Economic historians that study long-term changes during the nineteenth and twentieth century are fundamentally restricted by the availability of qualitative data. As a result, researchers are forced to either impute missing data, or otherwise combine datasets in some way. In this article, we demonstrate the versatility of state-space models in addressing these problems. Not only do they enable us to compose large data series of high quality, they also provide a clear estimate of how reliable this data is, allowing any subsequent analyses to take this reliability into account. We illustrate the advantages of a state-space model using the population of Belgian municipalities as a case study. By combining growth and level data, we are able to compute yearly population statistics of over 2600 municipalities from 1880 to 1970.