Temporal disaggregation of stock variables - The Chow-Lin method extended to dynamic models

Aurélien POISSONNIER

Since the seminal paper by Chow and Lin the literature on temporal disaggregation has focused on temporal disaggregation of flow variables. Moreover, this literature on optimal methods has traditionally emphasized static models and forced all dynamic dimension of the link between time series to be embedded in the unexplained component. Nevertheless, these techniques have proved particularly useful to compute quarterly national accounts in numbers of countries in Europe (France, Italy, Spain, Portugal, Switzerland). Following this literature, this paper builds an optimal method to derive higher frequency estimates of stocks variables using their annual value and related flow indicators at high frequency. An example of this method is given for quarterly non-financial corporations' capital in computers and communication equipment. Although presented in the case of stock variables, all the results from this paper hold in the case of flow variables disaggregated either with static or dynamic models.

Documents de travail
No G2013/03
Paru le : 27/02/2013