Contracts for adaptive programming

This paper surveys a branch of economics known as contract theory, with the objective of drawing attention to some of the challenges that emerge when viewing adaptive programming from a contracting perspective. We ask: what insights does contract theory offer to practitioners of adaptive programming?

The premise of contract theory is that, because the aid agencies or governments who commission the services of non-governmental organisations (NGOs) and consultants often cannot know what actually happens during experiments, service providers may not genuinely experiment or may conceal the results if it is in their interest to do so. The essence of adaptive programming is to learn and change. When contracts are subject to revision, service providers will worry about aid agencies using the results of experimentation against their interests. Contract theory is concerned with designing contracts that overcome this incentive to conceal information, and induce genuine experimentation and truthful disclosure. This paper sketches some of the solutions.