A Guide to Developing Financial Plans and Performance Measures for Transportation Asset Management (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.

41 Forecasting revenue and expenditures helps agencies plan ahead and anticipate their ability to accomplish their long term goals. With the sources and uses identified in the previous chapter, the next step is to project the revenues and expenditures into the future in order to build a sustainable financial plan. There are a number of different methods that can be used to forecast revenues and expenses. This chapter provides background on those methods commonly used by transportation organizations and outlines steps for TAM practitioners to follow when forecasting for the TAM financial plan. 4.1 Fundamentals of Forecasts Introduction In order to develop a comprehensive financial plan for asset management, agencies must identify actual and estimated revenue sources 10 years into the future. Fortunately, state DOTs are already in the practice of forecasting revenue sources and expenditures. In particular, state DOTs incorporate financial projections into their STIP and their long-range plan. The 10-year look into the future required by the FHWA rule on TAMPs goes beyond the STIP time horizon, which is four to six years, but is within the duration of the state’s long-range plan. TAM profes- sionals therefore do not have to develop forecasts on their own, but still need to understand the methodologies and assumptions used in forecasting. This chapter offers background informa- tion and key fundamentals of forecasting revenues and expenditures, but the TAM professional developing the financial plan for asset management is encouraged to utilize the efforts already in place within the agency to project revenue in the future. At this point, the TAM professional should be equipped with a list of asset management-related sources and uses, populated with the current-year budget values. The steps at the end of this chapter will guide the practitioner through the process of forecasting to obtain a 10-year forecast of agency revenues and non-asset management uses. The forecast of asset management uses will then be covered in the next chapter. The final product will vary by agency, but Table 4-1 shows an example of a forecasting worksheet based on data from the Texas Department of Transportation (Texas Department of Transportation 2017). The Texas DOT projects 11 years into the future, and Table 4-1 displays three of those 11 years. Forecasting Sources A variety of methods have been developed previously for forecasting revenue. NCHRP Synthesis 479: Forecasting Transportation Revenue Sources: Survey of State Practices provides C H A P T E R 4 Financial Forecasting

42 A Guide to Developing Financial Plans and Performance Measures for Transportation Asset Management key insight into the practices states use to predict future revenue. In general, the forecasting activities at the state level focus on “traditional revenue sources: disbursements of federal funds as reimbursements for expenditures, state fuel and excise taxes, state vehicle registration fees, tolls, and local taxes and fees” (Wachs and Heimsath 2015). The report describes three of the most commonly used methods of predicting future revenue from these traditional sources: • Trend Extrapolation, • Expert Analysis, and • Econometric Models. The following summarizes each of these methods from the NCHRP report. Trend Extrapolation Trend extrapolation “involves applying a mathematical formula to past revenue levels to determine future revenues” (Wachs and Heimsath 2015). Thus, historical funding trends can be extrapolated into the future in a “linear manner or with some adjustment to account for expected future changes in underlying conditions” (Wachs and Heimsath 2015). For example, historical data might indicate a particular growth rate on a sales tax used for transportation. Assuming this same growth rate into the future allows agencies to predict future revenue in a straightforward manner. Trend extrapolation is a straightforward application of basic statistical methods. In its sim- plest form, it may be just a univariate linear regression of the dependent variable on time, though multivariate and/or nonlinear models can be employed as well. According to the NCHRP Syn- thesis 479, states generally use trend extrapolation to predict revenue from funding sources that are “historically stable,” that is, the method of trend extrapolation is a reliable way to predict Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 Actual Actual Actual Projected Projected Projected Expenses Plan 1,420 1,496 1,731 2,037 1,837 1,539 Build 3,132 2,961 2,971 3,154 2,584 1,755 Maintain 3,586 4,035 4,002 4,885 5,166 4,427 Use 166 157 166 189 192 182 Manage 220 226 240 306 284 263 Project Development & Delivery 9 693 1,131 1,328 1,820 2,331 Short-Term Borrowing D/S 376 752 1 1 1 1 Inflation/Additional Programs - - - - - 165 Total Expenses 8,910 10,320 10,243 11,900 11,884 10,664 Funding Sources State Highway Fund 8,100 9,472 9,401 10,340 10,782 9,793 State Mobility Fund - Bond Proceeds 220 158 169 347 300 179 State Mobility Fund - Taxes & Fees 229 379 409 756 777 564 ARRA Funds 15 10 - - - - Build America Bond Subsidy 90 40 96 98 88 123 GR – Debt Service 250 214 134 309 -100 - General Revenue 3 5 4 2 2 2 State Infrastructure Bank (SIB) 3 42 30 48 35 3 Total Funds 8,910 10,320 10,243 11,900 11,884 10,664 Actual Projected Table 4-1. Example high-level overview of a forecasting worksheet ($ million).

