Guidance for Calculating the Return on Investment in Transit State of Good Repair (2019)

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.

19 C H A P T E R 3 This chapter details the steps involved in performing an ROI analysis for a proposed SGR investment or program of investments. The seven steps of an analysis are illustrated in Figure 3-1. The following sections describe these steps further and are intended to provide an overview of key concepts and specific calculation procedures. Where applicable, informa- tion is provided describing other resources with information on relevant supporting topics. ROI Calculation Guidance An Excel spreadsheet tool, termed the Return on Investment Calculator, has been developed to facilitate the calculation process and is described in the guidance. Further documentation on the tool is provided in Chapter 4. Step 1. Establish the Analysis Scope Overview The first step in analyzing the ROI of an SGR investment or program of investments is to define the analysis scope. The analyst must establish • Which assets are included, • What investments are being analyzed, and • What the timeframe is of the analysis. Figure 3-1. ROI calculation steps.

20 Guidance for Calculating the Return on Investment in Transit State of Good Repair These must be determined for both the investment case, which describes the investment being analyzed, and for a counterfactual base case, in which the investment is either not made or is deferred. Generally speaking, an analysis should extend to as broad a range of assets and as long a time period as is meaningful for decision making and for which the analyst has or can estimate data. Typically, this implies an analysis will include all of a transit agency’s assets of a given mode and asset class (vehicles, guideway, facilities, and/or systems), and over a period of up to 20 years. A broad range of assets and long time period are recommended so that all of the relevant effects of an investment or set of investments are addressed. Longer periods of time may be analyzed, but given future costs and benefits are discounted; from a practical standpoint, extending an analysis beyond a period of 20 to 30 years is unlikely to impact the overall results of the analysis. Note the Return on Investment Calculator is configured to include data for two transit modes. For analyses with additional modes, one should create multiple copies of the tool and sum the results. For instance, consider a hypothetical case in which a transit agency is evaluating a potential investment in a new fleet of buses to replace an aging fleet of buses. Although it may be possible to make the investment over a relatively short amount of time, a period of 10 to 20 years may be needed to realize the benefits of the investment in terms of reduced maintenance costs. Thus, in this example, the analyst would likely include all the transit agency’s buses in the analysis and may consider including any supporting facilities over a period of 20 years. The concept of having both investment and base cases is crucial. As discussed in Chapter 2, the approach recommended for the ROI analysis is adapted from a traditional BCA. For any such analysis, all benefits and costs of an investment are measured relative to a base case. The basic requirements for the base case are as follows: • That it represents what occurs if the investment is not made • That it should be realistic – It should describe a scenario one can readily imagine and describe • That, if one is analyzing multiple projects, the base case should be consistent between the different investments being analyzed – If different base cases are used for analyzing different investments, it can be hard to compare them In a traditional BCA, a common approach is to define as the base case a scenario in which the proposed investment is simply not made. With this approach, the effects of the investment over its life are analyzed relative to the case in which there is no investment. In the case of a hypothetical bus fleet replacement, if the fleet is assumed to last 15 years, then the investment case would be compared with a base case in which fleet replacement is deferred by 15 years, after which time the bus fleet would need to be replaced again even if the proposed investment is made. In analyzing other SGR investments, this approach may not suffice. For instance, take the example of an investment to replace a critical component of a track signaling system that, once replaced, is expected to last 20 years. In the event the replacement is not made, the system is bound to fail at some point, imperiling rail service on the impacted line. Unless the failed component is replaced, then rail service would effectively cease. This may not be a realistic scenario, however. In reality, in such a case the transit agency would presumably perform emergency work to keep the line in service. Thus, in this case the base case might be to assume that, in the absence of invest- ment, increased reactive maintenance work would be performed to keep the system in operation. As the above example demonstrates, the concept of not investing in maintaining or achieving SGR in any way often violates the requirement for defining a realistic base case. In defining the base

ROI Calculation Guidance 21 case, the analyst often must allow for some form of SGR investment in the base case, if only in the form of reactive maintenance. However, it is also important to keep the base case consistent between analyses that are being compared with one another. If reactive maintenance is considered in one case, it should also be considered in others. Alternatively, if an assumption is made that investments are deferred for a specific period, the assumption (e.g., five years) should be consistently applied. The issue of defining a consistent base case can be particularly vexing when comparing invest- ments made toward achieving SGR to investments made toward enhancing capacity, if maintaining SGR is taken as a given in the case of capacity enhancements. For instance, consider a case where the ROI of extending a light rail line has been calculated separately and is compared with an alter- native set of investments to replace the existing fleet and rehabilitate the guideway of the existing line. These cases cannot be readily compared unless the costs associated with maintaining the exist- ing line were incorporated in the analysis of extending the line. It is more likely that, in the analysis of the line expansion, it was assumed that the existing line simply continues in its current state. How, though, can a base case be defined to analyze a set of SGR investments if maintaining SGR is taken as a given? An alternative in this case is to define the base case as one in which the existing transit system is maintained, consistent with the assumptions made in the case of capacity enhancement, and the investment case as one in which the transit system is allowed to deterio- rate. This is the approach used by the Metropolitan Transportation Commission (MTC) in its analysis of the benefits of achieving SGR (1). One obtains from such an analysis the disbenefit of allowing the transit system to deteriorate. Though this may seem counterintuitive, it results in a consistent base case, facilitating comparison of different investments. Procedures 1. Define what modes, asset classes, and specific assets are included in the analysis. At a mini- mum, the analysis should include any assets being rehabilitated or replaced. The analysis may also be extended to other assets in the same class as those being replaced, and/or to other asset classes, depending on the expected impacts of the investment. 2. Determine the start year and time period of the analysis. A period of 20 years is recommended as the default timeframe to address all effects of an SGR investment. Note this is the period over which costs and benefits are calculated, not the assumed life of a transit asset (which may be much longer). However, a shorter or longer period may be warranted depending on the nature of the investment and the period over which it will have an effect. Enter these values in the Return on Investment Calculator, as shown in Figure 3-2. Figure 3-2. Enter analysis year and analysis period.

