Guide to Effective Methods for Setting Transportation Performance Targets (2022)

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.

102 Measures The traffic congestion measures—annual PHED per capita and non-SOV mode share—apply to each U.S. Census Bureau–designated UZA that contains an NHS road, has a population of more than 1 million, and contains any part of a nonattainment or maintenance area for ozone, carbon monoxide, or particulate matter. States and MPOs within an applicable UZA must coordinate to establish a single, unified target for the UZA for each of these measures. All states and MPOs with NHS mileage that overlaps within the applicable UZA must coordinate on a single, unified target and report on the measures for that area. Annual Peak Hour Excessive Delay per Capita The PHED traffic congestion measure is computed on the basis of excessive delay, where excessive delay is the extra amount of time spent in congested conditions defined by speed thresholds that are lower than the normal delay threshold. For TPM reporting on this conges- tion measure, the speed threshold is 20 miles per hour (mph) or 60% of the posted speed limit, whichever is greater. The PHED calculations need to be done only for the peak periods during weekdays (morning peak, 6–10 a.m.; afternoon peak, 3–7 p.m. or 4–8 p.m.). The performance measure per capita is computed as follows: 1. The annual excessive person hours of delay during the peak periods over all reporting segments in the UZA are calculated by taking traffic volumes and vehicle occupancy into account. 2. The aggregated annual excessive person hours of delay are divided by the total population of the UZA. Non-Single-Occupancy-Vehicle Mode Share For the non-SOV mode share performance measure, states and MPOs have an option of various data sources, including the U.S. Census Bureau’s ACS mode share data, local travel survey, or samples collected in the region. Most states choose the ACS-provided data on the percentage of workers ages 16 and older who used a transportation mode other than a single- occupancy vehicle to commute to work. Challenges Associated with Setting and Revising Targets Annual Peak Hour Excessive Delay per Capita Challenges with PHED targets included the following: • Messaging related to traffic congestion targets can be challenging, since it reflects only exces- sive delay, a concept with which many are unfamiliar. Target-Setting Methods for Traffic Congestion

Target-Setting Methods for Traffic Congestion 103 • Coordinating with multiple agencies under one UZA adds another layer of complexity to setting targets, especially if there are competing priorities. • During the pilot testing phase of NCHRP 23-07, the analyses revealed concerns regarding data quality, notably in relation to the data from the NPMRDS used for system performance and freight movement. State DOTs identified questionable data from NPMRDS for some roadways or some periods of the day, which created challenges for the analysis. • Many states cited uncertainties related to COVID-19 as a big limitation in setting targets in 2020 and beyond and in terms of making adjustments to targets. Non-Single-Occupancy-Vehicle Mode Share Challenges related to non-SOV mode share targets include the following: • For traffic congestion measures focused on the UZA level, the state DOTs faced challenges in coordinating with MPOs to conduct runs with the regional travel demand forecasting model, which was one of the recommended methods for testing. • MPOs and states noted challenges in considering capital improvements and attributing impacts of specific investments to performance of these measures. • Some of the MPOs interviewed (e.g., Philadelphia’s Delaware Valley Regional Planning Commission) wanted to use other data sources to supplement ACS data for non-SOV travel, but data from other modes were not comparable and were of varying quality and completeness. Some MPOs noted their data would improve with time as they completed more activities, such as sidewalk traffic counts, to better understand pedestrian activity levels. • At least one MPO noted the need for more guidance on whether to use overlapping 5-year ACS data. • Coordinating with multiple agencies under one UZA adds another layer of complexity to setting targets, especially if there are competing priorities. Summary of Target-Setting Methods for Traffic Congestion Annual Peak Hour Excessive Delay per Capita Most states rely on the NPMRDS data when setting targets for the PHED measure, as with the reliability and freight measures. Table 13 shows the categories into which these target-setting methods fall, along with their strengths and limitations. Non-Single-Occupancy-Vehicle Mode Share For target setting for non-SOV mode share, states and MPOs generally use one of the four methods shown in Table 14, which range from a simple trend analysis to a more-advanced travel forecasting model. States and MPOs coordinate to establish the non-SOV targets for relevant UZAs. Agencies have several options for data sources to use in developing the target, including the ACS, a local travel survey, or locally collected samples or continuous counts of travelers using non-SOV modes. Most agencies, however, used ACS data as the starting point for their analysis. The ACS provides information on the percentage of workers ages 16 and older who commuted to work using a mode of transportation other than a single occupancy vehicle.

