Empirical Analysis of Mobile Broadband Adoption in Major Six Countries

Abstract

ObjectivesCurrently the development of mobile broadband has been proceeding rapidly and its access speed amounts to 150Mbps for the latest 4G (LTE), which is the same as that of fixed broadband. In these circumstances, mobile broadband, consisting of 3G and 4G, expected to play more important role due to their wider coverage areas with relatively low cost.The purpose of this paper is to analyze factors promoting mobile broadband adoption focusing on smartphones such as Android and iPhone. In this study, due to the data availability, six major countries such as the U.S., UK, France, Germany, Japan, and Korea are selected which have the large number of mobile subscribers and well established database related to mobile phones. Previous studies regarding mobile adoption including smartphones were based the survey data which contained relatively small samples. This study also attempts to contribute the current issues related to fixed and mobile broadband such as FMC (Fixed Mobile Convergence), their substitutability, and the backhaul network. MethodologyThis paper employs panel data analysis regarding six major countries and postulates the estimation equation such that the number of mobile subscribers can be explained as follows: where Subscriberit denotes the mobile broadband adoption rate in country i at t (quarter in 2000 to 2012), Subscribert-1 one period lag for examining network effect, Price (Voice)it monthly charge of voice services calculated by (Voice ARPU)/MOU, Price (Data) calculated by (Data ARPU)/MOU for country i (the Price variable is divided into two parts, voice and data, which is due to the fact that monthly data price is fixed and its speed increases rapidly), Speed download speed of mobile broadband, HHI the Herfindahl-Hirschman Index of mobile broadband market, Income GDP per capita. Factor denotes dummy variables including the launch of Android, iPhone, and FMC (whether carriers providing fixed and mobile services together or not). Regulations such as MNP (Mobile Number Portability) and frequency auction are excluded, since they were not significant. The data were obtained from National Regulatory Authorities, telecom carriers’ homepages or IR documents, and international organizations such as OECD and ITU.Results and further issuesThe results obtained so far are as described in Table and summarizes as follows: (i) launch of Android is positively significant, while the launch of iPhone is not; (ii) competition among carriers in terms of HHI is positively significant; (iii) voice price is negatively significant, while data price adjusted by speed is not significant; and (iv) FMC is negatively significant, that is carriers which provide both of fixed and mobile tend to have more subscribers. iPhone is not significant, because it was initially provided to a single mobile carrier in each country. These results demonstrate that competition among carriers promotes mobile broadband diffusion. This study provides an important suggestion to the backhaul issue of mobile broadband, since the variable FTTx can be thought as a proxy of the backhaul network. The related estimation result shows that carriers own larger backhaul have more mobile subscribers. Further elaborations to this estimation however are required to satisfy sign conditions on coefficient, particularly those of price and access speed, and to take care of endogeneity. This analysis will provide an important basis for national broadband policy including mobile in each country.Table: Estimation resultsDependent variable: mobile broadband subscribers (Log)Mobile Broadband Adoption Ratio (one lag, Log) 0.7670 *** (136.2) Price (Voice, Log) -0.0651 *** (-3.486) Price/Speed (Data, Log) -0.000883 (-0.237) Income (GDP/Capita, Log) 0.0708 (1.085) HHI (Log) -0.304 *** (-14.38) FTTx (Adoption Ratio, Log) 0.0240 *** (5.065) Android (Dummy) 0.0344 *** (3.221) iPhone (Dummy) 0.0064 (0.562) FMC (Dummy) -0.0617 *** (-3.513) Constant 2.5260 *** (3.692) Observations 195 Number of country_id 6 R-squared 0.998 Note: t-statistics in parentheses *** p