The literature on economic growth in Latin America still presents mixed results. In the case of Mexico, the main findings show a similar pattern, highlighting the lower capacity of the southern and central states to take advantage of the export-oriented model. The present research analyses the economic growth path of the Mexican states over the period 1980-2018. Firstly, we analyse the income per capita distribution from a statistical point of view, seeking to better characterise the path followed by the national states in the past forty years. Secondly, we estimate a number of growth equations through panel data models, to scrutinise the role played by some factors, such as the level of education, public institutions, and external openness. As novel contributions, we provide a more robust way of estimating convergence clubs for the subsequent econometric approach; and we take into account a number of biases that properly improve the estimation of the growth equations. The main results indicate the presence of conditional convergence, pointing to the idiosyncratic nature of the state growth paths. These findings recommend the application of a customised policy at a state level, that could complement the federal one, helping to close the personal income gap across the country.