22 Guidance for Calculating the Return on Investment in Transit State of Good Repair 3. Define the base case against which the investment case will be compared. Examples of base and investment cases are described previously and listed in Table 2-1. Typically, the base case will be one in which the investment is not made over the analysis period being considered or is deferred for a specified period. The base case may include additional maintenance that is triggered as a result of the deferral of needed investments or, alternatively, may incorporate impacts from loss of service that may result if assets fail. 4. Document the definitions of the investment and base cases and the assumptions made concerning the analysis scope. 5. Compare the analysis assumptions to those made for any investments that will be evaluated together with the investment being analyzed. Revise the analysis assumptions as necessary, such that a realistic and consistent base case is used for all investments being compared. Additional Resources The FTA Asset Management Guide discusses different asset classes and their characteristics (2). TCRP Report 78 is a guidebook for performing benefit–cost analysis for transit investments and discusses key concepts in BCA (3). Step 2. Calculate Agency Capital Costs Overview Once the scope of the analysis has been established, the next step is to determine the capital costs for both the investment and base case. Here the analyst must specify • Costs by year for rehabilitation or • Replacement of transit capital assets. This will include the initial cost of the investment and may include other costs over time, depending on the scope of the analysis. As an example, in the hypothetical case of the replacement of an aging bus fleet, the agency capital cost would include the costs of • Bus replacement and • Any changes to existing facilities required to accommodate the new fleet. If the analysis period extends beyond the expected life of the fleet, then the cost of subsequent replacement of the fleet should be included as well. Alternatively, if the analysis period is equivalent to the expected life of the fleet, then subsequent costs would be excluded. Regarding the base case, if this case is defined to exclude any capital investments, then there are no capital costs to quantify. However, if the base case represents a deferral of needed work, then it may be necessary to include the investments being deferred, with these occurring at a

ROI Calculation Guidance 23 later time. Note that the cost of the investment is calculated as the difference between the capital costs of the investment and base cases. Thus, it is not necessary to quantify any capital costs that have the same value and occur at the same time in both cases, as the net cost of the investment is zero in such instances. Nonetheless, it is important to quantify costs that occur at different times in the two cases. Once the costs are quantified, the analyst can enter these into the Return on Investment Calculator to assist with the analysis. In quantifying future costs, it is important for the analyst to be familiar with the concepts of inflation and discounting. Inflation is the general increase that tends to occur over time in the prices of goods and services in the economy (4). That is, in general, the cost of a set of goods or services tends to be higher from one year to the next. For instance, when the New York City subway opened in 1904, the fare was $0.05 per ride. By 1986, the fare had increased to $1 per ride, and as of 2018, the cost was $2.75 (5). One can demonstrate similar increases in nearly any other good or service. The absolute price of a good or service, unadjusted for inflation, is its nominal or current dollar price. By contrast, the inflation-adjusted price is called the real or constant dollar price. A basic technique to account for inflation is to inflate historic costs to prices of a specific year using an aggregate measure of inflation such as the Consumer Price Index (CPI) or a construction cost index, and then express projections of future costs using that year’s prices. The U.S. Office of Management and Budget (OMB) recommends the use of such real values rather than nominal values for economic analyses. Specifically, OMB Circular A-94: Guidelines and Discount Rates for Benefit–Cost Analysis of Federal Programs states, “Future inflation is highly uncertain. Analysts should avoid having to make an assumption about the general rate of inflation whenever possible” (6). Thus, for an ROI analysis it is recommended that the ana- lyst document the year for which prices in the analysis are representative (e.g., “2018 dollars”) and conduct the analysis using real values. However, this approach implicitly assumes that the relative costs of different commodities are stable (that is, they increase at the same rate). An analyst may want to introduce consideration of inflation if he or she has data to suggest that the cost of a given commodity (e.g., fuel) is likely to change at a different rate from that of other commodities. If one is working with nominal values, it is possible to restate these as real values provided one knows what assumptions were made concerning inflation to calculate the nominal values. Specifically: R N I I R N (3-1)= where R is the real value, N is the nominal value, IR is the price index for the year for which prices are computed for real cost, and IN is the price index for the year in which the nominal cost is projected. Another important concept in performing an ROI analysis is that of discounting— or adjusting—future benefits and costs to account for the time value of money and risk premium. OMB describes the logic for discounting as follows: “Benefits and costs are worth more if they are experienced sooner. All future benefits and costs, including non-monetized benefits and costs, should be discounted.” Note that if one is analyzing an investment that will yield benefits over time, generally the lower the discount rate is, the more favorable the investment will appear. Thus, care is needed in setting the discount rate, and once a rate is set it should be used consistently across dif- ferent analyses. OMB Circular A-94 provides additional guidance on how to set the discount

