The uneven provision of digital infrastructure between regions in Germany can hinder digital participation. Using detailed regional data on wired and mobile internet availability, we show that the expansion of mobile data infrastructure in recent years has been able to mitigate existing regional differences in internet connectivity.
Internet-based technologies are used in most areas of our coexistence. In economic terms, for example, they help with job searches, facilitate the flow of information and data within and between companies and enable companies to offer innovative digital products. Internet-based applications also play a major role in social life. For example, a significant part of social communication takes place via internet-based news services and social media, we orient ourselves using digital maps or search, buy and evaluate offers on online platforms.
For all these applications, an adequate internet connection is the basic prerequisite that makes this social and economic digital participation possible in the first place. Internet access is provided by the local digital infrastructure, a complex system of various technical components. The local network infrastructure determines the availability of internet on site, for example the speed for data transfers or the frequency of disconnections. The supply situation with digital infrastructure differs greatly from region to region, which means that digital participation is not equally possible in all regions. It is the declared goal of politics to reduce this connectivity gap – regional differences in internet availability – and to provide all regions with adequate internet in order to enable the use of internet-based technologies everywhere.
One challenge to this objective is that the provision of internet infrastructure in cities and conurbations is much more cost-effective than in more sparsely populated areas, because less investment in physical infrastructure is sufficient to reach more people there. This market incentive is more pronounced for wired internet than for mobile internet, as wired provision (e.g. DSL, cable or fibre) requires physical infrastructure to be built in or at least close to the building being served. With mobile internet, on the other hand, the construction and connection of transmission masts is sufficient to provide internet in a larger area. For example, a transmission mast with frequencies in the 700 and 800 MHz range can achieve a range of more than 10 km under appropriate geographical and weather conditions. The market cost incentives for providing mobile internet in less populated – and therefore generally less well served – regions are more favourable than for wired technologies. Due to this economic incentive effect, broadband mobile internet connections have the potential to mitigate the regional connectivity gap.
This paper first presents the extent of regional differences in wired internet and then uses existing literature to examine the extent to which mobile internet can compensate for the lack of wired access. Then, in an empirical analysis, we ask whether mobile internet has actually been able to dampen the regional connectivity gap in Germany in recent years.
How large is the regional connectivity gap for wired internet in Germany?
The differences in regional connectivity in terms of wired internet availability are shown in Figure 1. The underlying regional data come from the Federal Government’s Broadband Atlas. They provide information on the availability of wired internet according to different download speeds as a proportion of households in a district. Figure 1 shows the deviation of broadband availability of German districts and independent cities from the average. On average, 79.2% of households in a district had wired internet availability of at least 50 Mbps in 2018.
The strong urban-rural divide in wired coverage is striking. The first forty places are exclusively occupied by cities without districts, with Regensburg in Bavaria as the front-runner. There, 99.6% of households have a wired broadband connection with at least 50 Mbps. The best-served district is the Hochtaunuskreis in Hesse with 96.3%. At the other end of the spectrum is the Eifelkreis Bitburg-Prüm on the Luxembourg-Dutch border in Rhineland-Palatinate with the worst wired broadband availability (36.3%). The strong correlation between population density and coverage also means that there are many districts with below-average coverage, especially in eastern Germany. But also in other rural areas, such as the Bavarian-Polish border region or some districts in central Germany, the coverage is strongly below average.
The reason for this situation, apart from demand-side factors, is above all the high supply-side provision costs per inhabitant. This is because line-bound connections require a physical connection to the individual buildings to be supplied by corresponding infrastructure. Even if the existing cable and telephone network can be used for DSL and cable connections, considerable technical investments are necessary, for example in technology at local distribution boxes. The infrastructure for fibre-optic connections, which is needed for particularly high and stable connections, often has to be created completely from scratch.
To what extent can mobile internet compensate for the lack of wired technologies?
