Comprehensive Disaggregation Framework with Information Loss Function

Inconsistent data frequency is a common problem in many research fields; therefore, it should be handled before a particular study is well under way. Many novel ideas including disaggregation techniques, which are the major interest of this study, have been suggested to mitigate the nuisances of different frequencies. Regarding the issue of disaggregation, there are two main purposes of this study: first, to suggest a generalized framework to disaggregate lower- frequency series, and second to assess the performance of disaggregation. For the first purpose, we construct a disaggregation framework based on a model that consists of two stages: a regression and disaggregation of the residual from the regression by employing a state-space formulation. The contribution of this modeling is to reduce the disaggregation problem to univariate analysis, and therefore it facilitates analysis of the effects of aggregation on the underlying series. For the second purpose, this study examines what effects take place during aggregation, and measures the loss of information that inherently occurs during temporal aggregation. Then, we provide a set of practical criteria for the disaggregation performance according to the relationship between the aggregation effects and the information loss. To verify the superiority of the two-stage model, we run a Monte Carlo simulation and compare it with a counterpart model in terms of the disaggregation performance. The simulation not only confirms that the suggested model brings better disaggregation results but it also shows how the aggregation effects damage the underlying series. The results of the Monte Carlo simulation support the theoretical ground for both the disaggregation process and the assessment procedure developed in this study. Additionally, to describe the entire disaggregation process, we implement an empirical study. In the empirical example, real retail sales is the target series, and it is disaggregated within the two-stage model by utilizing relates series: personal consumption expenditure and the unemployment rate. With this empirical study, we further discuss the aggregation effects and the disaggregation performance from a practical point of view. This article has supplement material online.