This paper investigates the contribution of immigration to income and productivity of host countries. Using a dataset constructed from census data and labor force surveys for 20 OECD countries in the period from 1960 to 2005, we explore the information on age and educational attainment of immigrants to assess the contribution of immigration to income components: changes in physical capital, human capital, employment, and total factor productivity. We combine level accounting approach with panel income regressions, and also account for the endogeneity of migration choices to productivity shocks. Our main findings are that, overall, higher shares of immigrants over natives have a positive effect on income and productivity of their host countries. Under the assumption that older immigrants are also the ones with the longest duration of stay, this effect is due to the long run changes in TFP, and is robust to educational disparities between immigrants and natives. The decomposition by age and education suggests that only unskilled immigrants have a non-neutral impact on income and productivity, which is negative in the short run but positive, and larger in magnitude, in the long run. We also find a dispersed impact of the presence of other immigrant groups on some income channels.