104 Guide to Effective Methods for Setting Transportation Performance Targets Method Strengths Limitations Other Considerations Building off the baseline with assumptions Maintaining the baseline level as the target or making an adjustment on the basis of judgement Simple, easy to communicate, and often brings in stakeholders There may be no rigorous methods for the adjustments Method for agencies with limited data; agency will need to decide which exogenous factors are relevant Time series trend analysis Forecast based simply on historical performance trend Simple approach; does not require special analysis tools; data driven No insights into causes of outcomes May result in a worsening target, which can pose communications challenges or conflict with stated goals Trend plus other factors Expands upon trend analysis to account for other factors that may shift future performance Still relatively simple, data driven, and brings in additional factors There may still be no rigorous methods for the adjustments; sometimes adjustments may not be data driven May result in a worsening target, which can pose communications challenges or conflict with stated goals; agency will need to decide which exogenous factors are relevant Travel forecasting model Uses regional travel model to forecast future congestion, often with anticipated change applied to baseline PHED Fuller understanding of causes of outcomes, fully data driven, and may support linking the target-setting process with decision-making by informing what factors can be influenced Models often do not account for all factors well, such as bicycle/pedestrian improvements and telework policies May result in a worsening target, which can pose communications challenges or conflict with stated goals; generally requires additional model analysis beyond what is typically conducted for (long-range) planning Table 13. Target-setting methods for annual PHED per capita measure. Method Strengths Limitations Other Considerations Time series trend analysis Forecast based simply on historical performance trend Simple approach; does not require special analysis tools; data driven No insights into causes of outcomes May result in a worsening target, which can pose communications challenges or conflict with stated goals Trend plus other factors Expands upon trend analysis to account for other factors that may shift future performance Still relatively simple, data driven, and brings in additional factors There may still be no rigorous methods for the adjustments; sometimes adjustments may not be data driven May result in a worsening target, which can pose communications challenges or conflict with stated goals; agency will need to decide which exogenous factors are relevant Policy based Target is set on the basis of a policy direction to increase non-SOV mode share Simple, easy to communicate, and brings in stakeholders; in line with agencies’ aspirations May not be realistic — Travel forecasting model Uses regional travel model to forecast future mode share, often with anticipated change applied to baseline Fuller understanding of causes of outcomes, fully data driven, and may support linking the target-setting process with decision-making Models often do not account for all factors well, such as bicycle/pedestrian improvements and telework policies May result in a worsening target, which can pose communications challenges or conflict with stated goals; generally requires additional model analysis beyond what is typically conducted for (long-range) planning Table 14. Target-setting methods for non-SOV mode share.

105   What It Is Building off the baseline with assumptions uses the baseline value with some assump- tions, either to maintain the baseline level or to adjust it on the basis of consideration of factors that might affect future performance. This method may be applied for its simplicity and ease of application, especially when the agency does not have access to historical data or data on external influencing factors. Though this method is quan- titative, it allows states to consider qualitative information on future plans and polices as well as expert professional judgment in the target-setting process. In some cases, these adjustments to the baseline account for analyses of traffic or policies and plans. When to Use It This method, though quantitative, is the easiest to implement. It works well for agen- cies that are not able to gather extensive historical data on factors that can influence performance. This method can be applied with the availability of some degree of future estimates of factors that may influence performance in the future. What Is Needed PHED data are needed to establish the baseline. Typically, the PHED data can be gathered from the RITIS NPMRDS analytics tool. The qualitative or quantitative data or information on relevant factors such as planned construction projects, nearly completed improvements, and socioeconomic trends and other related data can be used to inform whether an adjustment should be made to the baseline. How to Do It Step 1: Establish Baseline Establish the baseline for PHED with NPMRDS data. Step 2: Develop Adjustment On the basis of the data and information gathered on factors that might affect performance, come to a consensus and develop an adjustment to be applied to the baseline that is based on expert professional judgement. F A C T S H E E T Annual PHED per Capita Method 1: Building off the Baseline with Assumptions AT A GLANCE Ease of application: Technical robustness: Ease of communication: Allows for policy preference:

106 Guide to Effective Methods for Setting Transportation Performance Targets Step 3: Apply Adjustment Apply the adjustment to the baseline PHED measure. Step 4: Compute Target Compute the future year target on the basis of the adjustment applied to the baseline annual PHED per capita measure. Advantages This method is simple and easy to communicate. It often involves many stakeholders who could have different outlooks on the future performance and may bring in valuable perspectives. Limitations Although some quantitative or qualitative factors may be taken into account, no rigorous analytical methods are used. Examples The St. Louis, MO–Illinois UZA’s partners (the Illinois DOT, the Missouri DOT, and the East–West Gateway Council of Governments) agreed to set the target at the baseline level of 9.50 annual hours of PHED per capita (FHWA 2020). The Chicago, IL–Indiana UZA showed a 2017 baseline of 14.9 annual hours of PHED per capita from NPMRDS tools. The UZA’s respective agencies considered trend data and other factors in setting the target, including planned construction and agency goals of increasing transit ridership and transit-supportive land use and improving traffic operations (FHWA 2020). The Minneapolis–Saint Paul, MN–Wisconsin region had 8.65 annual hours of PHED per capita in 2017, with figures ranging from 10.83 to 11.00 between 2014 and 2016. An ambitious 4-year target of 8.5 hours per capita was selected to reflect the region’s desire to reduce hours of delay while also accounting for major projects that could negatively affect the measure (FHWA 2020).

107   What It Is A trend analysis relies upon a historical trend to forecast future performance and set the target close to the projected performance level. Time series trend analysis is a simple method that can be used for target setting for the PHED performance measure. State DOTs can choose to use a linear forecasting function or one of the other functions available in analysis software (e.g., exponential function) to forecast the target year’s performance. When employing this method, agencies should pay attention to significant changes in trends or inconsistencies in the data sources during the analysis period. For instance, recognizing the dramatic change in travel that occurred as a result of the COVID-19 pandemic, agencies could break the trends into pre-COVID, COVID, and post-COVID periods to better understand the trends and support forecasting for targets. Doing this requires some additional work to explore trends and make judgements about which trend is likely to occur in the future. When to Use It This method may be useful for agencies with limited data about exogenous factors. What Is Needed PHED data are needed to establish the trend line. Typically, the PHED data can be gathered with the RITIS NPMRDS analytics tool. A spreadsheet application or statistical software package is needed for the analysis. How to Do It Step 1: Collect Data for Selected Years Historical data on annual PHED per capita should be collected with the RITIS NPMRDS analytics tool. Step 2: Establish Trends and Projection After the historical data have been collected and plotted, a trend line needs to be established. This can be done with a spreadsheet application or another statistical analysis tool. Agencies F A C T S H E E T Annual PHED per Capita Method 2: Time-Series Trend Analysis AT A GLANCE Ease of application: Technical robustness: Ease of communication: Allows for policy preference:

108 Guide to Effective Methods for Setting Transportation Performance Targets can generate linear trends or use other functions (e.g., exponential, logarithmic) available in the analysis software. The projections can be developed by using either simple linear forecasting or nonlinear methods such as exponential smoothing or ARIMA. Following are some commonly used fore- casting techniques: • Linear trends, which are continuations of the line of best fit for historical data and are the most common approach to establishing a projection. Forecasts can be established through the FORECAST function in Excel (FORECAST.LINEAR in newer versions) or similar functions in other software. • Exponential smoothing, which weights recent observations more heavily than older ones in estimating future points. These models are particularly useful in handling seasonality in trends. They likely are not helpful if only limited years of data are available, but they may be a good option once more data become available. Excel has a function for exponential triple smoothing (FORECAST.ETS) that can be applied to historical data. • ARIMA, which uses correlations in past observations as the basis for future trends. Again, these models are likely not helpful for the limited years of data currently available, but they may be a good option once more data become available. Advantages This method is simple while still being data driven. Limitations This method generally assumes that historical trends will hold into the future, which may not be the case, and it does not provide insights into causes of outcomes. It misses incorporation of new factors that may influence targets in the future. It may result in a worsening target, which can pose communication challenges. Examples Atlanta, GA, UZA: During the first reporting period, regional partners used the monthly values for the total annual hours of PHED per capita measure from January 2016 to December 2017 to develop a linear trend line, which showed an increase over time. They selected the lower end of the projected 2021 range as the 4-year target. Memphis, TN-Mississippi-Arkansas UZA: Memphis MPO staff reported that they tested three regression formulas (linear, exponential, and logarithmic) to calculate 2018–2021 esti- mates. To understand future trends, they calculated estimates in several iterations: 1. Analysis of 2014–2016 HERE data applied to 2016–2017 INRIX data; 2. Analysis of 2014–2016 HERE data applied to the HERE data; and 3. Analysis of 2016–2017 INRIX data applied to the INRIX data. Through collaboration among the state DOTs, the Memphis MPO, and the West Memphis MPO, it was agreed to use the PHED per capita values obtained from the linear regression formula derived from analysis of the 2014–2016 HERE data applied to the 2016–2017 INRIX data to set the target. Seattle area UZA: As part of the pilot testing for NCHRP 23-07, the Washington State DOT tested the time series trend analysis method and used data on the annual hours of PHED per

