The experimentation-accountability trade-off in innovation and industrial policy: are learning networks the solution?

The exact nature of industrial/innovation (I/I) policy challenges and the best way to address them are unknown ex ante. This requires a degree of experimentation, which can be problematic in the context of an accountable public administration and leaves the question of how to reconcile the experimental nature of I/I policy with the need for public accountability, a crucial but unresolved issue. The trade-off between experimentation and accountability requires a governance model that will allow continuous feedback loops among the various stakeholders and ongoing evaluation of and adjustments to activities as programmes are implemented. We propose an ‘action learning’ approach, incorporating the governance mechanism of ‘learning networks’ to handle the problems of implementing experimental governance of new and untried I/I policies. We resolve the issue of accountability by drawing on the literature on network governance in public policy. By integrating control and learning dimensions of accountability, this approach enables us to resolve conceptually and empirically trade-offs between the need for experimentation and accountability in I/I policy.