Office of Operations
21st Century Operations Using 21st Century Technologies

Does Travel Time Reliability Matter? - Primer

What is Travel Time Reliability?

The U.S. Department of Transportation (USDOT) defines reliability as "the degree of certainty and predictability in travel times on the transportation system."

Defining Reliability

Travel time reliability is a measure of the consistency, timeliness, predictability, and dependability of a trip.

The term reliability has different meanings in different fields of study and even within the context of road transportation systems. For our purposes, however, this primer will adopt USDOT's definition of reliability, which defines it as "the degree of certainty and predictability in travel times on the transportation system. Reliable transportation systems offer some assurance of attaining a given destination within a reasonable range of an expected time. An unreliable transportation system is subject to unexpected delays, increasing costs for system users."(9)

Travel time reliability is a measure of the consistency or dependability in the travel time of a trip, or time to traverse a road segment, as experienced in different hours of the day and days of the week.(10) It is measured in terms of the additional time (i.e., time cushion or buffer) that drivers need to allocate to compensate for unexpected delays. Travel time reliability is an important measure for commuters, transit riders, shippers, and other road users because it allows them to make better decisions regarding the use of their time. For example, shippers and freight carriers require predictable travel times to deliver goods and services on time. The concern is not just that travel time is excessive due to rush-hour congestion (i.e., mobility being lower than the desired level) but also that travel time is unpredictable as a function of time or road segment.

Traditionally, travel time has been communicated in terms of historical averages calculated over periods of a year or longer (e.g., annual average peak hour travel time). However, road users experience variations in travel time on a daily or weekly basis due to traffic, roadway, and weather conditions. Travelers are more likely to remember the few days they suffered excessive delays versus their average travel times, as shown in figure 1.(10)

Figure 1
Figure 1. Average travel time and actual travel time experience.(10)

Variability in travel time can take two forms: (21) day-to-day and thin-day. Day-to-day travel time variation refers to changes in travel time when departing at the same time each day. Within-day travel time variability refers to changes in travel time between the same origin and destination at various times of the day. Typically, travel time is shorter during off-peak hours than during peak commuting hours, resulting in variation in travel time within the same day. Road users, passenger travelers, transit providers, and shippers are concerned with and interested in the time-of-day and day-to-day variation of travel time.

Reliability VS. Congestion

Congestion is a traffic condition characterized by slower speeds, longer travel times, and the occurrence of vehicle queues (wait lines). It can be recurrent (repeating often) or non-recurrent. Recurrent congestion includes delays that are predictable in frequency and extent (e.g., rush-hour traffic). Conversely, non-recurrent congestion is due to unexpected delays from temporary drops in road capacity (e.g., blocked lane due to a crash or work zone) or sudden surges in demand (e.g., planned special events).

Lack of reliability is different from congestion, though it is related to non-recurrent congestion. Reliability refers to the predictability of journey travel times. A highway prone to unexpected delays is unreliable. On the other hand, a highway that is typically congested and where traffic speed is consistently low can be reliable.

High levels of congestion, however, increase the likelihood of unreliability.(13) If roads are highly congested, a slight disturbance of traffic flow can result in excessive delays, have a greater impact, and take longer to recover than in a non-congested area. Therefore, lack of reliability in travel time is associated with delays caused by congestion, specifically with delays from non-recurrent events.

Congestion and reliability are so closely related that improvements in congestion can improve travel time reliability as well. For example, a bypass around Stockholm reduced delay and improved reliability. In fact, the monetary benefits of the reliability improvement were an additional 15 percent of the benefits of the delay reduction itself.(14)

How are Reliability and Congestion Related

Factors That Affect Reliability

Several factors can affect travel time reliability, namely the following sources of traffic congestion:(11), (12)

