The Internet, Search Frictions and Aggregate Unemployment

How has the internet affected search and hiring, and what are the implications for aggregate unemployment? Answering these questions empirically has proven difficult due to selection in internet use and difficulty in measuring the search activities of both sides of the labor market. This paper overcomes these challenges by combining plausibly exogenous variation in the availability of high-speed internet in Norway with large-scale survey and administrative data on hiring firms, job seekers, and vacancies. Our empirical analysis shows that the internet expansion led more firms to recruit online and caused 9% shorter vacancy durations and 13% fewer unsuccessful hiring attempts. While the expansion increased job-finding rates by 2.4% and starting wages by 6% among the unemployed, we find no evidence of changes in job-to-job mobility or wage growth for employees. To interpret these findings, we develop and calibrate an equilibrium search model with endogenous job creation and destruction where workers decide how much search effort to exert on and off the job. Through the lens of the calibrated model, we find that better search technology is the main driving force behind our quasi-experimental evidence. Our calculations indicate that the steady-state unemployment rate fell by as much as 14% due to the broadband internet expansion.