The aim of this work is to test to what extent the star cluster population of a galaxy can be utilized to constrain or estimate the star formation history, with the Large Magellanic Cloud as our test bed. We follow two methods to extract information about the star formation rate from star clusters, either using only the most massive clusters or using the whole cluster population, albeit this is only possible for a shorter age span. We compare these results with the star formation history derived from colour–magnitude diagrams and find good overall agreement for the most recent ≈1 Gyr. For later ages, and especially during the ‘cluster age gap’, there is a deficiency of star clusters in relation to the star formation rate derived from the colour–magnitude diagram. The star formation rates following from the whole cluster population lie a factor of ≈10 lower than the star formation rates deduced from the most massive clusters or from the colour–magnitude diagram, suggesting that only ≈10 per cent of all stars form in long-lived bound star clusters.