Estimating the External Costs of Driving in San Francisco
The hidden costs of drivingSeptember 1, 2005
Anyone who drives in San Francisco knows it's not cheap. Cars are expensive to purchase and maintain, and do not hold their value well. Gasoline and insurance are also expensive, particularly in San Francisco.
But what many drivers do not fully realize is that the cost to society of their driving is considerably higher than their own direct expenses—driving imposes many costs to people other than the driver. These costs are known as "external costs" or "externalities." Significant externalities include deaths and illnesses due to air pollution caused by vehicles; traffic congestion that causes lost time and wasted fuel; deaths and damages due to accidents; and the costs to government to provide road construction, maintenance, police, ambulances, street lighting, traffic signals, and more.
Earlier this year, SPUR asked me to prepare a study to estimate the magnitude of each of these external costs in San Francisco, and also to discuss the societal costs that occur due to inappropriate pricing of parking. While several studies of this nature have been done at the national level, no such analysis has been done recently for San Francisco. The analysis reveals that current policy fails to account for a substantial proportion of the external costs of driving. The full study is available online at www.spur.org.
Most drivers realize that driving creates air pollution and damages roads, and that there are costs associated with these outcomes, including congestion and accident externalities. However, most drivers would be surprised to learn the extent of these costs. The estimates indicate that the annual cost to the residents of San Francisco of each of these four externalities runs into the hundreds of millions of dollars. The total external cost to the City and its residents of driving probably exceeds $1 billion annually. The cost per gallon of fuel used is likely an additional $5 or more in external costs alone.
The external cost per vehicle-mile in different driving conditions and for different vehicles varies significantly. For example, a car driving down a deserted city road at 3:00 a.m. would impose no congestion externality at all, and a very low accident externality per vehicle-mile; conversely, a car on a freeway in rush hour would impose congestion and accident externalities that might be many times higher. Similarly, a large, heavy car with poor fuel economy would impose much larger air emissions and road maintenance externalities than a small electric vehicle. (The study only considers passenger vehicles; buses, freight trucks, and the like are not included.)
These costs are somewhat difficult to conceptualize, because it may seem that no one actually "pays" them. They may seem to be costs on paper only, that no individual need account for. Indeed, that is the very crux of the problem of external costs—absent policy intervention, no one is forced to account for them. But these costs are very real. Air pollution causes death (estimated at 50 to 90 deaths in San Francisco per year) and chronic illness, as well as minor ailments. Accidents also kill people, and the external cost of driving in congested conditions appears directly in car insurance rates, which are higher in San Francisco than in most other parts of the country. Time delays caused by congestion cause the average San Franciscan to spend about 40 hours a year on the roads—equivalent to a typical work week—that she would need not spend if traffic flowed freely. These hours could be spent earning money, spending time with family and friends, or doing work around the household, each of which has value. Total San Francisco City government expenditures on driving-related operations and services are about $180 million (about $240 per San Franciscan); only a fraction of this budget is paid by drivers in the form of user fees, and the rest is paid through taxes.
The following is a summary of the largest external costs of driving in San Francisco, along with a brief description of who bears those costs and how they are calculated. For a full discussion of each of these and other costs, along with detailed calculations, please refer to the full report at www.spur.org.
Harmful pollutants emitted by passenger vehicles include particulate matter (PM), carbon monoxide, oxides of nitrogen (NOx ), oxides of sulfur (SOx ), lead, various hydrocarbons (including reactive organics and carcinogenic compounds), and carbon dioxide. These air pollutants are associated with a wide range of health impacts, from headaches and eye irritation to respiratory illnesses, asthma, heart disease, cancer, and deaths. Carbon dioxide and other hydrocarbons, as well as NOx, contribute to global warming.
The literature on the external costs of air pollution from motor vehicle activities demonstrates that health costs associated with PM emissions account for well over 90 percent of the total costs of motor vehicle–related air emissions. Our study expresses the health costs of PM emitted from cars in San Francisco in economic terms by estimating the fraction of airborne PM that results from motor vehicle–related activities, the health effects of these emissions, and the value these health effects in dollar terms.
The analysis indicates that particulate matter pollution in San Francisco caused by motor vehicles is responsible for 50 to 90 deaths per year. Using standard economic techniques to convert this into dollar terms, these deaths have an economic value of $280 million to $550 million. This translates into an average cost per vehicle-mile of $ 0.07 to $0.13. Expressed in per-gallon terms, the cost is approximately $1.30 to $2.50 per gallon—only slightly less than the current sale price of gasoline. This is a conservative estimate of the total annual external costs of air pollution due to driving in San Francisco for two reasons: it accounts only for deaths due to particulate matter (and not for illnesses or deaths resulting from other pollutants), and it does not include the potentially significant harm done to people outside San Francisco, as some pollution is blown across the Bay and much of it contributes greenhouse gasses.
External Accident Damages
The presence of a car on the road affects the accident risks that other users of the road face. External accident damages are those accident costs that a driver causes and does not personally face. When discussing accident externalities, one must consider causation as distinct from fault or negligence. If a car careens wildly in the wrong direction on a one-way street, but no other cars are present, no accident will occur. The presence of at least one other car (or bus, or bicyclist, or pedestrian) is required. In that sense, while the driver of the first car is the negligent party, the second car contributes to causing the accident. The presence of that car markedly increases the accident risk for both cars. Thus, driving imposes costs on others as a result of the additional risks it creates.
The estimate of accident externalities in San Francisco is based on a previously published statistical model that estimates the effect that traffic density has on insurance premiums and on loss costs to insurers, controlling for other variables such as alcohol consumption and precipitation. This approach estimates the accident externality that is imposed on other drivers by demonstrating the extent to which the premium (or loss cost) depends on how many vehicles are on the road.
