Crowdsourcing mobile coverage

• Mobile coverage affects socio-economic communication and performance.
• Using crowdsourced information on mobile coverage and femto-cells the existence of mobile divide in the US is measured.
• Low-income regions are found to receive 15% less coverage compared to their affluent counterparts.
• Operators install two fewer mobile antennas per tract in lower income areas for equal distributions of subscribers.
• These findings have implications for universal service delivery policies.

Mobile coverage affects social and economic communication and performance. Looking at the determinants of mobile network coverage helps highlight both operators׳ strategies and individual users׳ decisions. Subscribers often install small wireless antennas called femto-cells within buildings with poor signal. The location and density of these antennas provides information for areas with significant demand but poor coverage-areas where coverage is effectively crowdsourced. In contrast to the development of fixed-line networks, average income, population, geography, and education are the main drivers of both telecom and femto-cell based mobile coverage. Low-income regions are found to receive almost 15% less coverage compared to their affluent counterparts and have two fewer telecom base stations installed for equal distributions of (potential) subscribers. Base stations tend to be at least three times further from lower-income subscribers. Within poorly-covered areas, wealthier households are able to compensate by investing in femto-cells, but poorer households are often excluded from communication networks and the internet. The results of this research has implications for information policy.