Guide for Conducting Benefit-Cost Analyses of Multimodal, Multijurisdictional Freight Corridor Investments (2017)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

76 S t e p 7 7.1 Goal The goal of this step is to quantify negative externalities associated with projects, including safety-related externality measures and environmental emissions, and conduct a valuation of these externalities. These are unintended consequences that are unpriced or those that occur from usage but are not paid for. These externalities can include: • Air pollution effects at the local and regional scale (sulphur dioxide [SO2], carbon monoxide, nitrogen oxides [NOx], particulate matter [PM10], volatile organic compounds [VOCs], and other gases). • Greenhouse gases at the global scale such as carbon dioxide (CO2) and other gases valued using the social cost of carbon (SCC) (53). • Noise pollution. • Water pollution. • Congestion. • Accidents/safety (example, grade crossings, highway projects). • Land use effects such as visual intrusion or habitat effects. • Health effects induced from air quality. • Security implications or risks from hazardous material routing. Air pollution, CO2 greenhouse gases, and accident-related safety effects have been well studied in terms of valuation methods. For many other externalities, non-market valu- ation methods have not generally established monetary valuation parameters, but they may be considered on a case-by-case basis. The state of the art continues to evolve in terms of developing suitable damage or external cost estimates. Health effects are increas- ingly becoming critical in congested ports and other regions for which the U.S.EPA has developed methods to calculate mortalities and deaths from pollution and monetize the health effects via open-source procedures. Security risks or implications of risks faced by the facility, region, or nation can be considered on a case-by-case basis separately. Secu- rity risks are hard to monetize in BCA. On one hand, a security risk may be considered from a loss reduction perspective, however, it must be evaluated within a probabilistic framework. On the other hand, the losses themselves can be short or long term, but often examined as part of an economic impact analysis using computable general equilibrium models rather than BCA. This is very different from measures taken to address these exter- nalities which are internalization approaches. Even if some externalities are difficult to quantify or difficult to monetize, BCA should document those and discuss them (via sepa- rate accounts). Analyze Public Externalities and Information Needs (Safety and the Environment)

Analyze public externalities and Information Needs (Safety and the environment) 77 7.2 Tasks Quantify Externalities The two main categories for externalities in a BCA are safety and the environment. Their treatment has been well studied and documented. A third category is noise pollution, but it is used much less in practice. Simpler ways of approximating externalities associated with safety, emissions, and noise rely on changes in ton-mile or distance traveled measures such as vehicle miles traveled (VMT). An alternate way relies on measuring the activity itself associated with the negative externality. Safety Safety has different dimensions including safety from accidents and safety/security of cargo. This step discusses the former, while the latter is assumed to be measured by cargo aspects (e.g., cargo mix) and costs associated with the mix (e.g., costs associated with demurrage). Cargo mix and costs can be determined from Step 5 benefit measures, assuming the inventory distribution of cargos is known. The most common mechanisms for quantifying accident-related safety risks are based on accident rates linked to vehicle miles or ton-miles as a metric to calculate safety-related fatality and injury costs. Use Table 12 to identify data sources for the development of accident metrics based on fatality and injury rates, along with safety statistics for the facility. Mode Fatality Statistics and Rates Valuation Parameters Data Sources (Value of a Statistical Life) per Fatality and Value of Injuries Highway— trucks and passengers State DOTs and metropolitan planning organizations (MPOs): statistics, Bureau of Transportation Statistics (BTS), National Transportation Statistics (several years) Kruse et al. (54): fatality statistics for highways, rail, and water comparisons (rates per million ton-miles) Government Accountability Office (GAO) study (55 ): fatalities and injuries (rates per billion ton-miles) USDOT TIGER guidance (value of a statistical life [VSL]) $9,200,000 (2013 $) (48). Used in all TIGER grant Rail FRA, Office of Safety Analysis, BTS, National Transportation Statistics (several years) Kruse et al. (54): fatality statistics for highways, rail, and water comparisons GAO study (55): fatalities and injuries (rates per billion ton-miles) USDOT TIGER guidance (VSL). Used in all TIGER grant studies. Updated yearly. This parameter is also used to value crash risks by category for grade crossings (56) Inland waterways BTS, National Transportation Statistics (several years) U.S. Army Corps of Engineers Kruse et al. (54): fatality statistics for highways, rail, and water comparisons GAO study (55): fatalities and injuries (rates per billion ton-miles) USDOT TIGER guidance (VSL). Used in all TIGER grant studies. Updated yearly. Ports American Association of Port Authorities USDOT TIGER guidance (VSL). Used in all TIGER grant studies. Updated yearly. Airports FAA USDOT TIGER guidance (VSL). Used in all TIGER grant studies. Updated yearly. FAA provides its own guidance on VSL and estimates it based on meta- analysis. The 2008 estimate for VSL was $5.8 million. studies. Updated yearly. GAO study (55) (fatalities and injuries). National Highway Traffic Safety Administration Fatal Analysis Reporting System (FARS). Table 12. Accident rates and valuation of fatalities across modes.

