exceptions, such audits have not been leveraged by scholars interested in American political development or political history. Examples of randomized controlled trials conducted by agencies or institutions throughout US history are offered, and a historical random audit of members of the US Congress by the Federal Election Commission is highlighted. We conclude with limitations and advice on how to analyze the effects of randomized controlled trials conducted by governments. Scholars can use historical randomization to enhance causal inference and test theoretical implications, though deep knowledge of descriptive historical data and events are required to discover historical randomizations within political and legal institutions.