Second thoughts on development accounting

We estimate the relative roles of factor inputs and productivity in explaining the level of economic development, which is measured as output per worker. For a large sample of countries, we show that alternative identifying productivity assumptions and alternative measures of human capital have a large impact on the relative weights of factor inputs and productivity in a decomposition of output per worker. For a sample of OECD countries, we find that productivity has almost no role in explaining cross-country differences in output per worker. This result supports the reasoning of a traditional neoclassical growth model.