78 Guide for Conducting Benefit-Cost Analyses of Multimodal, Multijurisdictional Freight Corridor Investments Environment Other negative externalities associated with freight investments are emissions and the envi- ronment, health-related impacts, and climate change effects. In the United States, a complete environmental effects screening is conducted as part of environmental assessments for federal projects. BCA seeks to internalize public externalities by monetizing those for which theoreti- cally agreed-upon methods and values exist for including them in the BCA, separate from the environmental assessments. Pollutants Emitted. The most typical ways of measuring impacts on the environment are in terms of tons of pollutants emitted or emissions exposure based on emission factors. CO2 greenhouse gas emissions are used to study effects on climate change. Emission factors are an important input to the valuation of emission impacts, along with activity measures such as truck volumes, freight tons, and ton-miles, to determine emissions. The simplest approach is to use default rates for PM10, NOx, CO2, SO2, and VOCs compiled in the literature for all modes. The use of these default rates should be examined in light of the specific context and/or facility. EPA provides guidance on emission factors for freight rail locomotives in its non-road vehicle current and projected emission factors directory for hydrocarbons, NOx, PM10, SO2, and CO2 (in grams per gallon) including procedures to convert to grams per ton-mile (57). Rail emissions can also be estimated by railroad- and terminal-provided information and from published informa- tion sources such as EPA’s Regulatory Support Document (58). Regional MPO models provide a source of emission factors for trucks. EPA’s Motor Vehicle Emission Simulator (MOVES) model allows estimation of emissions for any scale (local, regional, state, and national). The EPA emission factors for trucks are shown in Appendix H in grams per mile. Cross-modal emission factors are included in Argonne National Laboratory’s GREET model. The California Air Resources Board’s EMFAC model is another source of emissions factors. Fuel and Power Consumption. Activity measures of emissions for rail, for instance, can be developed in a variety of ways associated with a project when emission-reducing options are undertaken as part of a project (e.g., including sustainable or “green” solutions like the use of more fuel efficient engines by the providers, operators or carriers for internalizing emissions-related externalities as part of the project). For instance, they can be developed from changes in fuel consumption, changes in freight volumes, changes in container traffic, and changes in power consumption as well as simulation models. Fuel consumption activity measures are also useful for measuring variable direct operating cost changes. Using train trips as the measure of rail traffic, fuel consumption, and power consumption can be mea- sured in gallons per year by Equations 16–18. Similar measures can be developed for truck trips and barge trips. Fuel Consumption Year train traffic train trips year locomotives train distance miles speed miles hr fuel rate or intensity gallons hr (16) ( ) ( ) ( ) ( ) ( )= × × × × Fuel Consumption Year rail traffic containers year distance miles speed miles hr tons container fuel rate or intensity gallons revenue ton-mile (17) ( ) ( ) ( ) ( ) ( )= × × × × Power horsepower Year train traffic train trips year travel time hr trip average power horsepower hour (18) ( )( ) ( ) ( ) = × ×

Analyze public externalities and Information Needs (Safety and the environment) 79 The most recent emission factors are as follows: • For locomotives: for large line-haul and other categories and Tier 4 locomotives for power consumption, EPA in 2009 forecast (57). • For shallow-draft and deep-draft vessels: generally hard to come by. EPA published the most recent standards for inland marine vessels in 2008. • For oceangoing vessels: EPA in 2010. • For trucks: shown in Appendix H in grams per mile. • For different types of freight rail: shown in Appendix H in grams per ton-mile. Rail emissions can also be modeled at the national state, and regional scales following the project geography and then apportioned down to states and counties using rail density measures like gross ton-miles. Appendix H also shows emission rates (in grams per ton-mile) across modes for use with simpler ton-mile approximations instead of direct activity variable approximations or use of simulation models. The emission factors can be based on local values as available to account for non-attainment regions and geography. Noise Pollution The noise reduction effects of transport projects can be substantial based on project location and mitigation or abatement or internalization strategies that accompany the alternative being evaluated. Noise cost considerations can be important for