A method for temporal disaggregation of flow time-series based on high-frequency indicator data and movement preservation principle

The article presents a new method for transforming the low-frequency flow time-series into a consistent time-series with less step period according to given high-frequency indicator time-series. This method is founded on the framework of well-known movement preservation principle and the proportional first difference Denton method. The analytical solution for general problem of flow time-series temporal disaggregation is obtained in the form that demonstrates its linear dependency on an initial value of the instrumental disaggregation parameter. Some ways of incorporating the known Denton and Cholette initial condition to the analytical solution are considered. A notion of initial condition for preserving seasonal cycles in dynamics of multiplicative adjustments is proposed along with the certain way of its incorporation to analytical solution for general problem of flow time-series temporal disaggregation. The frontiers of idempotency and wide opportunities to formulate a recursive base for developed method are investigated. Special attention is paid to a sensitivity of the general problem’s analytical solution to small changes in high-frequency and given low-frequency data. Special techniques of applying the proposed method operationally are considered. The opportunities for two-stage implementation of the method are of great practical interest in the circumstances of the recent low-frequency data revisions and/or the new data arrivals. Computing efficiency of the developed tools for temporal disaggregation is quite high because associated calculations come to inversion of a square matrix of the order that equals a number of observations in low-frequency data set available. The paper contains a row of illustrative diagrams that represent the results of experimental calculations with the official statistical data on quarterly gross domestic product and monthly output index for main kinds of economic activities in 2011-2015.

интервальный динамический ряд, темпоральное дезагрегирование, принцип сохранения движения, пропорциональный метод Дентона, оптимизационная задача, метод множителей Лагранжа, flow time-series, temporal disaggregation, movement preservation principle, proportional Denton method, optimization problem, Lagrange multipliers

1. Miller R.E., Blair P.D. Input-Output analysis: Foundations and extensions: 2nd ed. N.Y.: Cambridge University Press, 2009. 782 p.

2. Motorin v. A generalization of initial conditions in benchmarking of economic time-series by additive and proportional Denton methods // Iranian Economic Review. 2015. Vol. 19. No. 3. P. 251-264.

3. Dagum E.B., Cholette P.A. Benchmarking, temporal distribution, and reconciliation methods for time series. Series: Lecture Notes in Statistics. Vol. 186. New York: Springer, 2006. 409 p.

4. Блум Э.М., Диппелсман Р.Дж., Меле Н.Э. Руководство по квартальным национальным счетам: концепции, источники данных и составление: пер. с англ. Вашингтон, округ Колумбия, США: МВФ, 2001. 222 с.

5. Denton F.T. Adjustment of monthly or quarterly series to annual totals: An approach based on quadratic minimization // Journal of the American Statistical Association. 1971. Vol. 66. No. 333. P. 99-102.

6. Handbook on quarterly national accounts / European Commission, Eurostat; Cat. No. KS-GQ-13-004-EN-N. Luxembourg: Publications Office of the European Union, 2013. 312 p.

7. Моторин В.И. Аналитические и вычислительные свойства пропорционального метода Дентона // Вопросы статистики. 2012. № 7. С. 66-72.

8. Magnus J.R., Neudecker H. Matrix differential calculus with applications in statistics and econometrics: 3rd ed. Chichester, UK: John Wiley & Sons, 2007. 450 p.