Peer Exchange Workshop on the "Perfect World of Measuring Congestion"
Workshop Summary Report
APPENDIX C: WHITE PAPERS
Paper #4: Operations Performance Management: How Should External Events and Trends be Considered?
Prepared for:
Peer Exchange Workshop on the “Perfect World of Measuring Congestion”
FHWA Office of Operations
Prepared by:
Texas A&M Transportation Institute and Battelle
FINAL
December 10, 2013
1. Introduction
There are many independent variables (or external factors) for which state DOTs, metropolitan planning organizations and USDOT have little control over. These variables may affect performance measures or trends. Therefore, when using performance measures for reporting and/or decision-making, these variables must either be incorporated or their impact recognized in the performance of the transportation system and in the results of project evaluations.
For many of these variables, the relative change may be more important than the actual value. For example, in an operations context, it may be more important to the evaluators to know the change in employment in a region and how that might have affected operations rather than knowing the base year number of workers. Adjustments in calculation procedures or communication can be made to incorporate some of these factors, but others can only be recognized and their possible effect communicated.
This paper identifies the key external influences that may lead to fluctuations in performance measure values. It also includes proposed adjustments to alleviate these fluctuations. These variables will be discussed in more detail at the workshop. When developing the proposed adjustments, the following elements were considered: Can we generally accommodate the variable? Can we connect it to congestion and reliability measures? Can we use the information to explain why the measure is changing? Can we use the measures to make investments, policy changes or practice adjustments that will reduce congestion? What needs to happen to achieve this level of analysis?
This paper is designed to help achieve the following two workshop objectives:
- Design of a “platinum standard” monitoring program; and
- Development of a practical FHWA research roadmap by the end of the workshop.
In the text below, a recommendation accompanies each variable to describe how the phenomenon or issue should be accommodated.
- Explanatory information – describes a variable that has an effect on transportation system performance, but cannot be included in the analytical construct of the measure.
- Modifying factors – includes variables that can be accommodated with a change in the measure or the calculation procedures
2. The Big Picture Variables
The independent variables included in this document may have an effect on performance measures. Specifically, the variables may impact the changes in performance measurement values between reporting periods. While these variables may not be reported in every case, they should be in the menu of questions that are asked. There may also be some contexts where the variables are more important: regional vs. corridor analyses, monthly vs. annual summaries, operating decisions vs. summary reports.
Economy –The 2008 recession once again illustrated the important role that non-transportation actions or occurrences play in congestion levels. Local recessions have caused drops in the congestion level in the past – see the big California cities in 2000 with the dot.com bubble burst. Measures such as those below can be included as a regular component of operations reports to explain the role of changes in the economy – both good and bad:
- Gross metropolitan product – the local version of gross domestic product
- Total salary and wages – essentially the job-related portion of GMP
- Employment - number of jobs
- Unemployment rate
- Population
- Gas and diesel fuel prices
Recommendation: Include an economy measure as an overarching system explanation unless there is significant sub-regional variation in employment changes (e.g., large assembly plant or corporation ceases operations).
Societal – Decisions made to improve the quality of life or enhance economic opportunity often have an effect on system operations. These may be regional, corridor or neighborhood level changes and may be accompanied by changes in operation. As operating policies, practices and technologies are deployed, it is also important to capture other ‘outside the right-of-way’ events or phenomena that affect operations. This need not be a comprehensive investigation of the urban condition, and it may be reported only occasionally or displayed as an appendix/additional information element; some examples include the following.
- Development Patterns – Density, magnitude and mix of land uses would be typical descriptors. Part of most long range metro plans is to concentrate more population and jobs into dense neighborhoods and centers and to move more people by transit and carpools. These changes are typically estimated to support more economic development and person travel for a given level of congestion. Vehicle ownership rates and vehicle use might be very useful to explain what is causing changes in several types of performance measures.
- Housing Cost – Separate from, but related to, development patterns is the issue of housing cost. This could be ownership or rental cost – the key element is to capture the changes in cost that might tend to shift commuting and other travel patterns.
- School Quality – Changes in school quality may cause changes in residential and job patterns that might result in vehicle travel changes. Much of the effect of this variable might be illustrated by changes in vehicle-miles of travel (see below) but the causation might be important. Suburban real estate agents reportedly (1) identify school quality as a cause of home purchases, explaining some of the ‘suburban sprawl’ that occurs despite long and unreliable commute times.
- Generational Differences in Travel – Travel data from the first decade of the 21st Century appear to indicate changes in the travel patterns of the younger end of the workforce. While some of this change is related to differences in employment numbers and rates, there also appears to be increased use of electronic means (such as telework or teleshop) for making trips and greater use of shared ride services among younger travelers. Additional descriptive information and alterations to trend calculation methods may be required if the ‘millennial travel shift’ is an enduring part of transportation.