Financial Forecasting 43 revenue from sources like general fund allocations and local contributions that are determined by a “fixed formula” (Wachs and Heimsath 2015). Expert Consensus Another method states employ to project revenue is that of expert consensus. This “relies on the professional judgement of a selected panel or conference of economists, analysts, academics, and others who discuss and try to agree upon future revenue projections or on critical inputs that will affect those projections” (Wachs and Heimsath 2015). NCHRP Synthesis 479 notes that this method can be used in combination with other methods as expert panels are asked to review forecasts already produced and to make necessary adjustments. For expert consensus, the agency can have a panel of experts meet together to discuss the forecast, or the agency can elect to do a Delphi survey. In a Delphi survey, the agency sends information regarding the forecast and asks various experts for feedback. This feedback is then summarized by the agency and sent out again to the experts. The iterative nature of a Delphi survey helps to achieve consensus on the forecast. Econometric Models Econometric models, like regression analysis, are used to quantify relationships between vari- ables. This involves identifying the independent and dependent variables that affect a particular revenue source and then establishing the relationship between the two. An agency might have a different set of regression equations for each revenue source (Wachs and Heimsath 2015). Econometric models can be as simple as linear regression that defines a relationship between a single independent variable and a dependent variable. For example, the relationship between the price per gallon of gasoline and the gallons of gasoline consumed could be described by linear regression. Models can also be more complicated by incorporating multiple independent variables (also called multivariate) to predict an outcome. These models find a fit for the input data and predict revenue from a source based on statistically estimated coefficients of the inde- pendent variables (Wachs and Heimsath 2015). NCHRP Synthesis 479 cites the Oregon Department of Transportation’s use of econometric models. The Oregon DOT uses more than 200 equations to forecast revenue. To get a sense of the kinds of variables included in an econometric model, the forecast equation for motor fuel, the sale of which is a main source of revenue for the state, includes the following independent variables: • Real retail price index for motor fuels, • Fuel efficiency of the existing stock of light-duty vehicles, • Oregon total non-farm employment, • Oregon real aggregate personal income, • Oregon labor force participation rate, • Consumer sentiment index, and • Variable for the implementation of an ethanol blend mandate. Again, an econometric model can be as sophisticated as an agency desires by incorporating independent variables to describe a particular outcome. Case Studies NCHRP Synthesis 479 (Wachs and Heimsath 2015) provides case studies on how different states do their projections. Table 4-2 sum- marizes three states from the report and includes the methods and data inputs they use to obtain future funding levels. TIP! All forecasting methods involve differing degrees of sophistication in the application of statistics. While trend extrapolation works with stable funding sources, expert consensus can work well when funding and expenses are subject to volatile market and political forces, among others, and econometric models are preferable when reliable data exists to predict funding and price level.

