Page 232 - Contributed Paper Session (CPS) - Volume 6

CPS1907 Klaudia M. T. et al.
Handling technological changes by time varying
coefficient model analysis in flash estimate of
gross value added in information and
communication industry
Klaudia Máténé Bella, Ildikó Ritzlné Kazimir
National Accounts Department, Hungarian Central Statistical Office, Budapest, Hungary

Abstract
This paper investigates the linkage of physical indicators as expression of
technological changes and gross value added of the industry
telecommunication by time varying coefficient model. Empirical results for
quarterly data from 2000q1 to 2018q3 in Hungary indicates that the physical
indicators follow a logistic curve and they change each other overlapping with
technological progress. We find the relationship that the most physical
indicators effect significantly the gross value added of industry
telecommunication until their growth reaches the inflection point. In each sub
period one or two variables could be detected to be driving force behind the
growth. The goal was the construction of a model that enables the forecasting
of gross value added of telecommunication using the selected physical
indicators. We argue that time varying coefficient model is able to handle
technological changes and the quick changes of the explanatory power of
exogenous variables. With the handle the full time series time varying
coefficient model is an effective method to flash estimate of gross value added
of the information and communication industry.

Keywords
Logistic curve; state space model; GDP estimation

1. Introduction
In the flash gross domestic product (GDP) estimation the Hungarian
Central Statistical Office (HCSO) applies a bottom up approach in the
production side (Cserháti et al. (2009)). The HCSO utilizes the most available
physical indicators, and fits autoregressive integrated moving average
(ARIMA) models completed with explanatory variables usually. In the case of
information and communication industry the physical indicators have low
explanatory power separately, and the significant multicollinearity hinder the
accurate estimation of gross value added. The strange and unusual
relationship between physical indicators is due to the rapid technological
changes in whole economy.
A radical change in circumstances of production results changes in
different areas of economic environment, the paradigms of production can
transform. For example, a significant new technology requires new or

221 | I S I W S C 2 0 1 9