The necessary investment costs for the provision of mobile internet, on the other hand, are lower. This is because there are no cost-driving individual connections to buildings. Mobile radio towers also have to be erected and connected to the network, but because of their range they can supply considerably more people in their vicinity. This technological advantage over grid-based technologies offers a “natural” economic incentive to supply less densely populated areas as well. However, the question arises to what extent wired and mobile internet are substitutes on the demand side, i.e. to what extent they represent an alternative for the user.
There are several studies in the literature on this question. Typically, in the years of the emergence of mobile internet, a complementarity to wired technologies is empirically established. For example, in their analysis of OECD countries in the period 2001 to 2009, Wulf et al. (2013) find that it is mainly people who already have a wired internet connection who also sign up for a mobile internet contract. Grzybowski (2014) confirms this in a study of EU countries in 2005 to 2012, but he also shows that with the advent of 3rd generation mobile communications (3G or UMTS), there are increasingly more households that rely only on mobile internet service. This shows that it is probably not so much the technology itself that matters, but rather how powerful mobile internet is compared to wired internet. What is important for the user is what speed, response time and connection stability is made possible by a technology compared to an alternative, and at what price.
The requirements with regard to these factors differ depending on the intended use. Some applications require a relatively high internet speed, for example video streaming, gaming or data-intensive cloud applications. In contrast, a lower data transfer rate is sufficient for other purposes such as messenger and news services, internet research or most social media. Lindlacher (2021) shows that in Austria, at low speeds (less than 20 Mbits), wired and mobile internet are substitutes. This is no longer true at higher speeds, which allow for other uses. As the data transfer rates of wired internet connections have typically been higher so far, Quaglione et al. (2020) accordingly find complementarity between mobile and wired internet in Italy for video streaming, gaming and cloud service users. For users of social networks or music streaming, however, the two technologies are rather substitutes, according to the study.
Furthermore, the use of mobile or wired technologies is also strongly related to the end devices used. For example, mobile reaches a broader segment of the population than wired internet, because smartphone use is higher in lower-income population groups than through PCs (Reisdorf et al., 2022; Marler, 2018). Another advantage of mobile over wired internet technologies is certainly their general location-independence.
In summary, it can be said that the internet speed made possible by mobile internet was lower than that of wired technologies during the period under review. Even if mobile internet is therefore not a perfect substitute for wired technologies, it was certainly an alternative for some important internet applications such as messenger and map services or internet searches in recent years and in this respect is fundamentally suitable for at least partially compensating for the lack of wired connections.
Has the connectivity gap been mitigated by the expansion of mobile data networks?
In order to have had a dampening effect on the regional connectivity gap, the provision of mobile data infrastructure must have improved disproportionately in regions that are less well served by wired internet. Whether this is the case is examined empirically below.
For this purpose, we use the location certificates of mobile radio towers of the Federal Network Agency. The database contains all radio systems of the German mobile radio networks with information on the exact location (coordinates), operating frequency (in MHz) and date of commissioning or change in operating status. As of 26 July 2021, there were over one million radio systems recorded at over 70,000 locations, which we aggregate to county level for the purpose of this analysis.
Figure 2 shows the number of mobile towers per 100,000 inhabitants at county level in 2018.
On average, there are 84.7 mobile towers per 100,000 inhabitants in a district in 2018. There are large regional differences here. Fürstenfeldbruck (Bavaria) has the fewest mobile towers with 45.1 per 100,000 inhabitants and the most with 154.5 can be found in Weimarer Land (Thuringia). It is particularly striking that in general more infrastructure for mobile internet is typically provided per inhabitant in more sparsely populated areas. Thus, 13 of the 15 districts with the highest number of mobile phone towers are located in the (more sparsely populated) new federal states. This is because conurbations tend to require fewer radio towers for high network coverage than rural areas. This means that, on average, the per capita investment in mobile data infrastructure is higher in rural areas.
Comparing the two maps in Figure 1 and 2 shows a strong (negative) correlation between wired coverage and mobile towers per capita. This is illustrated again in Figure 3, in which this correlation (population-weighted) is plotted as a red straight line and each blue circle represents a district or an independent city. The higher the number of inhabitants in a district, the larger the districts. Here it becomes clear once again that the cities (larger circles in the chart) tend to have higher wired broadband availability and at the same time have less investment in mobile internet infrastructure (radio towers) per inhabitant. Meanwhile, at the other end of the spectrum, on average, (rural) counties with few inhabitants are characterised by comparatively high per capita investment in mobile data infrastructure and low wired internet availability. This correlation precisely reflects the economic advantages of the respective technology described above.
