This paper examines the business cycle co‐movement in Mexican states over the period 2000–2014 by estimating an extended gravitational panel data model. Two different de‐trending filters are used to check the robustness of our results. The estimates suggest that the co‐movement increases as the size of the states’ economies does so as well as with the productive structure similarities and the relative level of development; however, the co‐movement decreases at a diminishing rate with geographical distance. There is also evidence of time‐dependent effects. In addition, the existence of moderate co‐movements among the states’ cycles suggests that common economic policies may not be suitable for all states, which implies there is a need for specific countercyclical policies to mitigate idiosyncratic shocks.