Methods for quarterly disaggregation without indicators; a comparative study using simulation

Abstract

Various methods have been proposed to estimate quarterly figures from annual time series, but these methods have not been systematically evaluated in order to know which to use with which kind of data. This study compares, with a Monte Carlo simulation, those methods that use only information from annual series. Differences in the results of the methods, analyzed as functions of characteristics such as the number of years or variability of the original annual time series, suggest which methods to use under which conditions.