However, the sheer number of mobile towers per capita does not, of course, indicate better mobile internet coverage in these areas. That is why we investigate where mobile data infrastructure has been particularly expanded in recent years, thus improving mobile internet coverage. In a preceding analysis, we show that in the last ten years the mobile network has been greatly improved with around 33,000 new mobile sites across Germany (Goldbeck, Lindlacher, & Schwarz, 2021). If we look at this expansion of the mobile internet infrastructure in relation to existing wired broadband coverage in Figure 4, we see a U-shaped correlation: the expansion between 2010 and 2020 was particularly strong in counties that were poorly served with wired internet (at least 30 Mbit/s) in 2013 and in counties (especially larger cities) that were already very well served with wired internet. Meanwhile, the expansion of mobile data infrastructure per capita was less pronounced in counties with average wired coverage. This suggests that the economic advantage in the provision of mobile internet relative to wired internet access operates in more sparsely populated areas and has triggered higher infrastructure investment per capita there. The strong expansion in urban areas during this period can be explained by the fact that the mobile network in metropolitan areas was already reaching capacity limits, which was addressed by redensifying the networks. However, Figure 3 illustrates that overall, even after this redensification, less investment per capita was made in radio towers in metropolitan areas.
Using data from the Broadband Atlas and from the Federal Network Agency, our paper shows that the connectivity gap has been mitigated by the increased roll-out of mobile data infrastructure in rural regions – typically less well served with wired internet – in Germany. In addition to regulatory requirements, the economic cost advantage of this type of coverage seems to play a major role in this increased per capita investment in mobile internet in rural areas. The regional differences in digital participation would therefore be greater today without this expansion of mobile networks.
Although mobile internet is already a substitute for wired internet access for some important applications – such as the use of map and messenger services, social media or for internet research – it cannot be an equivalent alternative for all applications due to the lower data transfer rates. However, this could change in the future if new mobile technologies enable higher bandwidths at lower provisioning costs compared to wired technologies. In order to enable full digital participation everywhere as cost-effectively as possible, future investments and funding measures in internet provision should be as open as possible to technology and results-oriented.
Goldbeck, M., Lindlacher, V., & Schwarz, S. (2021). Funkloch Deutschland? Der Ausbau der mobilen Datennetze in den letzten zehn Jahren. ifo Schnelldienst, 74(11), S. 33-37.
Grzybowski, L. (2014), “Fixed-to-mobile substitution in the European Union”, Telecommunications Policy 38(7), 601-612.
Lindlacher, V. (2021), “Low demand despite broad supply: Is high-speed Internet an infrastructure of general interest?”, Information Economics and Policy 56(3)
Marler, W. (2018), “Mobile phones and inequality: Findings, trends, and future directions”, New Media & Society 20(9), 3498-3520.
Quaglione, D., N. Matteucci, D. Furia, A. Marra und C. Pozzi (2020), “Are mobile and fixed broadband substitutes or complements? New empirical evidence from Italy and implications for the digital divide policies”, Socio-Economic Planning Sciences 71, Article 100823.
Reisdorf, B. C., L. Fernandez, K. N. Hampton, I. Shin und W. H. Dutton (2022), “Mobile Phones Will Not Eliminate Digital and Social Divides: How Variation in Internet Activities Mediates the Relationship Between Type of Internet Access and Local Social Capital in Detroit”, Social Science Computer Review 40(2), 288-308.
Wulf, J., S. Zelt und W. Brenner (2013), „Fixed and mobile broadband substitution in the OECD countries – A quantitative analysis of competitive effects”, 46th Hawaii International Conference on System Sciences, 1454-1463.
The blog posts published by bidt reflect the views of the authors; they do not reflect the position of the Institute as a whole.