This approach reveals that a typical car driven in San Francisco causes insurance premiums of other drivers to go up by $400 to $900 in aggregate. Expressed per vehicle-mile, this is a cost of $0.04 to $0.09. Per gallon, the cost is $0.80 to $1.80. However, this value includes costs only to the extent that they are compensated by insurance. Insurance often does not compensate accident victims for the pain and suffering they endure, but rather for their health expenses. Moreover, fatalities in accidents are significantly underinsured. One author estimated that the full cost of accidents is as much as 3.5 times the insured cost. Using this value, external accident costs could be as high as $0.15 to $0.33 per vehicle-mile, or $2.80 to $6.30 per gallon.
The act of driving imposes a time cost on all other drivers. The presence of a car in the traffic flow slows travel for all other drivers. Because this time cost is borne by other drivers, not the decisionmaker, this cost is external to the individual's decision to drive.
The relationship between congestion and costs is not linear: the amount of time a car adds to others' travel times in heavy traffic far exceeds the added time in light traffic. In fact, in light traffic conditions, cars may impose no additional travel time on each other. This is the case if traffic on a roadway is moving at free-flow speeds despite the vehicle's presence. For much off-peak driving, this is the case. On the other hand, driving done in extremely congested conditions—say, on the I-80 approach to the Bay Bridge in San Francisco during the afternoon commute—could create external costs several times as high as the average costs calculated here. Therefore, the congestion cost of a vehicle-mile will vary widely depending on time and place.
Driving in traffic also imposes costs on other forms of transit. It may be more difficult to walk or bike in areas of heavy traffic. In San Francisco, traffic certainly imposes travel costs on MUNI buses and light rail. This study only considers the time cost to individual drivers, and so the estimate of the congestion externality is a conservative one.
The external costs associated with congestion can be estimated by calculating the total time lost and the additional fuel cost due to congestion in San Francisco and placing a value on that time and fuel. Applying this methodology yields an estimated cost of $0.06 to $0.08 per vehicle-mile in time costs (depending on the value of time) and $0.01 per vehicle-mile in extra fuel. The total estimated cost of congestion per vehicle-mile is $0.07 to $0.09 for an average vehicle-mile, or $1.30 to $1.70 per gallon.
Road Operations, Road Maintenance, and Roadway Services
This category is composed of costs incurred to operate and maintain San Francisco roads and to provide services to roadway users. Costs in this category include street repair and resurfacing (done by the Department of Public Works), police, emergency response (fire department and ambulance costs attributable to driving), transportation planning, legal costs for traffic-related incidents, lighting, and parking enforcement. Most of these costs are incurred by City government, although roadway users do pay for some of the costs they create in the form of bridge tolls, fuel taxes, and registration fees.
Taxpayers bear this externality, since costs in this category that are not paid by user fees are paid by general tax revenues. To the extent that some of these items are funded by user fees that do not vary with the amount of driving, individuals who drive fewer miles are bearing some of these costs on behalf of those who drive more miles, create more need for maintenance, and demand more roadway services.
For the most part, costs in this category can simply be totaled from government agency budgets, adjusted for expenses are driving- or traffic-related. The total annual cost of the items discussed above is estimated at $180 million. Per vehicle-mile, this corresponds to a cost of $0.045. Per gallon, the cost is $0.84. This number is notably higher than gasoline excise taxes per gallon, which total $0.36. Although the gasoline tax was primarily instituted to internalize these costs, it actually internalizes less than half of their total for the average vehicle-mile.
There are many other categories of external costs of driving not detailed here. They include the ecological and financial costs associated with petroleum extraction; water pollution from leaking oil and other fluids; vibration damage to buildings resulting from cars passing by; waste disposal of old cars and parts; and noise pollution from vehicle traffic. While these costs are each likely smaller (in some cases, much smaller) than those discussed above, they are nonetheless real. Since the total costs reported in the study do not include these costs, they should be viewed as conservative estimates of the actual external costs of driving in San Francisco
Implications for Public Policy in San Francisco
The existence of these externalities creates a clear rationale for policy intervention to "internalize" them—in other words, to force the drivers who are imposing the costs to pay for their mitigation. The goal of such intervention would not be punitive; rather, drivers will recognize the true cost to society more accurately if they are forced to pay that cost. While San Franciscans are a socially minded lot, it is very difficult for even the most socially minded to mentally "charge" oneself an extra $5 for a gallon of gasoline and make driving decisions as if that were the case. Failing to adequately do so, the typical driver will undervalue the cost to society of her driving, and will drive more, and cheaper, than she would otherwise. Internalizing the costs would ensure that drivers consider these factors.
Current policy does internalize a small fraction of these costs through gasoline taxes, parking fees, and tolls. However, substantial external costs remain unaccounted for. There are many different possible interventions, and the City of San Francisco does not necessarily have the authority to impose many of them on its own. Perhaps the two most promising policies that the City might wish to consider on a local level are revising parking prices (particularly in areas where people park during the workday) and imposing a congestion charge to enter the city during high-traffic times. These policies would have a similar effect: raising the price of peak-hour commuting, which imposes external costs per mile driven that are far greater than the average figures discussed above (particularly for congestion and accidents) due to high traffic density. Careful analysis and policy design would be required to set appropriate prices.
Regardless of what form policy intervention eventually takes, it should recognize the problem of unrecognized external costs associated with driving in San Francisco, and encourage people to consider these costs each time they choose a mode of transportation.