Internet Use and Economic Growth

There is a widespread sense that the impact of the Internet has a far wider reach than the information-technology sector. The Internet began as an important tool for improving communication, but has since transformed into a ubiquitous technology supporting all sectors of the economy. Today, there is clear empirical evidence that the Internet has positive microeconomic effects, both at the individual and firm levels. The beneficial effects of the Internet at the microeconomic level mean that analogous effects should be detected at the aggregated, macroeconomic level. Indeed, several studies have examined the impact of the Internet on rates of economic growth. These studies employed broadband penetration rates as proxies for the Internet. This study verifies the magnitude of the potential impact of the Internet on economic growth by introducing an additional measure, the number of IPv4 addresses per capita, in the econometric analysis. Apart from its technical meaning, this proxy illustrates the number of separate devices connected to the Internet with their own unique IP address. Hence, it can be considered as an actual proxy for Internet usage by individuals and businesses. Moreover, this measure covers all the different ways of accessing the Internet (i.e. via broadband, dial-up and business leased lines). Of course, given that the Internet is relatively new and is growing relatively fast, the available data measure past developments and do not necessarily provide indicators of the Internet’s current use. For example, many countries seem to have reached saturation levels in terms of broadband penetration rates, and the number of IPv4 addresses has been recently depleted. However, there is no one single measure of the Internet that is free of bias. The advantage of employing the number of IPv4 addresses per capita is that it proxies the actual use of the Internet, whereas broadband penetration rates may undercount the total number of users if connections are commonly shared within a household or business. Therefore, it is useful to re-examine the question of how the Internet impacts economic growth by using alternative measures of the Internet. Clearly, correlation does not imply causation, and the basic quantitative association between the stage of Internet development and economic growth may be driven by reverse causality and other variables. To overcome the potential endogeneity problem, the econometric methodology used in this study is a two-step approach that involves constructing a predicted pattern for the evolution of the Internet. In the first stage, the evolution of the number of IPv4 addresses per capita is modelled across countries using the Gompertz function. This function is often used for modelling the evolution of technology markets, such as markets for digital television or mobile phones. The predicted evolution process is additionally explained with a set of indicators used as instruments that are not impacted by GDP (demographic and geographic indicators). In the second stage, the results of this modelling are used to produce the predicted evolution of the number of IPv4 addresses per capita. This predicted function is, in turn, applied to the growth equation. The conclusion that follows from the results of this exercise is that the growth of the Internet, expressed by the growth rate of IPv4 addresses per capita is one of the drivers of economic growth. This study suggests that, ceteris paribus, economies with faster Internet development are predicted to enjoy higher GDP growth rates compared to countries whose Internet development is slower.