Modelling and Forecasting Mobile Telecommunication Services: The case of Greece

In this paper we try to model the adoption pattern of mobile telecommunication services into the Greek market for the period from 1993 to 2005. Two separate sigmoid curves, the Gompertz and the Logistic, are fitted to the observed number of subscribers by means of non-linear least squares. The in-sample fit to data favoured the use of the Logistic curve in describing the diffusion process, fact which is further supported by Frances’ parametric test (1994b). The dominance of the Logistic curve over the Gompertz is also verified via a pseudo out-of-sample forecasting exercise. Furthermore, an attempt is made to predict the expected number of subscribers up to 2015, solely based on the Logistic curve. Taking into account the prediction’s uncertainty, the variance of the forecast errors is calculated utilising the non-parametric bootstrap method. Our empirical results reached to three conclusions. First, the introduction of the pre-paid mobile telephony in 1997 along with the entry of the third mobile operator in 1998 has boosted the diffusion process in Greece; second, the levelling-off process in the diffusion of mobile phones has already begun; finally, the average expected growth rate in new subscribers is less than half percent for the period between 2006 and 2015.