Does broadband internet reduce the unemployment rate? Evidence for Germany

- OLS results indicate negative association between DSL and unemployment rates.
- May be due to reverse causality: higher income leads to higher demand for DSL.
- IV approach is used that exploits path-dependencies in the roll-out of DSL networks.
- Instrumental-variable results do not confirm the OLS result.
- This suggests that the availability of DSL does not reduce the unemployment rate.

Broadband internet is widely considered an important determinant of economic growth that has a positive effect on various economic variables. This paper addresses the question whether differences in the availability of broadband internet between German municipalities can explain differences in the unemployment rate. OLS results indicate a negative association between DSL availability and unemployment rates. However, the roll-out of DSL networks is not random. To address concerns of reverse causality and omitted variables, an instrumental variable approach is used, which rests on impediments to the roll-out of DSL networks that stem from the structure of the pre-existing voice-telephony network. The results of the instrumental variable estimations do not confirm the OLS results suggesting that the availability of DSL does not reduce the unemployment rate.