Target-Setting Methods for Traffic Congestion 109 capita from 2013 to 2020 to forecast the 2022 value, as shown in Figure 24. The Washington State DOT used the yearly NPMRDS data and developed the 2022 forecasts by including and excluding the 2020 data. The agency anticipated utilizing this method with the latest available data when the targets were set for submission in October 2022. By then, data on the annual hours of PHED for 2021 would be available. The Washington State DOT planned to coordinate with the Puget Sound Regional Council and other agencies within the UZA to establish the CMAQ congestion target. 21.5 32.4 0 5 10 15 20 25 30 35 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 PHED - NPMRDS Forecast Includes 2020 Forecast Excludes 2020 Linear (PHED - NPMRDS) Figure 24. Time series trend analysis for annual hours of PHED per capita for the Seattle UZA.

110 What It Is A trend plus other factors analysis expands on a basic trend analysis by including other data, factors, or inputs that are determined to influence performance. These factors include economic indicators, travel projections, population growth, construction impacts, or network improvements that influence roadway congestion and delays. Inputs may also include stakeholder input in the form of professional judgement. The trends observed in the forecasted additional factors can guide the adjustments of the targets of the performance measures. This method is recommended when the agency has access to historical data on the performance measure and external influencing factors. The analysis might also include more qualitative stakeholder input or policy judgements (such as decisions to adjust the targets to be more conservative). The forecasted data on explanatory factors that are needed to apply the trend plus other factors method are sometimes not readily available. These data may be a chal- lenge. For instance, forecasts for explanatory factors developed prior to the COVID-19 pandemic may not hold true after the pandemic. It is generally accepted that the COVID-19 pandemic could have long-term effects on travel behavior and economic factors, but the extent and duration of these impacts are unknown. When to Use It This method can be applied when agencies have access to historical PHED data as well as data on factors influencing performance. What Is Needed PHED data are needed to establish the trend line. Typically, the PHED data can be gathered with the RITIS NPMRDS analytics tool. In addition, data on other factors, such as anticipated future funding levels, new or planned roadway construction, and VMT forecasts, are typically used. A spreadsheet application or statistical software package is needed for the analysis. How to Do It Step 1: Prepare Data Obtain the monthly or annual PHED data by using the RITIS NPMRDS analytics tool. F A C T S H E E T Annual PHED per Capita Method 3: Trend Plus Other Factors AT A GLANCE Ease of application: Technical robustness: Ease of communication: Allows for policy preference:

Target-Setting Methods for Traffic Congestion 111 Step 2: Identify Trend Use linear, exponential, or logarithmic analysis to identify a trend. Step 3: Identify and Analyze Factors In this step, the agency examines factors that may have an impact on vehicular congestion. This step may include analyses of historical patterns and forecasts of different parameters, such as VMT, or information on major transportation projects that would be anticipated to affect delay on the highway network (either by adding delay during construction or by alleviating delay after opening). Step 4: Adjust Forecast Trend The trends observed in the forecasted additional factors can guide the adjustments of the targets of the performance measures. This adjustment is typically developed on the basis of judg- ment of what is reasonable to expect. This step often involves soliciting input from partners to agree upon a reasonable adjustment to the trend line figures. Advantages This method is data driven. It allows for consideration of additional factors. Limitations There may still be no rigorous methods for the adjustments, and sometimes adjustments may not be data driven. This method may result in a worsening target. Examples Charlotte, NC–South Carolina UZA: The Metrolina Coordination Group met and reviewed possible targets based on a trend line forecast. Figure 25 shows that PHED increased year over year from 2014 to 2017. The North Carolina DOT explored various trend lines, including a linear trend and an exponential trend, to develop a range of possible targets. The agency also considered additional factors, including VMT growth, population growth, economic trends, and more, in selecting the target (North Carolina DOT 2018).