projects located in urban residential locations or congested regions. A few models can be used to estimate actual changes in decibels (the common measure of noise) associated with a project for use in BCA. Highways. Historically, noise pollution has typically been included as part of the environ- mental impact assessment and regulatory compliance. This is true for model-based methods such as the FHWA Traffic Noise Model (TNM), which simulates the noise reduction or noise measurement methods. FHWA also has a policy of including cost allowances for noise abatement aspects of proj- ects, which is handled as an implied BCA in determining the reasonableness of the mitigation or abatement method (e.g., sound walls). There is a cost allowance associated with any noise- impacted residence for a proposed highway project (59). These cost allowances range from a low of $10,000 per residence to a high of $50,000. The cost allowance for all benefited residences that receive a 3- to 5-dB reduction from a proposed sound wall are then totaled, and this cost is compared to the cost of the sound wall, using a process specific to individual states. Rail. FRA’s guidance on assessing noise and vibration impacts (www.fra.dot.gov) for rail projects is based on FTA guidance; however, FRA does not currently have an established proce- dure for valuing noise reductions. Airports. An ACRP project evaluated 24 airport BCAs and found that none include noise pollution (60), even though the FAA guidance is explicit in its consideration of hard-to-quantify benefits. FAA recommends inclusion of noise in BCA. It has now developed its own methods for including noise in BCA and provides its own models for estimating the noise impacts of capacity projects (the Integrated Noise Model). FAA’s methods are documented in a recent report Technology for a Quieter America (61). Much like EPA’s Benefit Analysis and Mapping Program (BenMAP) and Co-Benefits Risk Assessment Model, it theorizes that the metric used should be translated into an effect. The literature indi- cates that it is acceptable to use a day and night average sound level (DNL). The DNL value is a 10-dB increase in measuring system gain at night for the evaluation of community noise. FAA

80 Guide for Conducting Benefit-Cost Analyses of Multimodal, Multijurisdictional Freight Corridor Investments recommends using the percent of persons highly annoyed for urban areas. For rural or otherwise naturally very quiet areas, it suggests thresholds of 40 db. As far as noise impact or impact of noise mitigation measures in BCA, FAA recommends the use of hedonic studies using parameters of noise depreciation indices from the estimate equations as measures of willingness to pay. Select Metrics for Valuation of Externalities Accident costs can reflect several different costs including those inflicted on personal prop- erty, lost productivity, grief associated with loss, and minor to major disruptions to traffic, many of which are non-market costs. The most commonly used parameter for valuing costs associated with safety risks is the VSL provided by USDOT (48) and is updated every year. VSL, Value of Property Damage, and Value of Injuries USDOT provides the VSL, value of property damage, and value of injuries categorized by six different Abbreviated Injury Scale rating systems for highways. USDOT also indicates how to translate the KABCO scale to Abbreviated Injury Scale ratings. Use the same USDOT values across modes to ensure consistency in cross-modal usage. When safety benefits are the explicit drivers and are emphasized, a more detailed analysis may seek to quantify and monetize them using change in lost productivity (e.g., changes in effective work life or changes in life years).10 Emissions Consider using the values provided by USDOT for the monetization of emissions (48). The USDOT values are developed from the NHTSA’s documentation rule making.11 Using the national parameters provided by USDOT for all modes ensures consistency. These values are updated every year for the TIGER grant program. With respect to greenhouse gas emissions such as CO2, USDOT follows EPA in prescribing an approach to establish the SCC using a 3% discount rate. The SCC is an estimate of the mon- etized damages associated with an incremental increase in carbon emissions in a given year. The Whitehouse Technical Update (62) also recommends a 5–7% rate for situations where global emissions are impacted. Also called the shadow price of carbon, SCC is the monetary indica- tor of the global incremental damage done by emitting greenhouse gases today, CO2 being the major one. The values provided by USDOT come quite close to the values developed by Tol (63), who provided estimates from a meta-analysis of 203 studies and reported that marginal damage costs of CO2 are unlikely to be much higher than $50 per ton under standard assumptions and discounting. Locally adjusted parameters can be used