24 Guidance for Calculating the Return on Investment in Transit State of Good Repair rate, recommending a rate of 7 percent for investments that “displace both private investment and consumption” and the U.S. Treasury’s borrowing rates for other types of analysis, such as cost-effectiveness analysis or evaluation of internal investments. This resource is relevant, as it establishes the U.S. government’s policies for BCA of federal programs, but individual transit agencies are not required to follow this guidance. For ROI analyses, it is recommended that a transit agency use the rate established by the transit agency for other analyses if available. A default of 4 percent is recommended (between the two values described by OMB Circular A-94) where no value is available. If the analyst is using the Return on Investment Calculator, he or she needs to enter costs in real dollars, and specify the discount rate to use, if different from the default. The tool discounts all costs as follows: PV C r n1 (3-2) ( ) = + where PV is the present value of a cost or benefit, C is the real cost or benefit incurred in year n, and r is the discount rate expressed as a percentage. If one is calculating the present value of a stream of costs, the above formula is applied to the costs incurred each year of the analysis and the results are summed. The net cost of an invest- ment is then calculated as the difference between the present value of costs in the investment and base cases. Procedures 1. Determine the capital costs incurred by year for the investment and base cases. In most cases, this will entail analysis by the transit agency of the proposed investment. However, if either the investment or base case involves making all investment needed to achieve and maintain SGR, the analyst may be able to use a tool such as TERM Lite or TAPT to predict future capital costs. 2. If the capital costs are nominal rather than real costs, determine what assumptions regarding future inflation were made to compute the nominal costs and adjust the nominal costs to obtain real costs as outlined above. 3. Enter the real costs for each case in the Return on Investment Calculator, as depicted in Figure 3-3. 4. If there are other capital costs associated with the investment and/or base cases, these may be entered as well. Enter the real costs for each case in the Return on Investment Calculator, as depicted in Figure 3-4. 5. Establish the discount rate to use for the analysis. If the transit agency has established a spe- cific rate for analysis or used a given rate for a series of other analyses to be compared with the current analysis, enter that rate, as depicted in Figure 3-5. 6. Document the assumptions and adjustments made concerning inflation and discounting, as well as the cost year used. Additional Resources TCRP Report 78 is a valuable resource for guidance on how concepts such as inflation and discounting are addressed in BCA (3). OMB Circular A-94 provides the U.S. federal government’s guidance on performing eco- nomic analyses and describes different approaches to setting discount rates (6).

Figure 3-3. Enter costs for base case and investment case.

Figure 3-4. Enter other costs.

ROI Calculation Guidance 27 Figure 3-5. Enter discount rate. Step 3. Calculate Other Agency and Social Costs Overview To determine the return of an SGR investment, it is necessary to consider the impact of the investment on other agency and social costs besides capital costs. These might include the following: • Costs of routine maintenance, • Costs of reactive maintenance to replace failed components, • Changes in energy costs, • Changes in other operation costs, and • Social costs associated with fuel consumption. The basic rationale for including such costs in the analysis is that in the absence of needed SGR investments such costs are likely to increase. Thus, including these costs in the analysis provides a more accurate depiction of the effects of SGR investment (or lack thereof) and tends to strengthen the case for achieving and maintaining SGR. The difference in non-capital agency and social costs between the investment and base cases is included in the calculation of the Net Present Value (NPV) of the investment case. The NPV of the investment is the sum of the discounted benefits of the investment less its discounted costs. A reduction in other agency and social costs in the investment case relative to the base case is treated as a benefit, while an increase is treated as a disbenefit. As in the case of capital costs described above, other costs should be specified as real costs rather than nominal costs, and they are discounted prior to being summed over time. Other agency and social costs are handled differently from capital costs in the analysis in that they are not treated as costs in calculating certain ROI measures, including BCR, internal rate of return, and payback period. Only the capital costs determined in Step 2 are treated as a cost for these calculations. Increases in other types of costs are treated as disbenefits. Another value the analyst must determine at this step is the change in the residual value of the transit agency’s assets (also termed “salvage value”). This is approximated based on the change in investment needs as of the end of the analysis period between the base and

