The Effects of Broadband Internet on Mental Health

Project Information

BIMH

Grant agreement ID: 101027302

DOI

10.3030/101027302

Project closed

EC signature date 19 April 2021

Start date 1 August 2021

End date 31 July 2023

Funded under

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Total cost

€ 178 320,00

EU contribution

€ 178 320,00

Coordinated by

UNIVERSITE LIBRE DE BRUXELLES
Belgium

Periodic Reporting for period 1 - BIMH (The Effects of Broadband Internet on Mental Health)

Reporting period: 2021-08-01 to 2023-07-31

Summary of the context and overall objectives of the project

Since the late 1990s, governments around the world have allocated large amounts of public funds to develop high-speed internet access via broadband infrastructure. Between 2004 and 2011, the percentage of EU households connected to broadband internet rose from 15 to 67% (Eurostat, 2011). Residential high-speed internet has altered how, when, and where individuals conduct a wide range of activities, including work, search for information, sleep, and interact with others. Despite the growing number of studies analyzing the effects of broadband Internet access on a wide range of socioeconomic outcomes, there is scant causal evidence on the link between residential work-related internet use and mental health.

There is a growing concern that constant connectivity to work may disrupt work-life balance and be detrimental to workers' well-being. Understanding the causal link between work-related ICT use at home on workers' mental well-being is extremely important to design evidence-based public and health policy. For example, the “right-to-disconnect” policies—which allow workers to be unreachable after work hours—have been gaining traction in many European countries since 2016. In the case of Belgium, the analysis contributes to the analysis of the determinants of long-term disability insurance claims.

This project explores the relationship between access to broadband internet and workers’ mental health in Belgium. Our empirical strategy exploits a technological feature of the telecommunication infrastructure that generated substantial variation in the availability and quality of internet: the distance of a household from a network node determined internet access and speed. We merge information about the internet network topology with geocoded longitudinal data from the Belgian Crossroads Bank for Social Security, a rich collection of administrative microdata that collects information on Belgian workers’ labor market trajectories and basic demographics.

Using a dynamic difference-in-differences design, we compare mental health-related outcomes of individuals with access to residential internet to those without access. We find that access to broadband internet when initially deployed is associated with a 2-3% increase in long-term disability rates with respect to baseline values. This increase is driven entirely by increases in disability insurance claims due to mental health conditions. We find no changes in the probability of entering long-term disability due to other medical conditions. This project contributes to a broader understanding of the upward trend in work-related illnesses—burnout and chronic stress—and their role in the expansion of disability insurance programs.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

The potential effects of internet use on mental health are multidimensional because of the many ways internet is used. Examining the impact of internet access on mental health is challenging due to the potential endogeneity of internet diffusion. Internet subscriptions are correlated with other observable demographic characteristics (such as income, employment, and education) that are also correlated with mental well-being.

Our empirical strategy is based on exogenous discontinuities in internet access that stem from the characteristics of the Asymmetric Digital Subscriber Line (ADSL) technology. Initially, the connection of households to the ADSL internet network depended largely on the voice telecommunication infrastructure that preceded the arrival of the ADSL technology by several decades. In Belgium, the local distribution network consists of nodes, called local exchanges (LEs). We exploit the distance between individuals’ dwellings and their closest LE as a source of variation because the main challenge when transmitting a digital signal over the telephone network is the decay in strength associated with distance. Thus, the strategy for identifying the causal effect of high-speed internet access at home on mental well-being is based on a comparison between individuals who lived in areas with access to broadband internet, within 3km from a LE, and individuals who lived in areas with no access to high-speed internet.

We combine longitudinal administrative data between 1998 to 2015 with the topology of the internet network. We estimate a dynamic difference-in-differences model, an empirical approach similar in spirit to those in Amaral-Garcia et al. (2021), Ahlfeldt et al. (2017), Campante et al. (2017), and Falck et al. (2014). Our preliminary results, presented in Figure 1, indicate that residential access to internet is associated with a 2-3% increase in long-term disability rates (w.r.t baseline). However, there are many reasons why individuals may entry DI. To validate to some extent the mechanism that we have in mind, we look into the medical conditions that appear as reason for entering long-term DI. We find that the probability of DI due to mental health conditions increases for people living close to LEs relative to people living far from LEs (Figure 2). We find no statistically significant differences in the probability of DI due to musculoskeletal conditions (Figure 3) or all other conditions pooled (Figure 4). Overall, our preliminary results indicate that residential access to internet is associated with a 2-3% increase in long-term disability, driven by mental health conditions.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

This research project contributes to three strands of literature. First, the paper is related to the growing literature on the causal impact of broadband internet on a wide range of socioeconomic outcomes: wage and employment (Atasoy, 2013; Forman et al., 2012); search frictions (Bhuller et al., 2020); female labor force participation (Dettling, 2017); educational attainment (Faber et al., 2016); fertility (Billari et al., 2019); social capital (Bauernschuster et al., 2014; Geraci et al., 2022). We contribute to this literature by identifying the causal link between internet access and objective measures of adult mental health.

Second, this project is related to work studying the causes and consequences of (mental) health-related conditions. For example, the determinants of disability insurance claims (Autor and Duggan, 2003, 2006; Liebman, 2015; Fevang et al., 2017; De Brouwer et al., 2023), and the effects of business cycle (Avdic et al., 2021), globalization (Colantone et al., 2019), poverty (Ridley et al., 2020), and precarious employment (Moscone et al., 2016) on mental health. Here, we examine the link between access to broadband internet, workers’ well-being, the increase in disability claims due to mental health conditions.

Third, this research also adds to the growing literature exploring the relationship between digital technologies and (mental) health. For example, the effect of TV/internet on children’s obesity and mental health (Nieto and Suhrcke, 2021; Donati et al., 2022), the effect of social media on body image and self-esteem (McDool et al., 2020; Braghieri et al., 2021). Finally, studies have focused also on the impact of internet access on sleep duration (Billari et al., 2018), adult obesity (DiNardi et al., 2019), and self-reported mental health (Golin, 2021). We contribute to this literature by exploring the effects of work-related ICT use at home on health.

Figure 4

Figure 3

Figure 1

Figure 2