if local values differ significantly from national aver- ages, especially if counties are designated as non-attainment areas. State-of-the-art integrated assessment models such as the Air Pollution Emission Experiments and Policy (APEEP) analysis model for the United States12 (64) allow damages to be established at local levels, which internalize economic consequences of both air pollution and health (Appendix H). Another tool that allows detailed analysis of emissions, air quality, and implied health impacts is EPA’s open-source 10These are called disability adjusted life years, which are accepted global health indicators and are now being used to translate safety to health outcomes. These have yet to be applied in the United States but can be of value to examine when safety and emissions are impacted in major ways. 11USDOT values are based on Final Regulatory Impact Analysis of the National Highway Traffic Safety Administration’s Rulemaking on Corporate Average Fuel Economy for MY 2011 Passenger Cars and Light Trucks. 12Like EPA, APEEP (employing the EPA national emission inventory of air pollution emissions in the United States) calculates the resulting air pollution concentrations across the country. The model develops emission levels and expresses emissions in monetary terms. Models like these are time intensive to use in conceptual studies but could be important if detailed analysis is called for.

Analyze public externalities and Information Needs (Safety and the environment) 81 BenMAP, which relies on willingness to pay and cost of illness methods to determine the eco- nomic value of illness. Given the emission factors, CO2 emissions can be calculated using the following approach (Equations 19 and 20). Ton-mile estimates can be used to estimate emission costs for all categories (Equation 21). . (19)2 13CO Emissions Distance No of TEU Emission Factor= × × (20)2 2Grams of CO Distance miles TEU g of CO TEU miles[ ]( )( )= × × × - (21) , , 1 Emissions Related Benefits i VMT or FMT emf VMT or FMT emf Ci build k i dominimum k k i n ∑[ ]( ) = × − × × = In general, emission costs for an entire corridor can be developed for each pollutant category (j) using Equation 10. VMT or FMT is vehicle miles or freight ton-miles for all links (i) in the cor- ridor, emfk is the emission factor (pollutant k), and Ck is the damage cost value. Equation 7 should be summed over all pollutant categories. Similarly, emissions can be directly estimated using Equations 5–7 and either fuel or horsepower consumption. Several estimates of damage costs are provided in the literature for mobile sources (see Appen- dix H). If there are extenuating circumstances in a given scenario, these sources may provide better valuation methods. Noise Appendix F provides monetization parameters for use in noise cost valuation in cents per mile and cents per ton-mile. These parameters have been compiled in the literature. In most studies, distance and/or ton-mile impedance change is the simplest way to approximate the value; it can be used along with the social cost of noise parameters found in Appendix H. However, much like damage cost methods and health effects assessments, a detailed valuation of noise reductions is typically conducted using a contingent valuation study or a hedonic study, as recommended by FAA. Should a decibel measure be used and a noise depreciation index calculated, the economic cost of noise level increases is captured as a measured loss in house value per decibel per year. Tools for national analysis allowing the integration of such noise cost measures have been devel- oped as part of research conducted at the Massachusetts Institute of Technology (the Aviation Environmental Portfolio Management Tool). It has yet to be implemented in actual BCAs. The heart of FAA’s approach and the Portfolio tool to noise valuation is driven by a benefit transfer approach based on hedonic regressions of property values. Include the urban costs of freight rail noise in BCA; this allows quantification of noise abatement solutions. Rail routing decisions involve tradeoffs between moving trains on certain areas of track versus others; some cities, towns, and regions will experience more train traffic as a result of train re-routing, even if such routes pose the lowest overall risk to society. As a result, these adjacent areas (the affected community) will bear more relative risk and may experience secondary effects, such as increased air and noise pollution or reduced residential property values near freight rail corridors. This guidebook recommends that a promising approach to valuing noise costs of freight rail is also one that relies on examining the link between freight rail and residential real estate values. 13TEU stands for twenty-foot equivalent unit, which is a standard unit of measure for containerized freight. A container that is 20 feet long is 1 TEU; a container that is 40 feet long is 2 TEU.

