Competitive Learning in Yardstick Competition : Testing Models of Policy Diffusion

Organizations learning from others’ successful policies become more competitive not only because their policies improve but also because they avoid the costs of policy innovation. While latecomer advantage has been widely recognized by economists, the policy diffusion literature in political science has tended to neglect the connection between learning and competition. We distinguish competitive learning from learning that is not driven by competitive pressure, or ‘pure learning.’ We model policy diffusion as a game played on social networks governing competitive pressure and possibilities of information transfer. We develop an empirical test for competitive learning using spatial lags, which is applied to data on the performance of larger English local authorities from 2002 to 2006. We find evidence both for competitive learning and pure learning. The sharper distinction between causal mechanisms governing diffusion and our empirical approach should be widely applicable to diffusion across international boundaries and sub-national units