Does the Sample Size Matter in Estimating Toll Elasticity

Abstract

The authors discuss how San Francisco Bay area automobile traffic across the Golden Gate Bridge was impacted by the toll hike in September 2002. Using a recursive estimation technique, repeatedly updated toll elasticities for each day of the week are obtained. Results suggest that researchers analyzing given fixed sample size based toll elasticity estimates should be cautious, and that two or more years are needed for bridge users to fully adjust to a toll increase, since estimates begin to stabilize after 24 to 30 months of considerable instability being exhibited.