In many managed landscapes, low-productive land comprises most of the remaining relatively untouched areas, and is often over-represented within protected areas. The relationship between the productivity and conservational value of a site is poorly known; however, it has been hypothesized that biodiversity increases with productivity due to higher resource abundance or heterogeneity, and that the species communities of low-productive land are a nested sub-set of communities from more productive land. We tested these hypotheses for dead wood-dependent beetles by comparing their species richness and composition, as well as the amount and diversity of dead wood, between low-productive (potential forest growth < 1 m3 ha-1 year-1) and productive Scots pine-dominated stands in Sweden. We included four stand types: stands situated on (i) thin soils and (ii) mires (both low-productive), (iii) managed stands, and (iv) unmanaged stands set aside for conservation purposes (both productive). Beetle species richness and number of red-listed species were highest in the productive set-asides. Species richness was positively correlated with the volume and diversity of dead wood, but volume appeared to be a better predictor than diversity for the higher species richness in set-asides. Beetle species composition was similar among stand types, and the assemblages in low-productive stands were largely subsets of those in productive set-asides. However, 11% of all species and 40% of red-listed species only occurred in productive stands, while no species were unique to low-productive stands. We conclude that low-productive forests are less valuable for conservation than productive forest land. Given the generally similar species composition among stand types, a comparable conservational effect could be obtained by setting aside a larger area of low-productive forest in comparison to the productive. In terms of dead wood volumes, 1.8–3.6 ha of low-productive forest has the same value as 1 ha of unmanaged productive forest. This figure can be used to estimate the conservation value of low productive forests; however, as productive forests harbored some unique species, they are not completely exchangeable.