Safety Evaluation of Cable Median Barriers in Combination With Rumble Strips on Divided Roads

APPENDIX D. METHODOLOGY FOR CALCULATING CRASH COSTS

This appendix presents the methodology for estimating cost per crash for current year by severity, crash type, and speed limit based on value of a statistical life.⁽¹⁸⁾

The total cost of a fatal crash is larger than the value of a statistical life since there can be other injuries as well as property damage. The question is: how much larger? That is, how much is the cost of a fatal crash, given the value of a statistical life?

The relationship between value of a statistical life and cost of a crash can be derived by using latest comparable information from two sources involving the same research team. For 2001, the value of a statistical life (VSL) recommended to the National Highway Traffic Safety Administration was $3.8 million, and the mean comprehensive fatal crash cost irrespective of location or speed limit was $4,008,085, giving a ratio of cost per fatal crash to value of a statistical life of 1.055.^(19,20)

To update the cost of a fatal crash to 2014, for example, this ratio is applied to the VSL in that year ($9.2 million) as given annually in a USDOT memo to get an updated 2014 cost of a fatal crash of $9.7 million.⁽²¹⁾

For economic analyses in FHWA DCMF evaluations, researchers typically use total crashes (all severities combined), PDO crashes and fatal plus injury crashes combined, sometimes disaggregated for more precision by crash type, site type (e.g., signalized intersections, road segments), and environment (speed limit > 45 and > 45 mi/h). These costs are usually derived from the 2005 FHWA crash cost report using the 2001 costs and updating them using the procedure recommended in that report.⁽²⁰⁾

It is recommended that researchers continue to use the basic disaggregate cost from the 2005 FHWA report (since that is the latest source of disaggregate crash cost, and disaggregation is desirable) and update it to current year considering the prescribed USDOT VSL number for that year. For example, for 2014:

2014 disaggregate crash cost = (2001 disaggregate crash cost/2001 fatal crash cost) x (updated 2014 fatal crash cost derived from 2014 recommended VSL)

Substituting known values in millions, rounded to one decimal place per the USDOT memo, produces the following result:

2014 Disaggregate crash cost = (2001 disaggregate crash cost/4.0) x (9.7)

In other words, the 2001 crash costs are factored up by 9.7/4.0 = 2.425 to update them to 2014.

For example, the fatal and injury crash cost for all locations and speed limits in the FHWA report is $158,177 for 2001, which factors up to $383,579 for 2014.

Continuing the 2014 example, the USDOT memo also prescribes that sensitivity analysis be conducted based on low and high VSLs of $5.2 and $13.0 million for 2014. These translate to 0.57 and 1.41 times the $9.2 million VSL recommended for use. By inference, sensitivity analysis should also be done for DCMF evaluations by estimating B/C ratios for 0.57 and 1.41 times the 2014 crash costs derived using the method above. For example, for the fatal and injury crashes, B/C ratios would be estimated based on costs of $218,640 and $540,846.