TWO TYPES OF EXPERIMENTATION: Where do Institutional Bypasses Fit?

Abstract

Experimentation is in vogue in discussions of institutional design, but the concept contains an important ambiguity. That ambiguity is highlighted by Mariana Prado and Michael Trebilcock’s work on institutional bypasses. Prado and Trebilcock define an institutional bypass as a type of institutional reform that involves creating a separate institution which operates in parallel with and performs the same function as the original institution. They promote bypasses as ways of opening up room for experimentation. At the same time they distinguish institutional bypasses from randomized controlled trials, one of the best-known types of experimentation. In drawing this distinction Prado and Trebilcock highlight the fact that literature on experimentation in institutional design covers at least two different conceptions of experimentation. In some literature experimentation is exemplified by randomized controlled trials, in others it is exemplified by the more open-ended processes associated with experimentalist governance. This short essay elaborates on the distinctions between the two types of experimentation, argues that institutional bypasses are likely to fit best with the second type, and emphasizes that the two types of experimentation have different advantages and disadvantages as modes of learning and reform.