28 Guidance for Calculating the Return on Investment in Transit State of Good Repair investment case. This parameter is used to address instances where two cases result in very different estimates in asset condition as of the end of the analysis. However, this parameter is not needed if the base and investment cases include costs required to bring the asset inventory to the same state (e.g., to SGR) as of the end of the analysis period. Regarding social costs, the Return on Investment Calculator considers—as a social cost—the environmental cost of fuel consumption and resulting emissions. Changes in agency fuel consumption are determined by specifying the gallons of fuel consumed per year for the investment and base cases, multiplied by the environmental cost per gallon of fuel. Note that changes in automobile fuel consumption are calculated separately and described in Step 5. It is important to underscore that all costs and benefits of an investment are calculated by com- paring the investment and base cases. Thus, it is not important, or even particularly helpful, to quantify costs or fuel consumption values that are the same in both cases—they cancel out. The analyst should concentrate his or her efforts in quantifying those values that differ between the two cases. Furthermore, in calculating operating and maintenance (O&M) costs, the need is to calculate those costs specifically associated with a transit agency’s assets. In many cases, these may only be maintenance costs, but in some cases there may be an operating cost component associated with asset maintenance. For instance, the failure of a key track component may require establishing a temporary bus shuttle, which may be tracked as an operating cost rather than as a maintenance cost. The analyst has two basic options for estimating the various agency costs described here: estimate them through other means and enter them directly, or calculate them based on a set of baseline values and the predicted changes in asset ages over time. Two publicly accessible tools are available for estimating the required costs: FTA’s TERM Lite (7) and TAPT, which is detailed in TCRP Report 172 (8). Both of these tools estimate investment needs for achieving and maintaining SGR and can be used to estimate asset maintenance costs. TAPT includes changes in energy or fuel costs in its calculations as well. The following procedures describe how to obtain the needed values from these tools if a transit agency is using them for its analysis. Various other commercial off-the-shelf and custom systems have been developed or adapted for supporting analysis of SGR needs, and these can be used as an alternative. Whichever tool is used, the end result with this option is that the analyst obtains calculations of investment need and values for the investment and base cases by year for • Agency O&M costs, • Energy costs, • Gallons of fuel consumed, and • Other social costs, if applicable. If the analyst does not have data from an existing investment analysis tool, then they can estimate the required agency costs using the Return on Investment Calculator. In this case, the analyst enters a set of baseline values then estimates the average age by year for four asset classes and two modes. With this approach, baseline values are entered for up to two modes for the following asset classes for both O&M costs and average asset age: • Vehicles • Guideway • Facilities • Other Assets Next, the analyst specifies the average age by year for each of the above asset classes for up to two modes. The Calculator includes a set of prototypical cost curves for each asset class, with vehicles further split between buses, light rail, heavy rail, and commuter rail. To predict the cost as a function of average age, the tool uses its cost profiles to calculate the ratio between the age in the specified year and the baseline age and then applies this ratio to the baseline value. For

ROI Calculation Guidance 29 instance, if the average age of the bus fleet is predicted to increase from 6 to 12 years, the tool calculates a ratio of 1.24 for vehicle O&M costs and multiplies the baseline vehicle O&M cost by this value to calculate future costs. Use of asset ages to predict future costs is recommended only in cases where the analyst has no other estimates available, given that age is an imperfect proxy for asset state of repair. Even when asset ages are used, the analyst will still need to estimate the future change in investment need. In addition, if other social costs are being quantified in the analysis besides environmental costs of fuel consumption, these must be explicitly quantified. Procedures 1. Determine whether you will enter O&M costs by year or will predict future costs based on average asset ages. Enter the value for this parameter as shown in Figure 3-6. If you are predicting costs based on asset age, skip to Step 3. 2. If you are entering O&M costs by year, determine the O&M cost, energy cost, and gallons of fuel consumed for the set of assets established in Step 1. Note that in TERM Lite, O&M expenditures are determined by subtracting capital expenditures from the total expenditures shown in the TERM Lite expenditures report. TERM Lite does not provide explicit predictions of energy costs or gallons of fuel consumed. In TAPT, O&M costs are reported in the TAPT Summary as “Other Agency Costs,” and energy costs are explicitly reported. The tool calculates fuel consumption, but this is not shown in the summary report. If energy costs include only buses, one can estimate the gallons of fuel by dividing the energy cost by the cost per gallon of fuel. Once these values have been obtained, enter them in the Return on Investment Calculator, as shown in Figure 3-7, and skip to Step 5 (as in this case it is not necessary to provide additional information on predicted asset ages). 3. Select the type of mode for up to two modes. If using one mode, leave Mode 2 blank. Then enter the following initial values for each mode, as shown in Figure 3-8, leaving blanks if not applicable: energy cost; energy consumption in gallons; O&M costs for vehicles, guideway, facilities, and other assets; and average age for vehicles, guideway, facilities, and other assets. 4. Enter the average age by year for each asset class and mode, as shown in Figure 3-9. In this case, the average age is used to approximate annual O&M costs. In general, the age increments by one each year while replacement of all of the assets in the class resets the age to one year (assuming work is performed at the beginning of the prior period). If investments are being made to maintain SGR for a large set of assets (e.g., for facilities) and this is reflected in the capital costs, then the age may be assumed to be the same from one period to the next. Figure 3-6. Specification of approach for predicting year-by-year costs.

Figure 3-7. Enter costs.

ROI Calculation Guidance 31 5. Calculate the initial investment need for achieving SGR and the projected need as of the end of the analysis period for the investment and base cases. This value is predicted by tools such as TERM Lite and TAPT. As discussed above, here the change in need is used as a proxy for change in asset value as of the end of the analysis period. The need should include the set of assets established in Step 1 and expressed as a real cost. Enter these values, as shown in Figure 3-10. Note: one can enter zeros for the above values if predictions of investment needs are unavailable, or if the salvage or residual value of the asset inventory should be omitted from the analysis. 6. Review the environmental cost per gallon of fuel for agency fuel consumption and revise if different from the default value, as shown in Figure 3-11. Note the environmental cost for agency fuel consumption may differ depending on the type of fuel used by the transit agency (e.g., diesel versus gasoline) and other assumptions. 7. If other social costs besides environmental costs associated with fuel consumption are being included in the analysis, calculate these by year for the investment and base case and enter real costs in the Return on Investment Calculator, as shown in Figure 3-12. Additional Resources TERM Lite is an SGR analysis tool available from FTA (7). TAPT is an alternative tool detailed in TCRP Report 172 (8). Step 4. Calculate Effective Journey Time Figure 3-8. Enter values for Mode 1 and Mode 2. Overview SGR impacts can have two basic types of effects on transit users: they may affect users’ level of service, and/or they may affect ridership. In the ROI framework, level of service impacts

Figure 3-9. Enter average age.

