Regional temporal disaggregation on economic series with macroeconomic balance: An entropy econometrics-based model

Abstract

Regional information about gross domestic product or gross added value is usually available with annual frequency. However, higher-frequency data at regional level is not so frequently available. This type of information is essential in order to (i) monitor the economic situation and (ii) forecast future values of regional output. The application of statistical techniques developed to estimate quarterly regional economic series from annual aggregates has spread out, especially by statistical agencies. Usually, each region tries to estimate separately its quarterly series fulfilling its annual regional constraint. Furthermore, these regional figures are required to be coherent with the national aggregate. In this paper, we deal with the problem of deriving quarterly regional accounts (QRA) estimates for R regions simultaneously that are consistent with annual regional accounts (ARA) and quarterly national accounts (QNA) by using an information-theory-based model. More specifically, this chapter presents an estimator based on entropy econometrics for regional temporal disaggregation of time series. We test the validity of the proposed model by a simulation exercise with real data of the Spanish Regional Accounts. Our benchmark for the analysis consists of the Chow- and Lin (1971)-type models, since they are the most commonly used technique in this field. The results obtained in the experiment suggest that the proposed method could be a preferable alternative to Chow-Lin disaggregation in situations of limited information.