- Connected Travel – In placing a value on extra travel time due to congestion or unreliability, the role of travelers being wirelessly connected should be considered. The value adjustment(s) may be addressed in other research, but the penalty to a traveler, especially one on public transit or in a shared-ride, may not be as significant as in the past. The fact that travelers can work during this time may allow them to accept worse travel conditions, with no decrease in quality of life or economic condition.
Recommendation: Examine the potential role of these variables annually and include an explanation of the possible role in changing congestion or reliability levels and in altering the trends or values of the key performance measures.
3. Specific Variables
These variables might have a direct effect on the travel speed dataset. An operator typically has relatively little ability to affect these but performance measures can be significantly affected by them.
Changes in Travel Demand – More or fewer travelers and/or vehicles can alter the context of several measures. In addition to the broad measure of total travel, the following sources might also be considered to better understand the effect that changes in vehicle-miles or person-miles of travel may have on congestion and reliability:
- Latent demand is an associated issue; operations treatments can improve travel conditions and draw traffic volume from other routes, causing the improvement to appear to be less significant than if demand were constant.
- Effect of TDM programs which may change the vehicle volume, trip departure times or travel route; or trips that were not made due to electronic substitution
- Mode Share – If an agency is trying to get transit-oriented development and mode shift, mode share should be measured.
Recommendation: Examine the primary route and any routes that might be affected – typically parallel routes but if signal timing will be revised, significant crossing roadways should also be examined.
Person volume – If a particular approach or strategy will cause shifts to buses and carpools, person volume should be accommodated in the measure.
Recommendation: Include an estimate of person volume in all reports; collect data on person-volume to support estimates when usage changes (e.g., high-occupancy vehicle lanes).
Parallel routes – Major streets or other freeways should be included in the analysis if they will be substantially affected by an improvement. If the effect is not substantial they should be included in an explanation.
Recommendation: Examine congestion and reliability of any routes that might be affected by a corridor improvement.
Weather, Incidents and Road Work – The location and timing of events that affect system performance (including incident clearance data) should be linked with congestion data to improve performance management. The base case should include descriptive information as explanatory variables.
Recommendation: Initially, include explanatory variables of these events in any report -- this may be as simple as number of rain days or incidents. As the datasets for these elements are integrated, there may be an ability to parse the sources of congestion using a set of allocation rules.
Connected vehicles or autonomous vehicles – At some point these will begin to affect operating performance of the system and therefore the measures and the data.
Recommendation: Develop a description of the presence of infrastructure and vehicle attributes to characterize the reason for changes in operations performance measures.
Modal Accessibility – The ease of use of transit, walking and bike modes is an important aspect of describing the role of single occupancy vehicle (SOV) alternatives. Ease of transit use, etc.
Recommendation: An explanatory measure such as number or percent of urban residents with access to a nearby transit stop may be a useful initial measure. Other elements such as sidewalks, bike lanes and paths, bikeshare, car sharing should also be described if they begin to play significant role in the region or corridor.
4. Aggregation Level
The more disaggregate the measures are, the more likely they are to capture effects of improvements. But they cannot be too disaggregate or too narrow or the “spillover effects” may be missed. For example, freeway improvements could improve freeway speeds and throughput, which could draw traffic from a parallel arterial, thereby also improving the arterial operations.
Recommendation: There are a number of guidance documents that can aid performance measurement professionals in developing a consistent dataset. The specific actions depend on the uses for the measures and users should expect that some changes will be necessary as measurement, management and investment decisions evolve.
5. Customer Expectations and Target Setting
Comparisons of urban, suburban and rural congestion are often a part of regional congestion discussions. At the statewide level, similar comparisons are made between large and small urban area congestion levels. These often are generated by differences in expectations and, to date, all the technical ‘solutions’ have been found wanting in the public discussion of how the comparisons should be reflected in project or program investments.
Recommendation: The use of the data and measures will be key in the decision about this element. At a broad regional scale, the long range planning process should be used to investigate public opinion on expectations; the level of ‘unacceptable congestion’ can be used to develop performance measure targets. For system reporting – such as a freeway operations report, a good first step may be choose some target level (for example, the speed at which the maximum volume occurs) and explain the use of that target. It appears that a map of acceptable congestion levels will be needed to identify where system improvements are needed, and where the community has decided to not aggressively attack congestion in lieu of other attributes.
REFERENCES
- Nonlinear Effects of School Quality on House Prices. Abbigail J. Chiodo, Rubén Hernández-Murillo, and Michael T. Owyang. Federal Reserve Bank of St. Louis Review, May/June 2010, 92(3), pp. 185-204.
