In this paper we derive a method for the estimation of symmetric input–output tables (SIOTs), which makes it possible to use the commodity technology assumption even when use- and make tables are rectangular. The method also solves the problem of negative coefficients. In the empirical part we derive annual SIOTs in order to evaluate the differences between SIOTs calculated with different methods and the change in technical coefficients over time. Our results, based on data for Sweden, show that the impact of using different technology assumptions is rather large. However, in a factor content of trade application the impact of different technology assumptions does not seem to be very important. Also the size of the changes in the technical coefficients over time is found to be quite large, indicating the importance of calculating SIOTs annually.