ROI Calculation Guidance 33 Figure 3-10. Enter other parameters. Figure 3-11. Enter environmental cost. are captured using the Effective Journey Time (EJT) measure. This measure captures the total time a transit user must budget to make a trip, with adjustments for the quality of that time. It includes wait time, in-vehicle time, and the buffer time the transit user must budget to ensure that they arrive at their destination on time. In cases where significant time is required to negotiate transit stations, transit users’ in-station conveyance time may be included as well. While no single measure can fully describe the many dimensions of transit quality of service, EJT represents an attempt to characterize many of the dimensions that may vary as a function of asset condition. An important part of the process of calculating EJT is making adjustments for transit users’ perceptions of their journey. For instance, by default, actual wait time spent standing is multiplied by a factor of 1.9 to account for the increased discomfort compared with sitting. Like- wise, the in-vehicle time on a vehicle that has reached or exceeded its useful life is multiplied by a factor of 1.2 to account for the increased discomfort of riding on a deteriorated vehicle. Other adjustments are made to buffer time, time spent walking up and down stairs, extra time spent as a result of vehicle failures, etc. However, the analyst has discretion in how EJT is calculated and can calculate an unadjusted time without corrections for transit user perceptions, if desired. Procedures for calculating EJT and a set of two EJT calculation tools are provided in TCRP Research Report 198: The Relationship Between Transit Asset Condition and Service Quality (9). This report provides two different tools for calculating EJT: 1) a comprehensive tool that calculates EJT for each passenger on a transit line and considers data on passengers boarding and alighting at each stop or station, and 2) a simplified tool that yields a single EJT calculation for a line, making a series of simplifying assumptions. In theory, the comprehensive tool provides a more accurate calculation, but it can be used only if one has sufficient data— including information on ridership by stop or station—and specific assumptions regarding changes over time in vehicle headways and run times. If the analyst lacks such data, then the alternative is to use the simplified tool. Figure 3-13 is a screenshot showing the data entry fields in the Simplified EJT Calculator. Whether the analyst uses one of the tools from TCRP Research Report 198 or calculates EJT through some other means, the result is a prediction of existing EJT and EJT at the end of the analysis period for the investment and base cases. These values are used to determine the impacts of SGR investment on transit users. In the case that a travel demand model is unavailable for the analysis, changes in EJT are used to predict future changes in transit ridership in Step 5.

Figure 3-12. Enter other social costs.

ROI Calculation Guidance 35 Procedures 1. Review the data requirements in Chapter 5 of TCRP Research Report 198 to determine whether to use the simplified or comprehensive version of the EJT Calculator. 2. Follow the procedures in TCRP Research Report 198 to calculate existing EJT and future EJT per passenger for the investment and base cases. Note that the EJT models apply to a single line or route. If the analysis includes multiple lines, routes, or modes, it may be necessary to make multiple EJT calculations and then to combine them to obtain overall EJT. 3. Enter the resulting EJT values in the Return on Investment Calculator in the fields shown in Figure 3-14. Additional Resources TCRP Research Report 198 details the EJT calculation tools described here and includes a review of relevant research on transit SGR and quality of service (9). TCRP Report 165: Transit Capacity and Quality of Service Manual, Third Edition details the dimensions of transit service quality (10). Figure 3-13. Simplified EJT Calculator (9). Figure 3-14. Enter EJT results.

36 Guidance for Calculating the Return on Investment in Transit State of Good Repair Step 5. Calculate Change in Travel Demand and Its Effects Overview Ultimately, if the impact of SGR investment (or disinvestment) is great enough, it may have an impact on transit ridership. The analyst has two basic alternatives for calculating ridership impacts: determine these separately using a travel demand model, or estimate ridership changes as a function of change in EJT using an arc elasticity. Estimating changes in travel demand using a travel demand model is preferred if the following criteria are met: • The investment being analyzed is system-wide and expected to have impacts over a period of 10 years or more (e.g., analyzing overall impacts of investing in SGR as opposed to a one-time investment in a selected set of assets); • The investment may result in shifts between single occupancy vehicles and/or between different transit modes (e.g., between bus and rail); and • The transit agency has access to a regional travel demand model that will support the analysis. In the event the previous criteria are met, the analyst should work with staff familiar with the region’s travel demand model, providing data on expected changes in EJT as appropriate. The following should be obtained from the model: • Annual transit trips modeled in the initial period, by mode; • Future annual transit trips by mode at the end of the analysis period for the investment and base cases; • Annual auto vehicle miles traveled (VMT) and vehicle hours traveled (VHT) in the initial period; and • Future annual VMT and VHT at the end of the analysis period for the investment and base cases. If use of a travel demand model is deemed infeasible or unnecessary, then the Return on Investment Calculator will estimate the change in travel demand by mode based on the elasticity of travel demand with respect to travel time. Thus, in this case, the change in EJT drives the change in travel demand. To the extent that travel demand changes, an additional parameter specifies the percentage of trips that result in a shift to or from automobiles. By default, it is assumed that 56 percent of new transit trips result in a shift from automobiles, and likewise 56 percent of lost trips result in additional automobile trips. Given predictions of EJT per passenger and transit trip, the Return on Investment Calculator calculates the following to determine user costs and benefits and associated social costs: • Transit travel time costs based on EJT, transit demand, and the transit user’s value of time. • Automobile travel time costs either taken directly from the travel demand model results or based on the predicted mode shift between transit and autos if an elasticity is used to predict