82 Guide for Conducting Benefit-Cost Analyses of Multimodal, Multijurisdictional Freight Corridor Investments This approach is similar to the FAA’s benefit transfer approach, which is essentially applying values from a meta-analysis of empirical studies. Futch (65) notes that an increase in rail traffic of 10 million gross ton-miles per mile causes a 0.7 percentage point lower growth in home values within a 1⁄3-mile band around the tracks along the Alameda Corridor. These estimates merely provide a starting point for establishing a noise cost value estimate for freight rail noise in urban areas. These reports do not, however, explicitly link home values to a noise level analysis, which suggests that these methods still need to be developed further. Updating Valuation Parameters The default cost values from USDOT should correspond to the year closest to the base period used. USDOT does not recommend updating emissions costs, VSL, or other parameters such as value of time since the USDOT undertakes these updates using procedures it has established. However, valuation parameters for externalities such as noise costs may be done using consumer price index or gross domestic product deflators. Damage costs from emission models should note the base year for the model costs. If the tool recommends procedures for updating values, it is best to follow them to update the costs. Updating willingness-to-pay estimates for noise costs such as those provided by the FAA tool is a non-trivial task and often requires separate stud- ies commissioned specifically for that purpose. In the interim, the analyst can use procedures similar to other parameter updates. However, the analyst should invest in developing a first cut estimate of noise valuation potentially based on willingness-to-pay measures analogous to FAA’s, particularly for urban areas. Monetary values of externalities can also be depicted as a probability spectrum to account for uncertainty in the range of monetary values of damage costs. The divergent perspectives on cost values per ton of pollutant can also be represented as a distribution of potential values. The same can be said of operating costs and shipping costs as well as values of freight time and reli- ability across modes. Review Federal Funding Guidelines for Reporting of Specific Externalities For projects that receive federal grant funding, review the specific criteria and reporting guide- lines of the grant program. For example, if grade crossings are part of a project, federal guidelines require that safety effects be quantified. For such projects, Section 130 (23 U.S. Code § 148), which is continued under MAP-21 under the Highway Safety Improvement Program, provides a data-driven funding tool that is guided by public benefits in terms of safety applicable to grade crossing projects. Check on the need for National Environmental Policy Act (NEPA) compliance. In the United States, most large-scale projects with federal aid are required to include an environmental justice analysis that covers the specific categories via environmental impact statements. An environ- mental impact statement is, however, separate from a BCA. Screening based on project types and regions traversed can help with identification of equity issues (disadvantaged groups and land uses). The screening can be done using public domain tools such as Census Tiger Files (STF3), aerial imagery data, and geographic information system-based accessibility measures as well as using detailed models discussed earlier (noise, emissions, and health effects). This helps identify issues associated with the specific siting of a facility and its locational context, and can be conducted during project studies early on and considered more rigorously as part of the feasibility analysis. If such compliance is needed, an initial BCA could anticipate such public costs by including triggers in the BCA. A feasibility study, on the other hand, should draw from alternatives compared in the environmental documentation. Many tools discussed in this sec- tion are sometimes used as part of preparation of environmental documentation but are rarely

