The need to combine data from different frequencies plays an important role for many economic decision-makers
and economists. The process, which consists in using higher frequency data to construct a higher frequency
indicator from its lower frequency counterpart, is called temporal disaggregation. In this paper, we propose a
new temporal disaggregation technique based on MIDAS regression using time series data sampled at different
frequencies. We first propose a simple disaggregation procedure more flexible than the more traditional
approaches, such as Chow–Lin (1971), and we extend the procedure to a dynamic setting. The proposed
procedure is flexible enough to take into account seasonality or calendar effects. An extensive simulation study
examines the performance of the new approach compared to alternative approaches.