112 Guide to Effective Methods for Setting Transportation Performance Targets Source: North Carolina DOT (2018). Figure 25. North Carolina DOT trend analysis for annual hours of PHED per capita for the Charlotte UZA.

113   What It Is Agencies can utilize regional travel demand models covering UZAs to help account for factors that may affect excessive delay on the highway network. The travel demand fore- casting model can be used to estimate excessive delay for the base year and the future forecasted year. To use this method, the region should have access to a travel demand forecasting model that covers the entire UZA. If the travel demand forecasting model covers a different geography (an area smaller or larger than the UZA), then adjustments may be needed to estimate excessive delay for the UZA region. Typically, a regional travel demand model or a statewide travel demand forecasting model can be used to obtain an estimate of exces- sive delay for the UZA for the base year as well as the forecasted year for which the target is to be set. The trend observed in the estimated delay from the base to the future year can be used to inform the target-setting process for the PHED performance measure. When to Use It This method can be utilized when a regional travel demand forecasting model that covers the UZA is available. What Is Needed PHED data are needed to establish the baseline. Typically, these data can be gathered with the RITIS NPMRDS analytics tool. A regional travel demand model covering the UZA is also necessary. How to Do It Step 1: Prepare Data Gather data for the annual PHED per capita for the UZA for the base year from NPMRDS. Step 2: Gather Estimates Obtain the results of the regional travel demand model covering the UZA for the base year and the forecast year. Ideally, the year for which the model results are collected should be the same as the base year of the model. If the model results are not readily available, the regional model may F A C T S H E E T Annual PHED per Capita Method 4: Travel Forecasting Model AT A GLANCE Ease of application: Technical robustness: Ease of communication: Allows for policy preference:

114 Guide to Effective Methods for Setting Transportation Performance Targets have to be run for the base year and the future year. If the model covers a larger area than the UZA, further adjustment may be needed to obtain the sum of excessive delay over the modeled area. Most travel demand models estimate delays in the modeled area over each segment. By using a threshold over which delay is considered to be excessive, estimates of excessive delay can be computed from the delay estimates from the model (excessive delay from the model = modeled delay − threshold). The threshold may need to be based on the thresholds set by the RITIS NPMRDS tools or could be based on the way the model handles delays. The excessive delays are then summed over the entire UZA for the base year and the target year separately. Step 3: Compute Change Compute the growth or reduction (or percentage change) in the estimated excessive delay between the base year and the future year. Step 4: Estimate PHED Per Capita Apply the percentage change in growth or reduction either directly or indirectly by reducing it or increasing it to the PHED base number obtained from NPMRDS to obtain an estimate of the PHED per capita for the year for which the target is to be set. Advantages This method provides a fuller understanding of causes of outcomes. The method takes into consideration many factors, such as changes in development, traffic volumes, and transporta- tion projects, that are accounted for in the established model in the region. Limitations Travel demand models are typically utilized for forecasting over longer periods; their use in forecasting over a shorter period may not have been studied sufficiently. Moreover, these models often do not account well for all factors, such as bicycle/pedestrian improvements and telework policies. Examples Milwaukee, WI, UZA: The Southeastern Wisconsin Regional Planning Commission used its regional travel demand model as a proxy to estimate change in total annual average delay per capita. The percentage change from the model was applied to baseline performance based on the 2017 NPMRDS data; however, the percentage change was tempered because there are differ- ences between modeled delay and actual excessive delay based on collected data. As a result, the stakeholder workgroup selected 8.6 hours PHED per capita as the target for 2021 (FHWA 2020). Washington, DC–Virginia–Maryland UZA: The National Capital Region Transportation Planning Board (NCRTPB) used forecasts from the MPO’s travel demand model to produce outputs of congestion for modeled years 2016, 2020, and 2025 based on VMT estimates for the a.m. peak hour. The forecasts accounted for anticipated changes in population, employment, land use, and other factors that affect travel demand, such as changes in the highway and transit network. The forecast percentage change in congestion was then applied to the baseline measured performance. [Note: NCRTPB then averaged the results of this modeled forecast analysis with an extrapolation of measured data from the NPMRDS to establish the targets (NCRTPB 2018).]

