A rising tide: Measuring the economic effects of broadband use across California

Abstract

This research estimates the economic effects of broadband use across 39 California counties between 2001 and 2006. Broadband usage data are obtained from household surveys that identify how individuals access the Internet: through dial-up or broadband, where broadband is defined as Internet access through cable modem or digital subscriber line (DSL). The surveys were conducted in California counties twice each year from 2001 to 2006. Estimates from panel regression analysis demonstrate that increased broadband use within California has had a positive and significant effect on growth in employment and total payroll.

Acknowledgements

I am grateful to Stephen Perez, Ryan Sharp, and two anonymous referees for providing useful comments and feedback. I thank Helen Schaubmayer, Angela Owens, and Lindsay Cox for excellent research assistance. This research was made possible by the Sacramento Regional Research Institute, a joint venture of the Sacramento Area Commerce and Trade Organization (SACTO) and California State University, Sacramento. I am indebted to Kearns & West and AT&T for providing access to proprietary Scarborough Research data.

Notes

1 The 39 counties covered in the Scarborough Research data account for 92% of California's population. Excluded counties are primarily comprised of rural agricultural communities.

2 The remaining share of the adult population answered either “none,” meaning no Internet usage, or “other” meaning the individual accessed the Internet through some other source. The survey data used in this study does not distinguish between “none” or “other” at the county level.

3 The surveys ask households about their Internet access between February–July and August–January for some counties. Other counties use a March–August and September–February time frame.

4 Given the availability of data and the limitations on the time dimension of the panel data set, only county fixed effects are reported. The Hausman test for the regressions uniformly rejects the random effects specification.

5 The survey data used here do not distinguish between household broadband use for work versus leisure activities. Since households likely use broadband access for both, the estimates provided in the following section may underestimate the true effect of work-related broadband use on economic activity.