44 A Guide to Developing Financial Plans and Performance Measures for Transportation Asset Management Forecasting Uses While the TAM professional of a state DOT may gather revenue forecasting information from colleagues across state and local government, the projection of uses—or the “estimated cost of future work types” as described in the TAMP Rule—often demands greater involvement. The TAM professional may have to coordinate with bridge engineers and pavement engineers on construction cost inflation assumptions and discount rates (see discussion of these below). The TAM professional will likely play a role in the setting of performance targets that help determine funding needs, and even help achieve consistency between various program managers in their forecasting and reporting methodologies. There are several uses of funds that are commonly prioritized above asset management uses. For example, TAM professionals should look to budget officers to provide known annual debt service payments in building the forecast. Likewise, dedicated federal programs such as state planning & research or transit funding may be shown as a source in a comprehensive forecast but then should be allocated to their corresponding uses. Comparable state programs may also be delineated. Existing guidelines from FHWA on forecasting expenditures for the STIP can be applied to predicting expenditures for a financial plan. For capital costs, FHWA recommends basing State Forecasted Funds Methods Data Inputs Arizona • Federal funds allocated to state • State gas tax • State diesel fuel tax • State vehicle registration/license fees • Local sales tax • Local property tax • Expert Consensus • Econometric • Nominal personal income growth • Population growth • Construction employment growth • Phoenix Consumer Price Index • Sky Harbor passenger traffic growth • Non-farm employment growth • 30-year mortgage rate Minnesota • Federal funds allocated to the state • State gas tax • State diesel fuel tax • State vehicle registration/license fees • State sales taxes for transportation • Trend Extrapolation • Historical trends • National macroeconomic forecasts of fuel consumption • Regional forecast information from the federal Energy Information Administration Oregon • Federal funds allocated to the state • State diesel fuel tax • State vehicle registration/license • Weight-mile tax • Econometric • Uses national values and Oregon-specific inputs • Consumer Price Index • Employment • Housing starts • Real GDP • Real fuel price • New automobile sales • Fuel efficiency Table 4-2. State forecasting examples.

Financial Forecasting 45 a forecast on “historical costs for projects of comparable scale and design.” For operating and maintenance costs, FHWA suggests using “historic data applied on a per-lane mile and functional classification basis or an annual lump sum basis” (FHWA Office of Planning, Environment, & Realty 2017). FHWA also acknowledges the iterative nature of forecasting expenditures. Scenarios allow certain key assumptions or variables to change from iteration to iteration. Those changes—such as the anticipation of the repeal of state legislation or the acceleration of freight vehicle miles traveled—may impact revenues, expenses, or both. Scenario planning, the process of creating scenarios that assume different levels of funding to see the impact on overall agency asset man- agement, can help inform the projection of revenues and expenditures (Twaddell et al. 2016). Developing investment strategies impacts the allocation of the expenditures. Both concepts are detailed in the following chapter on Investment Strategies and Scenarios; it is important for an asset manager developing a financial plan to consider scenario planning and investment strate- gies even in the stages of projecting future revenues and expenditures. Assumptions, Inflation, and Uncertainty Assumptions, inflation, and uncertainty are all inherent in the process of forecasting. The individual or group that develops the forecasts within a state DOT makes certain assumptions about funding levels, selects the inflation rates and how to apply them, and deals with uncer- tainty. Below are some common issues that TAM professionals may encounter as they handle assumptions, inflation, and uncertainty in the TAM financial plan. Assumptions Dealing with assumptions in forecasting is another aspect of financial planning that can draw from the practices agencies already use to develop the STIP. The key is to determine which assumptions are reasonable and which are unreasonable. In guidance for developing the STIP, FHWA notes two important considerations in determining whether an assumption for includ- ing a revenue source in a forecast is reasonable: 1) “There is evidence of review and support of the new revenue assumption by state and local authorities; 2) There is documentation of the rationale and procedural steps to be taken with milestone dates for securing the funds” (FHWA Office of Planning, Environment, & Realty 2017). In order to include a revenue source in a financial plan and proceed with a projection of funds into the future, it should be clear that the funds will be available. Evidence of support from the government and/or documentation provides the clarity necessary to consider the assumption of funds reasonable. FHWA provides a few examples of reasonable and not reasonable funding sources to assume when developing a financial plan. These are summarized in Table 4-3. This is not a comprehensive list of assumptions, but gives asset managers a flavor of the assumptions that are reasonable to make. It is important to document all the assumptions that are made during the process of pro- jecting revenues and expenditures. This will ensure clarity when developing other parts of the financial plan and consistency when making future projections. Inflation Inflation, the general increase in prices over time, is an important consideration in making financial projections. Financial projections should be presented in year of expenditure dollars— also called nominal or current dollars. In many cases it is necessary to make assumptions about inflation to project future costs and revenues. Even when it is not necessary to make explicit