115   What It Is A trend analysis relies upon a historical trend to forecast future performance and set the target close to the projected performance level. Time series trend analysis is a simple method that can be used for target setting for the non-SOV mode share. State DOTs can choose to use a linear forecasting function or one of the other functions available in their analysis software (e.g., exponential function) to forecast the target year’s performance. Trend analysis uses historical trends to see where performance is expected to go. A target is then set at or close to the projected performance level. States typically utilize ACS data to conduct a trend analysis. In some cases, linear, exponential, and logarithmic trends are analyzed to identify the best fit. When to Use It This method may be useful for agencies with limited data about exogenous factors. What Is Needed A spreadsheet application or statistical software package is needed for the analysis. How to Do It Step 1: Collect Data for Selected Years Collect the non-SOV mode share data from the ACS for the selected number of years for which the trends can be developed. Step 2: Establish Trends and Projection After the historical data are collected and plotted, a trend line needs to be established. This can be done with Excel or another statistical analysis tool. Agencies can generate linear functions or other functions (e.g., exponential, logarithmic) available in the analysis software. The projections can be developed by using either simple linear forecasting or nonlinear methods such as exponential smoothing or ARIMA. Following are some commonly used fore- casting techniques: • Linear trends, which are continuations of the line of best fit for historical data and are the most common approach to establishing a projection. Forecasts can be established through the F A C T S H E E T Non-SOV Mode Share Method 1: Time-Series Trend Analysis AT A GLANCE Ease of application: Technical robustness: Ease of communication: Allows for policy preference:

116 Guide to Effective Methods for Setting Transportation Performance Targets FORECAST function in Excel (FORECAST.LINEAR in newer versions) or similar functions in other software. • Exponential smoothing, which weights recent observations more heavily than older ones in estimating future points. These models are particularly useful in handling seasonality in trends. They likely are not helpful for the limited years of data currently available, but they may be a good option once more data become available. Excel has a function for exponential triple smoothing (FORECAST.ETS) that can be applied to historical data. • ARIMA, which uses correlations in past observations as the basis for future trends. Again, these models are likely not helpful for the limited years of data currently available, but they may be good options once more data become available. Advantages This method is simple while still being data driven. Limitations This method does not provide insights into causes of outcomes. This method may result in a worsening target, which can pose communications challenges or conflict with stated goals. Examples Seattle area UZA: As part of the pilot testing for NCHRP 23-07, the Washington State DOT used the historical data on non-SOV mode share from the ACS (2012–2016) to forecast the 2022 and 2026 non-SOV mode share targets. Both the linear and exponential functions available in Excel were used. For the pilot, the Washington State DOT developed target estimates both for the Seattle area UZA and the entire state of Washington, as shown in Table 15. Year Non-SOV Mode Share (%) Linear Forecasts Exponential Forecasts Washington State Seattle UZA Washington State Seattle UZA 2012 27.6 31.2 27.6 31.2 2013 27.3 31.1 27.3 31.1 2014 27.3 31.3 27.3 31.3 2015 27.4 31.5 27.4 31.5 2016 27.7 32.0 27.7 32.0 2017 27.6 32.0 27.6 32.1 2018 27.6 32.2 27.7 32.3 2019 27.6 32.4 27.7 32.5 2020 27.6 32.6 27.7 32.7 2021 27.7 32.8 27.8 32.9 2022 27.7 33.0 27.8 33.1 2023 27.7 33.2 27.9 33.4 2024 27.8 33.4 27.9 33.6 2025 27.8 33.6 28.0 33.8 2026 27.8 33.8 28.0 34.0 Note: Five-years of ACS data were used. Table 15. Results of time series trend analysis and forecasting method used by the Washington State DOT for non-SOV mode share.

117   What It Is The trend plus other factors method expands on a trend analysis by including addi- tional data, factors, or other inputs that are expected to influence performance beyond trend line perfor mance levels. This information can include relevant external factors such as economic indicators, travel projections, and population growth as well as fac- tors that could affect non-SOV travel patterns in the future. These other factors can include future funding, the impacts of new or planned transportation projects, gas prices, development patterns, and more. In some cases, agencies made policy judgements or assessments of the likely effects of these factors rather than account for them through quantitative or other statistical analysis. When to Use It This method can be applied when the agency has access to data on historical non-SOV mode share and factors influencing performance. What Is Needed Non-SOV mode share data for the UZA region from the ACS or a regional survey are needed to establish the trend line. In addition, other data elements, such as data on future funding, new or planned transportation projects, gas prices, development patterns, population and employment growth, and economic trends, are typically used. How to Do It Step 1: Collect Data Collect the non-SOV mode share data from the ACS for the selected number of years for which the trends can be developed. Step 2: Conduct Analysis Use linear, exponential, or logarithmic analysis to identify the trend. Step 3: Forecast Performance Use the trend line to forecast anticipated future performance. F A C T S H E E T Non-SOV Mode Share Method 2: Trend Plus Other Factors AT A GLANCE Ease of application: Technical robustness: Ease of communication: Allows for policy preference:

