Part I contrasts the general types of policy advice that follow from three different approaches to understanding economic growth and technological change. Neoclassical theory gives policy advice that is assumed to be relevant for all countries at all times: remove sources of 'market failures'. Romer's branch of new macro growth theory stresses the nature of knowledge, non-rivalrous and partly appropriable. Structuralist-evolutionary theory is micro based and stresses the uncertainty that is associated with technological advance. Both of the latter approaches conclude that the neoclassical optimal allocation of resources is unachievable and hence the policy advice of removing impediments to achieving that optimum is not well grounded. As a result, policy advice for enhancing technological change must rely on a mixture of theory, empirical analysis and policy judgement. Part II deals with the large amount of context-specific policy advice that follows from structuralist-evolutionary theories.