46 A Guide to Developing Financial Plans and Performance Measures for Transportation Asset Management inflation assumptions, it is still important to estimate future inflation when using management system models, as these work either in real or constant dollars (dollars that exclude inflation) or incorporate inflation explicitly. There are multiple inflation indices as prices of different goods increase at different rates. For instance, the Consumer Price Index is used to establish the infla- tion rate for consumer goods; and the National Highway Construction Cost Index is often used to predict construction inflation, as are various other industry and state-specific indices. FHWA allows considerable flexibility in estimating inflation. For STIP financial projections FHWA requires that “cost and revenue estimates . . . use an inflation rate to reflect ‘year of expenditure dollars’ based on reasonable financial principles” (23 C.F.R. § 450.324 2016). Con- cerning which inflation index to use, FHWA suggests examining past trends in tax receipts and cost of living indices to forecast revenue sources. For future expenditures, it is suggested that agencies use a construction cost index for projections. When possible, the FHWA encourages states and metropolitan planning organizations (MPOs) to use local cost data to develop cost inflation indices. However, if this information is not available, a four percent annual inflation rate is acceptable (HWA Office of Planning, Environment, & Realty 2017). Inflation rates are different from discount rates and cannot be used interchangeably. Inflation rates measure historic growth or anticipate the future growth in unit costs. They help demon- strate the loss in purchasing power over time (assuming inflation prevails over deflation). By contrast, the discount rate is used to quantify the time value of money and is used to compute the present value of a stream of future revenues or costs. Uncertainty Due to the nature of financial projections and the assumptions inherent in the process, there will be uncertainty in the estimates of revenue and expenditure in the financial plan for asset management. Two methods for understanding the impact of uncertainty on financial projec- tions are to consider different funding scenarios and to perform a sensitivity analysis. Again, scenario planning is covered in more detail in the next chapter of this guide but can be helpful in the stage of projecting revenue and expenditure. Asset managers may develop scenarios to illustrate different levels of funding over the long term. For instance, one scenario might maintain current levels of funding, another might show an increase in funding, and a third Reasonable to Assume Unreasonable to Assume A new toll or other user fee, if there is clear evidence of support by the governor, legislature, and/or other appropriate local/regional decision maker. New funds from upcoming statewide, regional, or local ballot if polls indicate strong likelihood of defeat or there is past history of repeated defeat of similar ballot initiatives. A new tax for transportation purposes, if there is clear evidence of governmental and public support to enact the tax and a strategy exists for securing approvals. An increase in gas tax revenues if there is not a history of consistent gas tax increases. An increase in the state gas tax, if there is historical success in increasing the gas tax and it is clear that the trend will continue. Funding that would result in one metropolitan area receiving a disproportionately high percentage of the total national program dollars to support multiple large-scale transportation projects. Table 4-3. Differences between reasonable and unreasonable assumptions regarding revenue sources.

Financial Forecasting 47 could illustrate the effects of decreased transportation funding. Different scenarios help identify gaps in funding expectations versus needs. They illuminate the financial situation of the agency further in order to overcome the uncertainty in financial projections. Similar to scenario planning sensitivity analysis is a method to help understand how differ- ent inputs to the projection of revenues and expenditures impact the outcome. Changing the revenue sources, different independent variables contained within an econometric model, or the inflation rate, for example, can reveal what projection inputs influence the outcome the most. Other Variables A number of resources exist to help strengthen the use of economic and financial principles in the financial plan. The appendix provides further explanation on the time value of money and points the reader to additional literature. 4.2 Steps for Forecasting Revenues and Expenditures At this point you now have a spreadsheet listing your agency’s sources and uses of funding for asset management activities with dollar values for a single fiscal year. Now it is time to forecast revenues and expenditures for the next 10 years. The following steps guide you in understanding how your agency does forecasting for other financial planning activities and what information you need to forecast revenues and non-asset management expenditures for the TAM financial plan. Forecasting TAM-related expenditures is detailed in the next chapter. Step 1: Establish Roles and Responsibilities Your agency already does forecasting for other financial planning activities. The STIP and the long-range plan are two common plans, but there may be other purposes for which your agency does some form of forecasting and looking ahead at revenues and expenditures. As a TAM prac- titioner, you likely will not have to forecast revenues and expenditures for the TAM financial plan on your own. Take the time to determine who in your agency is responsible for forecasting, or those individuals outside the agency who may be able to contribute. These may include: • Budget officers; • Accountants; • Long-range planners and local planning partners; • Economists, demographers, and statisticians. You may also want to identify any senior managers responsible for presenting and defending your forecast when it is complete in addition to government relations officers, public informa- tion officers, and policy makers willing to help you communicate your forecast. Establish the gaps between the forecasting efforts already in place and your needs for the TAM financial plan. Specifically, you need a forecast of revenues and expenditures for TAM-related activities over a 10-year period. Then coordinate with the appropriate individuals or group to utilize forecasts from other financial documents and get the projections necessary for the TAM financial plan. Step 2: Review Prior Forecasts Once you have established who does the forecasting in your agency and what other finan- cial planning documents incorporate projections, familiarize yourself with the techniques and assumptions used in those forecasts. Does your agency use basic trend extrapolation or is there

