This paper measures the impact of a freeway management system (FMS) on the incidence of reported motor vehicle crashes in Phoenix, AZ. Using a fixed effects negative binomial regression model, I find that the FMS reduces the frequency of crashes involving property damage only, possible injury, and minor injury by 25, 30, and 21%, respectively. I find no evidence for an effect on the frequency of major injury crashes or fatal crashes, although such accidents account for less than 5% of the total crashed in Phoenix. Classifying the data by the type of crash rather than by severity, I find that the frequencies of rearend crashes and sideswipe crashes are reduced by 25 and 37%, respectively, and I find no evidence of an effect on single vehicle crashes. The results are robust to many different model specifications, including a variety of functional forms, covariates, and data. A conservative estimate of the annual crash benefits of the FMS in Phoenix ranges from $4.8-13.2 million, depending on various assumptions about the value of pain and suffering and about the extent of crash underreporting. These annual crash benefits far outweigh the $1.6 million in annual operating and maintenance costs of the Phoenix FMS, and they offset considerably the approximately $47 million invested to date in the design and construction of the system. This paper measures the impact of a freeway management system (FMS) on the incidence of reported motor vehicle crashes in Phoenix, AZ. Using a fixed effects negative binomial regression model, I find that the FMS reduces the frequency of crashes involving property damage only, possible injury, and minor injury by 25, 30, and 21%, respectively. I find no evidence for an effect on the frequency of major injury crashes or fatal crashes, although such accidents account for less than 5% of the total crashed in Phoenix. Classifying the data by the type of crash rather than by severity, I find that the frequencies of rearend crashes and sideswipe crashes are reduced by 25 and 37%, respectively, and I find no evidence of an effect on single vehicle crashes. The results are robust to many different model specifications, including a variety of functional forms, covariates, and data. A conservative estimate of the annual crash benefits of the FMS in Phoenix ranges from $4.8-13.2 million, depending on various assumptions about the value of pain and suffering and about the extent of crash underreporting. These annual crash benefits far outweigh the $1.6 million in annual operating and maintenance costs of the Phoenix FMS, and they offset considerably the approximately $47 million invested to date in the design and construction of the system. (Author/publisher).
Abstract