Digital Divide: Evidence from a Federal Natural Experiment

Most previous research into the digital divide has relied on archival or primary surveys of a cross-sectional nature. Their findings are sensitive to endogeneity and cannot identify variations in the digital divide over time. Other research focused on the digital provide, focusing on the impact of new technologies in the continuing presence of the old. We exploited a natural experiment in implementation of a service to identify the causal effect of communications media on consumer choices. For 10 days, the service was provided through telephone only to the west but through the Internet to all states. Comparing consumer responses in the east (only Internet service available) and west (both Internet and telephone service available), we found robust evidence of digital divide. Absence of telephone service resulted in 27% fewer consumers. The loss was concentrated among the elderly, but was unrelated to income and education.