More than half of total congestion can be attributed to non-recurrent sources.
  1. bottleneck logo Bottlenecks refer to road segments that exhibit reduced traffic capacity when compared to the capacity of upstream road segments. Typical bottlenecks include lane drops, changes in alignment (e.g., horizontal curves), presence of merge and weave sections, changes in physical road characteristics (e.g., tunnels), freeway-to-freeway interchanges, hills, geometric changes, and access points to residential or commercial developments. Strategies that mitigate the effect of bottlenecks include demand management at upstream locations, efficient signage treatments, and provision of real- time traffic information and alternative routes for travelers.
  2. stoplight logo Traffic-control devices are used to inform, guide, and control the flow of traffic (both vehicles and pedestrians). Malfunctioning, ineffective, or inefficient traffic-control devices cause intermittent traffic flow disruption, which, in turn, causes delays and unreliable travel time. Common problems with traffic-control devices include use of improper devices; improper device placement; use of wrong color, shape, and size; poorly timed devices; poor maintenance; and device failure. Proper maintenance and use of traffic-control devices can significantly reduce delays from this type of congestion.
  3. cloud logo Weather includes adverse weather conditions such as heavy rain, fog, snow, and wind, as well as seasonal variations such as glare from the position of the sun. These weather conditions interfere with the visibility of traffic-control devices and lane delineations, which can lead to significantly reduced capacity and result in non-recurrent congestion and delays. Weather also affects driver behavior (e.g., leave longer following distances, slow down, and change lanes). Strategies to address this non-recurrent congestion include adjusting signal timing and using efficient road clearing practices to return roadways to full operating capacity both during and after adverse weather.
  4. orange cone logo Work zones are areas where roadway construction activities result in temporary physical changes to the highway environment. Factors that create delays in work zones include lane closures, lane width reductions, lane configurations, type and duration of work, work intensity, presence of police enforcement, pavement condition, and work zone length. Long-term work zones tend to have less effect on disrupting traffic compared to short-term work zones, since road users become familiar with the changed traffic pattern. Mitigation strategies for congestion due to work zones include reducing work zone duration, keeping lanes open, and reducing other incidents induced by work zones, such as crashes.
  5. car crash logo Traffic incidents are random events that disrupt the regular flow of traffic, causing a reduction in roadway capacity. These events include crashes, vehicle breakdowns, spilled loads, and debris. In addition to reduced road capacity due to lane blockage by the incident or the responders, drivers tend to slow down near an incident to observe it (i.e., rubbernecking), which further exacerbates delays due to the incident. Time delayed due to incidents depends on the number of lanes blocked, duration of the incident, and level of travel demand at the time of the incident. On average, during a multiple-lane incident, travel time increases by as much as 205 percent compared to traffic conditions with no incidents.(15) Minimizing the lanes blocked by the incident, decreasing clearance time, and shielding incidents from the view of drivers are some techniques that decrease the effect of non-recurrent congestion due to traffic incidents.
  6. two vehicles logo Travel demand fluctuations are the daily and seasonal variations in travel demand that result in increased travel compared to regular traffic. For example, seasonal variations could be due to holidays, part-year residents moving out or tourists coming in, or school-related traffic during the school year at the beginning and end of the school day. In some studies, an additional travel demand of more than 5 percent of the average traffic volume is considered as higher demand than regular traffic. Demand management strategies, such as diverting excess traffic to alternate routes and promoting use of public transportation facilities, can help alleviate congestion from increased demand.
  7. special events logo Special events can lead to sudden changes in travel demand. The resulting traffic condition will be significantly different from regular trends, causing unexpected delays. Non-recurrent congestion due to special events happens near arenas, conference centers, stadiums, and other public gathering places due to a sudden surge in demand during a short period of time (typically shortly before the event starts and after it finishes). Strategies to mitigate this type of congestion include diverting the non-event traffic, temporarily increasing capacity in the major direction of travel, and controlling on-off ramps to limit traffic entering or leaving a freeway.

The first two sources can be treated as recurrent congestion (i.e., repeating often) and the rest as non-recurrent congestion.

According to USDOT's 2013 Status of the Nation's Highways, Bridges, and Transit: Conditions & Performance, more than half of total congestion could be attributed to non-recurrent sources.(16) Similar statistics were also reported by the 2015 Urban Mobility Scorecard.(15)

Measuring Reliability

Spotlight Atlanta, GA Peak Period Unreliability.  Non-recurrent congestion (unreliability) 51.6%, Recurrent congestion - 48.4%

Data from 2008 in Atlanta, GA, showed that more non-recurrent congestion (unreliability) existed in peak periods than recurrent congestion. Primary causes included crashes, debris, and weather.(11)

Travel time reliability is a key indicator of the efficiency of a transportation system, and improving it is a prime objective for transportation agencies. The first steps toward improvement include measuring travel time reliability and quantifying it objectively to determine the current level of reliability.

Measurement Data

Travel time reliability studies require quality travel time data. Travel time reliability cannot be directly measured; rather, it is calculated from travel time data. Travel time is the time taken by a vehicle to traverse a route between two points of interest (e.g., the origin and destination). It is measured by noting the time a trip began and the time the destination was reached. It can also be measured by dividing the distance between the two points of interest by an average speed. Travel time includes running time, which is time spent while a vehicle is in motion, and stopped delay time due to traffic-control devices (e.g., red lights) and other operational delays (e.g., congestion).