118 Guide to Effective Methods for Setting Transportation Performance Targets Step 4: Consider Other Factors Consider other data and influencing factors and how they might move the trend line. Advantages This method provides a fuller understanding of the causes of outcomes. It takes into consid- eration the expected growth from already established models in the region. Limitations There may still be no rigorous methods for the adjustments, and sometimes, adjustments may not be data driven. This method may result in a worsening target, which can pose communica- tions challenges or conflict with stated goals. Examples Chicago, IL–Indiana UZA: Targets were set by Illinois and Indiana in coordination with the Chicago Metropolitan Agency for Planning (CMAP) and the Northwestern Indiana Regional Planning Committee on the basis of ACS trends between 2012 and 2016. The process also accounted for the anticipated effects of CMAP’s ON TO 2050 goal of doubling transit ridership in the region by 2050 for non-SOV travel, and so adjusted the non-SOV mode share to account for the increase in ridership (FHWA 2020). Charlotte, NC–South Carolina UZA: For the non-SOV mode share, the North Carolina DOT and other agencies in the Charlotte UZA created a trend line with ACS data from 2012 to 2016. The trend line showed a slight downward trend of 21.8% to 21.5% between 2012 and 2016, but with year-by-year fluctuations. In addition to the trend analysis, the group discussed other factors that could influence non-SOV travel. The factors identified included “VMT growth, population growth, economic trends, ongoing highway construction, and completed projects that add transit or highway capacity” (North Carolina DOT 2018). The partners recognized that major projects anticipated to open would shift some travel from auto to transit, but the magnitude of the shifts over the 2 to 4 years was unclear. As a result, the Metrolina Coordination Group selected a conservative approach and set a target of 21.0% for both 2020 and 2022 (North Carolina DOT 2018).

119   What It Is An agency can utilize a regional travel demand model covering a UZA to help account for factors that may affect the non-SOV mode share. The travel demand fore casting model can be used to estimate non-SOV mode share for the base year and the future forecasted year. To use this method, the region should have access to a travel demand forecasting model that covers the entire UZA. If the travel demand forecasting model covers a different geography (an area smaller or larger than the UZA), then adjustments would be needed to estimate non-SOV mode share for the UZA region. The resulting estimated change in non-SOV mode share from the model analysis is typically applied to the base- line non-SOV value. When to Use It This method can be utilized when a regional travel demand forecasting model that covers the UZA is available. What Is Needed A travel demand model covering the UZA and a spreadsheet application or statistical software package are needed for the analysis. How to Do It Step 1: Establish Baseline Establish the baseline for the non-SOV mode share from the ACS data. Step 2: Forecast Use the travel demand model to forecast the non-SOV mode share for base year and target year. Care should be taken that the non-SOV modes from the model are consistent with the non-SOV modes from the ACS. Some regional models may not be able to estimate nonmotorized or telework mode shares. Adjustment or assumptions may need to be made to account for these modes. F A C T S H E E T Non-SOV Mode Share Method 3: Travel Forecasting Model AT A GLANCE Ease of application: Technical robustness: Ease of communication: Allows for policy preference:

120 Guide to Effective Methods for Setting Transportation Performance Targets Step 3: Calculate Target Value Calculate the target value by applying the change in the non-SOV mode share to the baseline figure. Advantages This method provides a fuller understanding of the causes of outcomes and accounts for planned investments (e.g., transit service improvements, changes in roadway infrastructure) that may affect mode shares. The method takes into consideration expected changes in land use, population, and employment from an already established model in the region. Limitations Gathering data from a regional travel demand model managed by partnering agencies can take some time and needs to be planned in advance, as the regional model may have to be run for the selected base or forecast year for the exact area of the UZA under consideration. If data from the regional model are not available for the years or the area under consideration, adjust- ments may be needed. In addition, because models often do not account for all non-SOV mode shares, such as nonmotorized modes and telework, additional analyses may be needed. Most models also do not account well for investments such as bicycle/pedestrian improvements. This method may result in a worsening target, which can pose communications challenges or conflict with stated goals. Examples Milwaukee, WI, UZA: The Wisconsin DOT and the Southeastern Wisconsin Regional Plan- ning Commission (SEWRPC) used ACS data to establish a trend line and then used SEWRPC’s travel demand model to separately forecast travel. The agencies decided to use the midpoint between the two analyses as the target. To inform their analysis, they also considered projects in SEWRPC’s fiscally constrained plan and identified those that were likely to be funded during the relevant period (FHWA 2020). Washington, DC–Virginia–Maryland UZA: NCRTPB used the region’s travel demand model as an input to set the non-SOV target for the UZA. The region’s travel demand model produced outputs of SOV mode share for commute trips for modeled years 2016, 2020, and 2025, accounting for anticipated changes in population, employment, and land use and changes in the highway and transit network. The forecast percentage change in SOV travel was then applied to the baseline measured performance. Staff noted that a midpoint of the forecasts from the travel demand model and the ACS historical trend line was used to establish the target.

