The molecular clock hypothesis is fundamental in evolutionary biology as by assuming constancy of the molecular rate it provides a time frame for evolution. However, increasing evidence shows time dependence of inferred molecular rates with inflated values obtained using recent calibrations. As recent demographic calibrations are virtually non-existent in most species, older phylogenetic calibration points (>1 Ma) are commonly used, which overestimate demographic parameters. To obtain more reliable rates of molecular evolution for population studies, I propose the Calibration of Demographic Transition (CDT) method, which uses the timing of climatic changes over the late glacial warming period to calibrate expansions in various species. Simulation approaches and empirical datasets from a diversity of species (from mollusk to humans) confirm that, when compared to other genealogy-based calibration methods, the CDT provides a robust and broadly applicable clock for population genetics. The resulting CDT rates of molecular evolution also confirm rate heterogeneity over time and among taxa. Comparisons of expansion dates with ecological evidence confirm the inaccuracy of phylogenetically derived divergence rates when dating population-level events. The CDT method opens opportunities for addressing issues such as demographic responses to past climate change and the origin of rate heterogeneity related to taxa, genes, time and genetic information content.