Methods for measuring travel time include vehicles equipped with Global Positioning System (GPS) and cellular devices, license plate matching, aerial photogrammetry, vehicle-detecting devices, and road-user experience surveys. Travel time over longer periods of time is required to capture daily and seasonal variations and adequately calculate travel time reliability measures.(11) In combination with other basic traffic information, such as volume, travel time data can also be used to estimate other variables (e.g., total delays).

Some of the data-capturing techniques that can be used for calculating travel time reliability include the following:(18)

  • Field data collection: The collection of traffic volume, speed, occupancy, and vehicle classification data, which are used to determine various measures of travel time reliability for different vehicle classes. Recently, probe vehicle data collected from a fleet of vehicles equipped with special instruments (e.g., GPS and cellular phone) have been widely used to derive reliability measures. The advantage of traffic count–based data is that real traffic conditions can be observed without relying on road-user interviews. However, the identification of user characteristics, their behavior, and choices is difficult.
  • Simulations: Computer-based products that mimic vehicle and pedestrian movement and interaction with each other and traffic-control devices. Such models can simulate traffic at micro-, meso-, and macro-levels. Simulations are used to capture changes in the behavioral responses to differing travel and infrastructure options.

Each data type has its own advantages and limitations, which should be considered in the design of any reliability study. The data source has an impact on travel time reliability assessment.(19)

Probe Vehicle Data Collection

Map of Atlanta
Routes I-295 and NJTP for common trip origin and destination.(21)

Travel time data of two probe vehicles traveling along I-295 and the New Jersey Turnpike (NJTP) were collected for trips made on an hourly basis from 10:00 to 19:00 on three different Sundays: May 24, June 7, and July 19, in 2009.(20) The probe vehicles were dispatched simultaneously from a common trip origin point to a common destination. One used I-295 (64 miles), and the other used the NJTP (60 miles). Average travel time for route I-295 was 58.6 minutes with a standard deviation (or measure used to quantify variation) of 2.89 minutes, while the average travel time for route NJTP was 59.5 minutes with a standard deviation of 5.53 minutes. This shows that I-295 was a more reliable route than the NJTP during this time due to its smaller variation.

Reliability Metrics

Several metrics have been developed to measure travel time reliability, six of which are commonly discussed in literature.(11), (22)

Table 1. Six commonly used metrics to measure travel time reliability.
Metric Description Example
1. Travel time index Ratio of travel time observed during peak periods compared to free-flow travel time (or steady rate at which traffic traverses a freeway segment).
This metric indicates how much longer travel time is during congested conditions relative to light traffic. It is usually computed separately for the morning and afternoon peak hours on weekdays.
If travel time between two points of interest during afternoon peak hour averages 45 minutes, and travel time during free-flow conditions is 30 minutes, then the travel time index is 1.5.
Travel time index equals 55 minutes in peak hour divided by 30 minutes in free flow equals 1.5
2. Buffer index Additional time cushion that road users must budget to ensure an on-time arrival in 95 percent of trips, which is equivalent to being late for work once a month.
Buffer index is expressed in terms of a percentage of average travel time.
Assuming a buffer index of 50 percent and an average travel time of 20 minutes, then travelers must plan for an additional 10 minutes of time cushion to ensure on-time arrival in 95 percent of trips.
Buffer index equals allowance of 10 extra minutes divided average time of 20 minutes equals .5
3. Planning time index Total time a road user should allocate to arrive on time for 95 percent of trips. If the planning time index is 1.5, and the average travel time is 20 minutes, then the planning time is 30 minutes.
Planning index equals 30 minutes succeeds for all but 1 day a month divided by average time of 20 minutes equals 1.5
4. Failure and on-time measure The fraction of trips that are on time or late.
A successful, on-time arrival has a travel time no more than a pre-established threshold. The threshold can be the median travel time of a trip plus an additional 10 percent.
If the median travel time between a certain origin and destination is 30 minutes, and 18 of 100 trips between those points take longer than 33 minutes, then the failure rate is 18 percent.
Two formulas: 1) Failure rate equals Number of failed trips divided by number of all trips equals 18 divided by 100 which equals 18%.  2) On-time rate equals number of on-time trips divided by number of all trips equals 82 divided by 100 which equals 82%
5. Skew statistic Shape and size of travel time distribution as a measure of travel time reliability.
This measure relates a specified percentile travel time (e.g., 40th percentile below and above the median travel time), thus indicating the magnitude and direction of the skewness (or asymmetry) of the travel time distribution.
None
Misery index Ratio of excess travel time to average travel time.
It measures how much worse the longest trips are than normal trips. The misery index is the ratio of the average travel time of the 20 percent worst trips minus the average travel time of the all trips to the average travel time of all trips.(11) A modified formula for the misery index is the average of the worst 5 percent of travel times divided by free-flow travel time.
If the average travel time for a certain route is 30 minutes, but the average travel time for the worst 20 percent of trips over a week's time is 45 minutes, then the misery index 0.5.
Misery Index equals (Average travel time for worst 20% of trips minus Average travel time of all trips) all divided by average travel time of all trips equals (45 -30) divided by 30 equals .5