ROI Calculation Guidance 37 changes in demand. Additional parameters specify the value of time and average duration of an automobile trip that substitutes for a transit trip and average vehicle occupancy. For instance, if an investment results in 1,000 more transit trips per day, by default it is assumed that 560 of these trips will substitute for automobile trips. If we further assume that each automobile trip has an average duration of 22 minutes and there are 1.1 occupants per vehicle, then the savings in automobile travel time will be approximately 187 vehicle hours per day (560 passenger trips / 1.1 passengers per vehicle × 22 minutes per trip / 60 minutes per hour). • Automobile operating costs taken are determined by multiplying the automobile operating cost per mile by the change in VMT either taken directly from the travel demand model results or based on the above change in trips if an elasticity is used. An additional parameter specifies the average length of an automobile trip. In the above example, the 560 fewer auto trips per day results in a reduction of approximately 2,545 VMT per day (560 passenger trips / 1.1 passenger per vehicle × 5 miles per trip). • Reduced consumer surplus. This calculation is needed to account for new transit users who shift from other modes besides automobiles and transit users who depart the system but do not shift to automobiles. We know that these users’ costs are changing enough to trigger a mode shift, but we do not really know by how much. Thus, the change in consumer surplus is estimated as being half of the change in travel time for each of these users, consistent with standard BCA convention. • Auto emissions costs calculated based on VMT and the emissions cost per VMT. • Congestion costs calculated based on VMT and the congestion cost per VMT. This mea- sures the additional congestion caused by each automobile added to a network. The analyst can specify whether these costs are included in the analysis. Typically these costs should be excluded if a travel demand model is used—a travel demand model would typically capture congestion effects—and included if an elasticity is used. All of the above costs are classified as user costs, with the exception of emissions and conges- tion costs. Emissions costs and congestion costs (where used) are classified as social costs. Procedures 1. Enter values for the following parameters used for calculating user and social costs, or use the model defaults for these: – Annual transit trips for up to two modes – Personal value of time – Elasticity of transit demand with respect to travel time – Percentage of the change in transit trips to/from autos – Auto operating cost – Environmental cost of autos – Additional auto congestion cost – Average auto occupancy – Average auto trip length – Average auto trip duration These are entered as shown in Figure 3-15. 2. Determine whether you will use a travel demand model or elasticity to estimate travel demand. Also determine whether you will calculate additional congestion costs resulting from changes in automobile travel (typically the answer is “no” if using a travel demand model and “yes” otherwise). Enter the value for these parameters as shown in Figure 3-16. If you are using elasticity, then skip Step 3. 3. Work with your transit agency’s travel demand modeling staff to obtain the following from the travel demand model: – Annual transit trips in the initial period for each mode, as well as at the end of the analysis period for the investment and base cases.

38 Guidance for Calculating the Return on Investment in Transit State of Good Repair Figure 3-15. Enter required, general, and automobile- related parameters. Figure 3-16. Travel demand model dropdown.

ROI Calculation Guidance 39 – Annual auto VMT in the initial period, as well as at the end of the analysis period for the investment and base cases. – Annual auto VHT in the initial period, as well as at the end of the analysis period for the investment and base cases. These are entered in the fields shown in Figure 3-17. Additional Resources The previously described calculation of user and social costs in BCA are referenced in TCRP Report 78 (3). Litman reviews typical values used for various parameters when analyzing costs and benefits of transit investments (11). NCHRP Report 716 provides an overview of travel demand forecasting techniques, as well as their data requirements and assumptions (12). Step 6. Evaluate Results Overview At this point, the analyst has specified all the data required for an ROI analysis and can now review the results. In the Return on Investment Calculator, year-by-year calculations are detailed on the Calculations tab, and the results are summarized on the Results tab. The tool provides the following summary results: • Investment benefits. The sum of the benefits for each year of the analysis. The benefit is the savings in non-capital agency costs, user costs, and social costs for the investment case com- pared with the base case. These are presented as undiscounted and discounted sums. Positive values indicate a benefit is obtained for the investment case, and negative values indicate a disbenefit. The grand total is shown, as well as the totals for agency, user, and social benefits. Each of these categories is further subdivided as follows: – Agency benefits include savings in agency energy costs, savings in O&M costs, and increased residual value. – User benefits include reduced transit travel time, reduced automobile travel time, reduced vehicle operating costs, and increased consumer surplus. Figure 3-17. Enter travel demand model results.