48 A Guide to Developing Financial Plans and Performance Measures for Transportation Asset Management an econometric model? What assumptions are made about forecasting federal, state, and local funding? How about assumptions for fixed costs versus variable costs? Is the forecast in current dollars or year of expenditure dollars? What inflation rate(s) are used? Did the state government recommend certain assumptions for state funds in the coming years? Is the financial planning done by federal or state fiscal year? You should be able to reuse many of the same assumptions in the projections for the TAM financial plan and it is a good idea to maintain consistency across financial planning documents within the agency as well. While you are not on your own to forecast revenues and expendi- tures for the TAM financial plan, it is important to understand the methods forecasters use to obtain the result since you will likely be responsible for communicating the financial plan to stakeholders when it is complete. Step 3: Forecast Revenues Obtain a 10-year revenue forecast for sources related to asset management. Again, you should not be required to do this task individually, but you may find it necessary to extend the forecasts beyond what already exists at your agency. For example, the STIP is a 4- to 6-year forecast of the capital program at a DOT. You need a 10-year forecast of revenues for both capital and non- capital funds. To continue the projections for a longer timeline and include relevant revenues, use the methods and assumptions for forecasting you reviewed in the previous step. This way you can fill in the gap between what projections already exist and what you need for the TAM financial plan. If possible, you can utilize the same format for forecasting as other financial planning docu- ments in your agency. This will help maintain consistency. Included in the next section is a template for forecasting in the case that it is not efficient to use other formatting techniques. In addition, Figure 4-1 provides an example from the Minnesota Department of Transportation (MnDOT), which shows a sampling of their revenue projections (MnDOT 2017a). Figure 4-1. MnDOT forecast (MnDOT 2017a).

Financial Forecasting 49 Also, at this stage it is important to ensure agreement on the level of funding to include in the financial plan. For example, some managers may want to demonstrate drastic underfunding (e.g., a cut in funding of 50%) in order to emphasize the need for additional resources. Others may want to be aggressive about revenue assumptions (e.g., assuming an increase in funding of 25%) in order to build a full capital program that meets the needs of everyone and sup- ports robust regional TIPs and STIPs. Get input from managers and other relevant stakeholders throughout the process to achieve consensus on the level of funding. Step 4: Forecast Non-asset Management Uses Next obtain information on non-asset management expenditures that you are including in the plan. This may include, for example, salary increases over the next 10 years as well as other fixed costs. Again, these projections should already be available at your agency. Note that these expenditures may be ranges of expected costs for the various categories. Determine how to best convey these ranges and fill in the appropriate columns in your financial plan. Step 5: Determine Available Funding for Asset Management Following the projection of revenues and planned expenditures allocated for uses other than asset management, determine how much funding is available for asset management. You may have this information readily available from the previous chapter or it may be the straight- forward difference between the total available funding you have documented and the non-asset management uses you have determined. The figure could be a range as well. This step helps build the foundation for the following chapter where you will determine how to allocate the available funding for asset management. Step 6: Document Key Assumptions Finally, it is important to document the assumptions you made throughout the process of forecasting. Some of the key variables, methodologies, and assumptions you want to note include: • Rates of growth and inflation; • Anticipated changes in legislation, funding formulas, debt service schedules, or other timing issues; • Sources of your data and information if not derived from your own group; • Fixed costs or minimum funding thresholds for certain uses; and • Methods for forecasting beyond the last year included in the forecast, if the method is different from simply applying an inflation or growth rate.