Figure 2 shows the relationship between the travel time index, buffer index, and planning time index.(10)

Travel Time Index, Buffer Index, and Planning Time Index
The difference between buffer index and planning time index is that the former represents the additional time cushion, whereas the latter represents the total travel time necessary for on-time arrivals for 95 percent of trips. Both travel time index and planning time index have similar numeric scales. However, travel time index is for peak hours, while planning time index is for any time of day.

Graph of the relationship between travel time index, buffer index, and planning time index
Figure 2. Relationship between travel time index, buffer index, and planning time index.(10)
2015 Estimates - 10.8 billion tons, $11.6 trillion; 2024 Projection - 14.2 billion tons, $18.7 trillion
Total freight transported by road in the United States.(25)

FHWA recently issued a final rule with two new travel time reliability metrics as the Moving Ahead for Progress in the 21st Century (MAP-21) performance measures.(23), (24) The objective of MAP-21 performance measures is to assess the performance of the National Highway System (NHS), freight movement on the interstate system, and the congestion mitigation and air quality improvement program. The two new reliability metrics are:

  • Level of travel time reliability (LOTTR): The ratio of the 80th percentile travel time to the normal travel time (i.e., the 50th percentile occurring throughout a full calendar year) using data from FHWA's National Performance Management Research Data Set (NPMRDS). NPMRDS includes travel time data on the NHS, and LOTTR is used to assess the performance of the NHS.
  • Truck travel time reliability (TTTR): The ratio of the 95th percentile travel time of trucks to the normal truck travel time, (i.e., the 50th percentile using a full calendar year of truck travel time data). TTTR assesses the freight movement on the interstate system and is reported for five time periods depending on time of day and day of the week.

Various agencies use different metrics of travel time reliability as a portion of their mobility performance assessment depending on their needs. Table 2 shows the reliability metrics used by several agencies.

Table 2. Reliability metrics used by various transportation agencies.(19)
Agency Reliability Metrics Used
Georgia Regional Transportation Authority and Georgia Department of Transportation (GDOT) Buffer index and planning time index
Florida Department of Transportation (FDOT) Buffer index and on-time arrival
Southern California Association of Governments (SCAG) Buffer index
Washington State Department of Transportation (WSDOT) 95th percentile travel time
Maryland State Highway Administration (MDSHA) Travel time index and planning time index

Benefit-Cost: Value of Travel Time and Value of Reliability

The value of travel time (VOT) and the value of reliability (VOR) are key factors in benefit-cost studies.

VOT =
monetary value of reducing travel time

VOR =
monetary value of reducing variability of travel time

VOT is the monetary value that road users place on reducing their travel time. VOR is the monetary value that road users place on reducing the variability of their travel time. The concepts of VOT and VOR involve estimating the magnitude of the benefits gained by saving in travel time and reduction in travel time variability or the penalties incurred by longer travel times and greater travel time variability.(26)

Value of Reliability
Map of Western US

When moving goods from one end of I-5 to the other, a commercial vehicle operator has to add about 6 hours of buffer time to ensure on-time delivery with 95 percent confidence.(10)

Recently, the valuation of travel time reliability in a benefit-cost analysis of transportation projects has been gaining attention. Some U.S. and international research projects have been initiated to provide guidance on how an agency can include the value of reliability (VOR) in a benefit-cost analysis when making congestion reduction-related project investment decisions.(2733)

Most benefit-cost studies of mobility solutions require the monetary value of travel time (VOT) and the monetary VOR. Understanding VOT and VOR can inform decisionmakers about the value of potential strategies for improving reliability, such as adding tolled roads or high-occupancy lanes or reducing incident clearance times.