40 Guidance for Calculating the Return on Investment in Transit State of Good Repair – Social benefits include reduced emissions costs for transit and automobiles, reduced congestion costs, and reduced other social costs. • Costs. The sum of the capital costs for each year of the analysis. The cost shown is the increased capital cost of the investment case relative to the base case. These are presented as undiscounted and discounted sums and further subdivided by agency and other capital costs corresponding to the categories of costs entered in Step 2. A positive value indicates a greater cost for the investment case than the base case. • Net Present Value (NPV). The sum of discounted benefits, less the sum of discounted costs. An investment with positive NPV is generally deemed worthwhile. • Benefit/Cost Ratio (BCR). The ratio of discounted benefits to discounted costs. Generally, an investment with a BCR greater than one is deemed worthwhile. • Internal Rate of Return (IRR). This is the value for the discount rate for which the benefits of the investment equal its costs. If this value is greater than the discount rate used for the analysis, the investment will have a positive NPV and thus be deemed worthwhile. In industry, this value is often compared with an organization’s interest rate on loans to determine whether it is worthwhile to borrow money to make an investment. This measure is computed using the IRR function in Microsoft Excel. • Payback Period. This is the number of years required for the sum of undiscounted benefits to equal or exceed the sum of undiscounted costs. That is, it represents the time required to recoup the investment. This measure can be useful in interpreting the results of an analysis and for comparing two investments: all things being equal, the investment with the shorter payback period is preferred. However, it can be difficult to interpret the measure when the investment involves a stream of varying costs over time. The Return on Investment Calculator also lists some of the key analysis assumptions and shows graphs summarizing total costs and benefits. In interpreting the results, the analyst may conclude that an investment is worthwhile if it has a positive NPV. In such cases, the BCR is always greater than one, and the IRR should be greater than the discount rate. If one is compar- ing two different investments, then, strictly speaking, the investment with the greatest NPV is preferable given it has greater value. However, if a transit agency has a limited capital budget and is seeking to make the investment with the greatest benefit per dollar of investment, then selecting the investment with the greater BCR achieves this—provided the two investments are analyzed with a consistent base case. In interpreting BCR, one should be mindful that seemingly small changes in how costs are classified could have a significant impact on this measure. For instance, if an investment of $2 million in capital funds yields agency O&M cost savings of $1 million and user cost savings of $3 million, then the NPV of the investment is $2 million and the BCR is 2. If, however, the change in agency cost is classified as a capital cost, then the investment is determined to cost $1 million, yield no savings in agency costs, and still yield $3 million in user cost savings. In this case, its NPV is still $2 million as before, but its BCR is now 3. This is a significant change! To avoid such issues, the analyst can either rely on NPV rather than BCR, or, if using BCR, be careful and consistent in classifying costs. Although the ROI analysis yields specific values for the measures described above, in reality there is inherent uncertainty in both the individual analysis parameters and any projections of future conditions. A basic approach to addressing this uncertainty is to perform a sensitivity analysis to determine the degree to which the results of the analysis change as a result of changes in model assumptions and parameters. To perform such an analysis, the analyst must deter- mine which parameters to vary and how much to vary them and then calculate a set of results using revised values for the selected parameters. For instance, one might quantify the effects of changes to the discount rate, value of time, and agency O&M costs, in each case testing halving and doubling of each parameter.

ROI Calculation Guidance 41 To address uncertainty in a more comprehensive manner, the analyst can perform a Monte Carlo simulation to obtain a distribution of results for a large set of simulated cases. Such an analysis can be structured to consider cross-correlations between different modeling parameters in determining how to vary parameters for each case. For instance, if automobile operating costs are higher than projected, then it is likely that transit agency energy costs are higher as well. Procedures 1. Review the initial analysis results in the Results tab of the Return on Investment Calculator. The key results generated by the tool are depicted in Figure 3-18. Figure 3-18. Summary results.

42 Guidance for Calculating the Return on Investment in Transit State of Good Repair 2. As necessary, review the analysis assumptions and calculations to gain insights into the results. Year-by-year calculations are detailed on the Calculations tab of the Return on Investment Calculator. As needed, revise the analysis assumptions and review the revised results. 3. Perform a sensitivity analysis in which key model parameters are revised, yielding revised results. The analyst should consider which model parameters are most uncertain and vary these over the range of likely values. If desired, perform a Monte Carlo simulation in which results are determined for a set of test cases, with key parameters varied by random in each case. 4. Document the analysis results and the results of the sensitivity analysis. Additional Resources TCRP Report 35 discusses how to interpret results of an economic analysis, as well as the strengths and weaknesses of different types of analyses (13). The final report for NCHRP Project 8-36, Task 62, “Best Practice Methodology for Calculating Return on Investment for Transportation Programs and Projects,” discusses the use of sensitivity analysis and other approaches for addressing uncertainty in ROI projections (14). Step 7. Summarize and Supplement Results Overview The financial measures described in Step 6 are the primary output of the ROI analysis. However, no one measure can fully describe the effects of investing in achieving and main- taining SGR for transit assets. In this step, the analyst should prepare a supplemental narrative description of the investment to address any factors not appropriately captured in the analysis and document the analysis results. If desired, the analyst may supplement the ROI analysis with additional supporting analyses. The supporting narrative should describe the proposed investment, the rationale for what needs to be spent and why, and any additional factors that lend extra credence to making the investment, such as the need to replace a set of assets likely to fail in the near term. Factors that the analyst may wish to introduce at this point that are not well addressed in traditional BCA include but are not limited to • Criticality of the investment, • Community impacts, • Effects on land values, • Legal requirements or other commitments, and • Impacts on the transit agency’s reputation. In this step, an analyst may additionally prepare an economic impact analysis of the proposed SGR investment. As discussed in Chapter 2, an economic impact analysis relies on a somewhat

