Government policy, network externalities and mobile telecommunication services: evidence from OECD countries

Abstract

This article analyses the effect of government policy on the diffusion of mobile telecommunication services in member countries of the Organization for Economic Cooperation and Development (OECD). Specifically, we examine how the competition in and standard policies of each country affect cellular diffusion through interactions with positive and negative network externalities. The empirical analysis shows that significant network externality effects exist in a cellular market. Although the positive effects dominate the negative effects initially, the negative network externality effects become larger and outweigh the positive externality effects after a certain level of diffusion rate has been achieved. In particular, a single standard policy and the speed of technological innovation combined with the previous penetration rate generate a positive network externality on the diffusion of mobile telecommunications. However, a competition policy that solely increases the rate of new subscriptions does not generate any interacting effects.

Keywords:

• mobile telecommunications services
• penetration rate
• network externality
• cellular policies

JEL Classification:

• L10
• L96
• O33

Notes

¹  Numerous studies have attempted to verify the significance of network externalities using analytic models. For further details, see Katz and Shapiro (1985), Economides (1996), Swann (2002) and Farrell and Klemperer (2007).

² On the supply side, diminishing marginal growth rate of diffusion can be explained by low margins that discourage the cellular service providers to expand their services at the end of product life cycle. But we would like to mainly focus on the demand side for this phenomenon.

³ Implementing the single standard policy increases the possibility of selecting an inferior technology, which is known as the lock-in effect. For further details, see Gruber (2005).

⁴ ${\tau }_i$ depends mainly on the consumer’s preferences, GDP per capita and various socio-economic features in each country. For simplicity, however, we assume ${\tau }_i={\tau }_j=1$ for all countries. Since all 30 countries considered in this study are relatively homogeneous OECD members, we are justified in supposing that consumers in each country have similar living standards and preferences towards mobile telecommunication services. For further details, see Frank (2004) and Gruber and Verboven (2001a).

⁵ Equations 4 and 5 do not present network value, but denote the impact of the relative network size on new subscription rates. The value of a network is determined by the network type and the absolute network size (Swann, 2002).

⁶ For specifying $a$, we adopt the variables that capture innovation activities. In a similar vein, $b$ includes a demand-side variable and a standard policy dummy variable. For further details, see Gruber and Verboven (2001a).

⁷ This study considers a market with four or more firms as a competitive market instead of an oligopoly. Although this terminology is untraditional, we adopt this measure for simplicity.

⁸ Since there is no constant term in Models 1 and 3, we are able to conduct a panel analysis only for Models 2 and 4. Following the Hausman test results, we carry out panel fixed effects model analysis and confirm that the fixed effects estimates do not seem very different from the GLS estimates. As shown in Table 2, it is also worth mentioning that the panel fixed effects estimator cannot estimate the coefficient on a time-invariant variable such as the speed of technological innovation in this study. Thus, we mainly consider the results of panel GLS analysis for further discussions.

⁹ We obtain these penetration rates by differentiating Equations 6 and 7 with respect to $GD{P}_{it}$. That is, when the first derivative of each equation satisfies the conditions below, the impact of GDP per capita on new subscriptions changes from the negative to the positive value. Specifically, these conditions can be denoted by $\frac{\mathrm{d}{y}_{it}}{\mathrm{d}GD{P}_{it}}={\delta }_2{S}_{it-1}-{\delta }_5{S}_{it-1}^2=0$ in Model 3 and $\frac{\mathrm{d}{y}_{it}}{\mathrm{d}GD{P}_{it}}={\gamma }_6{S}_{it-1}-{\gamma }_{13}{S}_{it-1}^2=0$ in Model 4. Since Model 3 does not consider innovative adopters, the critical penetration rate in Model 3 is much higher than that in Model 4.

¹⁰ If the absolute value of the coefficient $S_{it-1}$ is identical to the absolute value of the coefficient $S_{it-1}^2$, then $S_{it-1}-S_{it-1}^2>0\mathrm{w}\mathrm{h}\mathrm{e}\mathrm{n}0<S_{it-1}<1$. If this condition holds, we are able to state that the previous penetration rate encourages rather than discourages new subscriptions when the penetration rate is greater than 0 and less than 1. In other words, new subscribers of cellular services always exist when the penetration rate is less than 1. In addition, we can suggest that the number of new subscribers would decrease when the penetration rate is greater than 0.5 and approaches 1. This statement can be easily proven mathematically. By differentiating Equation 4 with respect to $S_{it-1}$, we obtain $\frac{\mathrm{d}{y}_{it}}{\mathrm{d}{S}_{it-1}}=1-2S_{it-1}$. It is easily proved that $\frac{\mathrm{d}{y}_{it}}{\mathrm{d}{S}_{it-1}}<0$ when $\frac{1}{2}<S_{it-1}<1$ and ${\beta }_1={\beta }_2$.

Additional information

Funding

This work was supported by the National Research Foundation of Korea Grant funded by the Korean Government [NRF-2010-411-B00028].