This paper analyses the policy and institutional determinants of long-run economic growth for a sample of OECD and non-OECD countries, with two objectives. First, it assesses the extent to which the main findings from growth regressions covering industrial countries are robust to a larger sample covering lower-income OECD and non-OECD countries. Confirmation is found from pooled mean group estimates for the larger sample of countries that long-run GDP per capita levels are increased inter alia by education policies, trade openness, R&D expenditures and policy frameworks that are conducive to low inflation, although the estimated effect of education is implausibly large. Second, the paper proposes a new growth regression framework that explicitly models technology diffusion and allows exploring the growth effects of a wider set of policies and institutions, while alleviating some of the constraints of the pooled mean group estimator. Under this approach, the estimated return to education is more in line with available evidence from microeconomic studies. Regulatory barriers to entrepreneurship, explicit barriers to trade and – especially – patent rights protection appear to be fairly robust determinants of long-run cross-country differences in technology. Some other policies and institutions such as trade liberalisation are found to speed up technology convergence. There is limited evidence here that the effects of policies and institutions vary depending on countries’ level of development. These findings are subject to the usual limitations of growth regression analysis.