Why econometricians don't replicate (although they do reproduce)

This article attempts to give a more formal statement of the model of the problem of replication given in Harry Collins' Changing order, and to apply the model to a recent controversy over the lack of replication of econometric research in the economics literature. This paper shows that due to problems in defining ‘asymptotically satisfactory replication’ in conjunction with strong incentives to ‘reproduce’ (i.e., to extend any empirical result in a ‘constructive’ manner), it is highly unlikely that scientists will engage in replication behaviour. We conclude with some observations on the need to restructure the social structures of science in order to encourage a higher level of replication.