Analyze public externalities and Information Needs (Safety and the environment) 83 considered as part of a feasibility BCA. As an example, a BCA that pays due attention to emis- sions and their distribution on communities using suitable tools listed in this guide can be used for environmental justice analyses and for addressing equity. 7.3 Inputs: Recommended Tools and Data Sources A number of tools can assist the analyst with analyzing public externalities and informa- tion needs: • Appendix F, Table F1, noise social cost values. • Table 12 and Appendix H for emission factors. • Railroads and terminals (direct contact). • EPA Regulatory Support Document (revised 1998) (58). • EPA Motor Vehicle Emission Simulator (http://www3.epa.gov/otaq/models/moves/). • EPA marine information (http://www3.epa.gov/otaq/marine.htm). • EPA 2009 Emission Factors for Locomotives (https://www3.epa.gov/nonroad/locomotv/ 420f09025.pdf). • California Air Resources Board (CARB) Emission Factors (EMFAC) Model Truck and Auto Emissions (http://www.arb.ca.gov/emfac/). • BLS Consumer Price Index for updating emission cost estimates when using past studies. • FRA’s GradeDec for grade crossings (https://www.fra.dot.gov/Page/P0337). • EPA’s BenMAP for evaluating health effects (based on concentration response functions) (http://www.epa.gov/benmap). • Integrated assessment models—APEEP (https://sites.google.com/site/nickmullershomepage/ home/ap2-apeep-model-2). • EPA’s Co-Benefits Risk Assessment Model (COBRA) (https://www.epa.gov/statelocalclimate/ co-benefits-risk-assessment-cobra-screening-model). • Geographic Information System mapping and aerial imagery. • FHWA’s Traffic Noise Model. • FAA’s Integrated Noise Model. • Train Energy Model. • Aviation Environmental Portfolio Management Tool (http://partner.mit.edu). • FRA guidance and CREATE toolkit (http://www.fra.dot.gov/Page/P0216). • Guidance on Treatment of the Economic Value of a Statistical Life (VSL) in U.S. Department of Transportation Analyses-2015 Adjustment (https://www.transportation.gov/sites/dot.gov/ files/docs/VSL2015_0.pdf). More detailed local modeling could use public or private data sources on fleets and volumes and rely on detailed modeling for safety, emissions, or noise and their economic effect for consideration in BCA, for environmental justice considerations, or for specific public funding resources. 7.4 Best Practices and Examples Best practices for Step 7: • Clearly present all data, assumptions, and rates. • Match tools to the context. For instance, non-attainment areas must provide more in-depth analysis than attainment areas. • Choose whether to undertake a more detailed emissions and health assessment. This choice is driven by the context and extent to which emissions are known to be impacted and can be documented.

84 Guide for Conducting Benefit-Cost Analyses of Multimodal, Multijurisdictional Freight Corridor Investments • When relying on damage cost methods, ensure they are appropriate for the context and clearly present the methods used. • If the build alternative has specific mitigating measures or technology considered to internal- ize external costs, using an activity measure specifically linked to that technology is good practice for demonstrating the community benefits. Example 1: Lee et al. (66) studied the health impact of emissions from freight corridors using EPA’s BenMAP tool. The San Pedro Bay Port of Los Angeles and Long Beach is the largest container port complex in the United States. Although the benefits of handling and hauling freight are enjoyed by the nation as a whole, the negative effects of congestion and air pollution created by the port fall mostly on the people who live and work nearby and along connecting freight corridors. These corridors include two busy freeways—I-710 and I-110—and an active rail link—the Alameda Corridor. The study used an integrated model framework by drawing on traffic models, emission models, BenMAP, and a pollutant dispersion model. Example 2: Harris and Hanson Inc. (67) developed a spreadsheet tool for quantifying noise levels for the Chicago CREATE rail corridor. This tool is now included in FRA’s noise guidance methodology (http://www.fra.dot.gov/Page/P0216). It allows noise calculations from eight dif- ferent sources but does not consider them as part of a BCA. These calculations are, however, part of the National Environmental Policy Act regulatory compliance process. Methodological improvements in rail noise valuation could allow such measurements to be directly included in a feasibility BCA. Lack of valuation or monetization measures like this are cases of epistemic uncertainty. 7.5 Common Mistakes Common mistakes occur when the project team: • Uses unit values for a given safety or environmental metric that are not consistent across modes. • Uses units of emission factors and freight volumes that are not consistent and does not conduct suitable conversions. • Does not use an SCC approach to monetize CO2 emissions.