ROI Calculation Guidance 43 different set of assumptions to calculate a different set of measures from BCA. One cannot add the results from an economic impact analysis to the results from BCA. However, an eco- nomic impact analysis may provide supplemental information of value for decision making. The approach and specific values from the final report from TCRP Project J-11, Task 7, “Economic Impact of Public Transit Investment” (15), are recommended in the procedures for estimating the following economic impacts: • Cost reduction impacts. These are generated by cost savings and improved productivity projected to result from increased investment in transit. – Personal cost savings – Business cost savings – Equivalent wages from cost savings – Equivalent jobs from cost savings – Tax revenues from cost savings • Spending impacts. These are multiplier effects resulting from spending money, either on capital investments or O&M costs. – Total economic impact – Equivalent wages from increased spending – Equivalent jobs from increased spending – Tax revenues from increased spending Procedures 1. Document the analysis assumptions and results and save a copy of the Return on Investment Calculator with the analysis results. 2. As needed, prepare a narrative description of the investment documenting additional economic and social impacts not captured in the ROI analysis. 3. If desired, prepare a supplemental economic analysis. Table 3-1 provides estimates of cost reduction impacts (in 2012 dollars) as a function of new transit trips and spending impacts as function of increased spending. Use the increase in transit trips and undiscounted agency costs totals with the parameters in this table to estimate economic impacts. Additional Resources TCRP Reports 35 and 78 describe additional factors not included in an analysis performed using a BCA framework (13, 3). The report for TCRP Project J-11, Task 7, “Economic Impact Type of Impact Effect Approximate Value ($ 2012) Cost Reduction Personal Cost Savings $4.63 per new trip Business Cost Savings $2.56 per new trip Wages $5.50 per new trip Job Equivalent 1 job per 9,649 new trips Corresponding Tax Revenue $1.11 per new trip Spending Impact Total Economic Impact $3 per $1 spent Wages $1.30 per $1 spent Job Equivalent 21.8 jobs per $1 million spent Corresponding Tax Revenue $432,000 per $1 million spent Table 3-1. Estimates of cost reduction and spending impacts (15, 16).

44 Guidance for Calculating the Return on Investment in Transit State of Good Repair of Public Transit Investment,” details an economic impact analysis for transit investments (15). The parameters published in the report were subsequently updated in a 2014 report prepared for APTA (16). References 1. Paterson, L., and D. Vautin. “Evaluating the Regional Benefit/Cost Ratio for Transit State of Good Repair Investments.” Journal of Public Transportation, 18(3), 2015. 2. Parsons Brinckerhoff. Asset Management Guide. FTA Report 0027. Federal Transit Administration, 2012. 3. ECONorthwest, and Parsons Brinckerhoff Quade & Douglas, Inc. TCRP Report 78: Estimating the Ben- efits and Costs of Public Transit Projects: A Guidebook for Practitioners. TRB, National Research Council, Washington, D.C., 2002. 4. Federal Highway Administration (FHWA). Economic Analysis Primer. Washington, D.C., 2003. 5. Pham, D. “All the MTA Fare Hikes of the Last 100 Years.” 2015. https://www.6sqft.com/all-the-mta-fare- hikes-over-the-last-100-years-plus-a-video-of-when-it-cost-just-15-cents/. 6. Circular A-94: Guidelines and Discount Rates for Benefit–Cost Analysis of Federal Programs. Office of Manage- ment and Budget. 1992. 7. TERM Lite User’s Manual. Federal Transit Administration. 2015. 8. Robert, W., V. Reeder, K. Lawrence, H. Cohen, and K. O’Neil. TCRP Report 172: Guidance for Developing a Transit Asset Management Plan. Transportation Research Board of the National Academies, Washington, D.C., 2014. 9. Spy Pond Partners, LLC; AECOM; McCollom Management Consulting, Inc.; H. Cohen, and S. Silkunas. TCRP Research Report 198: Relationship Between Transit Asset Condition and Service Quality. Transportation Research Board, Washington, D.C., 2018. 10. Kittelson & Associates, Inc.; Parsons Brinckerhoff; KFH Group, Inc.; Texas A&M Transportation Institute; and ARUP. TCRP Report 165: Transit Capacity and Quality of Service Manual,Third Edition. Transportation Research Board of the National Academies, Washington, D.C., 2013. 11. Litman, T. Evaluating Public Transit Benefits and Costs: Best Practices Guidebook. Victoria Transport Policy Institute, 2017. 12. Cambridge Systematics, Inc.; Vanasse Hangen Brustlin, Inc.; Gallop Corporation; C. Bhat; Shapiro Trans- portation Consulting, LLC; and Martin/Alexiou/Bryson, PLLC. NCHRP Report 716: Travel Demand Fore- casting: Parameters and Techniques. Transportation Research Board of the National Academies, Washington, D.C., 2012. 13. Cambridge Systematics, Inc.; R. Cervero; and D. Aschauer, D. TCRP Report 35: Economic Impact Analysis of Transit Investments: Guidebook for Practitioners. TRB, National Research Council, Washington, D.C., 1998. 14. Cambridge Systematics, Inc.; and Economic Development Research Group. “Best Practice Methodology for Calculating Return on Investment for Transportation Programs and Projects.” Contractor’s Final Report for NCHRP Project 8-36, Task 62. 2008. 15. Weisbrod, W. and A. Reno. “Economic Impact of Public Transit Investment.” Contractor’s Report prepared for APTA through TCRP Project J-11, Task 7. 2009. 16. APTA. Economic Impact of Public Transportation: 2014 Update. 2014.