The empirical literature on development has labelled as “middle-income trap” (MIT) the fact that many developing economies struggle to adjust to new sources of growth after reaching middle-income levels. For Latin America and the Caribbean, this is an especially challenging scenario, as only Chile, Trinidad and Tobago, and Uruguay have become high-income economies in the last six decades while several other LAC countries, already middle-income as early as 1950, stayed in that income range. This paper analyses empirically the main policy areas explaining the MIT, based on the experiences of 76 emerging economies and OECD countries, comparing those which evaded it and those which stayed there since the 1950s. Based on more than 200 000 estimations using a linear discriminant analysis, we identify institutional, social and economic features that help characterise policy priorities to overcome the middle-income trap. Furthermore, using the Synthetic Control Method, we present for selected Latin American countries their main policy gaps according to their unique characteristics.