121   What It Is A policy-based target is one that has been selected to reflect regional policy goals instead of relying (solely) on historical data or projections. Historical data and trend lines are explored, but the target itself is based on policy considerations. When to Use It This method may be appropriate in contexts in which targets are used as a communi- cations tool to motivate policy change. What Is Needed Policy-based target setting for non-SOV mode share requires non-SOV data from the ACS or a local survey, policy goals, and information about other influencing factors. How to Do It Step 1: Establish Baseline Establish the baseline for the non-SOV mode share from the ACS data. Step 2: Identify Relevant Policy Goals This method is accomplished mostly by qualitative analysis of policy, particularly if the region has already set goals for transit ridership or non-SOV mode share as part of long-range planning. In addition to considering existing long-range goals, agencies can consider relevant regional policy goals that could affect future mode shares (e.g., GHG reduction goals). Investment priori- ties, such as increased funding for transit and nonmotorized modes, as well as travel demand management policies, should also be considered. Step 3: Coordinate with Relevant Stakeholders As the non-SOV mode share performance measure is calculated for a UZA, it is important to coordinate with the agencies within the UZA. The state DOT, MPOs, local governments, transit agencies, and other partners should coordinate in developing these goals. It is all the more important for various agencies in the region to coordinate when a policy-based method is being used, as different agencies may have different policies, outlooks, and priorities. Non-SOV Mode Share Method 4: Policy Based F A C T S H E E T AT A GLANCE Ease of application: Technical robustness: Ease of communication: Allows for policy preference:

122 Guide to Effective Methods for Setting Transportation Performance Targets Step 4: Establish a Target After coordination of the relevant agencies, the target is set on the basis of collective judge- ment about how the policies in the region improve the usability of the alternative modes and reduce the attractiveness of SOV mode share. The target should be set by adjusting the baseline non-SOV mode share on the basis of qualitative analysis. Advantages This method is simple, easy to communicate, and brings in stakeholders in line with the agencies’ aspirations. Limitations This method is largely qualitative and does not include any robust quantitative analysis of the cause and effect of the factors influencing the targets. Examples Denver–Aurora, CO, UZA: The Colorado DOT worked with the Denver Regional Council of Governments (DRCOG) to develop a target based on current ACS data and goals in the Regional Transportation Plan. According to the CMAQ Performance Plan, “In Metro Vision, DRCOG has established a 2040 performance measure target of 35% non-SOV mode share to work. This was used as a basis for setting the 2020 and 2022 non-SOV unified target” (DRCOG 2020). Minneapolis–Saint Paul, MN–Wisconsin UZA: Between 2012 and 2016, the percentage of non-SOV travel incrementally increased from 22.89% to 23.23% on the basis of 5-year rolling averages. Partners developed 2- and 4-year targets of 25% to represent an increase in non-SOV travel in the region over the performance period. This target reflected a desire to improve non-SOV travel in the region. The partners noted, “Even if the region does not meet this target, the increasing rate of non-SOV travel over the past four years indicates that the region likely will make significant progress on this measure by matching or improving upon baseline results (23.23 percent) for this measure” (FHWA 2020). San Francisco–Oakland, CA, UZA and San Jose, CA, UZA: The non-SOV mode share targets for the San Francisco–Oakland UZA and the San Jose UZA were selected to align with the Plan Bay Area 2040 target. The UZA targets apply a consistent +1% target increase for 2020 and +2% target increase for 2022 to roughly align with the +10% mode shift target by 2040, which was a board-adopted target of the Metropolitan Transportation Commission (MTC) (MTC 2018).