In the trucking industry, shippers and carriers value travel time at $25 to $200 per hour (depending on the product being carried).(2) A recent study in the area of freight transportation found that the VOT ranged from $12.80 to $283 per shipment hour, and the average value was $37 per shipment hour. VOR ranged from $51 to $290 per shipment hour, and the average of the distribution of VOR was $55 per shipment hour.(34, 35) This indicates that freight shippers valued travel time reliability 1.5 times as much as travel time savings.

VOR to travelers can be estimated using data from congestion pricing strategies. This includes strategies like tolled roads and managed lanes (e.g., high-occupancy lanes). Such strategies offer reduced travel time and increased reliability to users who are willing to pay. Based on an analysis of congestion pricing data, many studies have reported that drivers value increased travel time reliability more than travel time savings.(3638) Increased travel time reliability accounted for 68 percent of the benefits of using congestion priced lanes.(38) Another study estimated that users of congestion priced lanes value travel time to be $11.63 per hour, while they value reliability to be $25.45 per hour.(36)

This monetary cost associated with travelers and shipments being late or unreliable arrival time should lead to prioritizing travel time reliability in transportation planning and management. And, the significance of travel time reliability should be taken into account in the appraisal and valuation of infrastructure development projects.

VOR has increased in recent years. More reliable transportation systems are needed to address changes in business, industrial, and personal travel patterns and the emergence of more complex and interrelated scheduled activities (e.g., the just-in-time freight distribution and arrival systems). In 2015, the total weight of freight transported by road in the United States was estimated to be more than 10.8 billion tons and a value of $11.6 trillion.(25) In 2024, the freight figures are projected to increase to a weight of 14.2 billion tons and a value of $18.7 trillion.

To support on-time delivery of freight, a reliable transportation network is critical. For example, for frequent truck drivers who are familiar with the local traffic patterns of a region, information on travel time reliability was found to be a more significant factor in route selection than was historical travel time information.(39) Expanding national and international trade creates greater goods volume, further destinations, and increasingly complex and interdependent supply chains. Successful interdependent and interconnected systems require reliable travel times between origins and destinations.

Methods for Measuring the Value of Travel Time Reliability

Attitudinal survey: Document commuters' perceptions about the relative importance of travel time reliability.

Stated preference survey: Ask respondents how much they value travel time reliability compared to other hypothetical scenarios. (This method represents bias due to respondents not replicating their stated choices as observed in real situations.)

Revealed preference study: Capture behavioral responses concerning the actual choices made by participants to travel time reliability measures. (The advantage of this technique is the reliance on actual choices, e.g., the technique avoids problems associated with stated preferences or a failure to properly consider behavioral constraints.)

Changing patterns of individuals' employment opportunities, increased income, as well as increased recreational choices have affected passenger miles traveled.(40) Accommodating the busy calendar of an individual demands a more reliable transportation system so that delays in one activity do not affect the next activity. Essentially, road users value travel time reliability for several key reasons:

  • Unreliable travel time forces road users to plan for extra time to avoid late arrivals. In economic analysis, the value of the extra time road users plan to avoid late arrivals (VOR) is generally more than the average VOT.
  • Given that the VOR is different from VOT, transportation planners have to incorporate the costs of unreliable travel in project planning, programming, and selection processes.
  • Reliability becomes an additional factor that influences where, when, and how road users travel.(41) Road users demand and expect a reliable transportation system. For example, given recent advances in communication technologies, the traveling public expects agencies to make traffic data available from a variety of devices so that they can see real-time traffic conditions. They also expect that agencies address traffic disruptions in a timely and efficient manner regardless of who owns the roads.

Travel time reliability is a key factor to the success of the U.S. economy and those who participate in it, including businesses, the traveling public, and local governments. Next, we'll dig a little deeper into the impacts of an unreliable transportation system, which includes negative effects on businesses, the traveling public, and local governments as well as impaired safety and security.

boat and airplane logo

Importance of Reliability Across Transportation Modes

On-time arrival of people and freight to their destinations is an important factor in all modes of transportation. Complex global and intermodal logistics chains require minimal disruption to ensure on-time delivery, which makes reliability critical. A recent survey of ocean shippers revealed that reliability of waterborne freight is considered to be the most important factor in maritime transportation compared to cost, speed, safety, security, and trackability.(42)

Similarly, in air transportation, on-time departures and arrivals are key performance indicators for airline service quality, which determines customer satisfaction and loyalty, market competitiveness, economic benefits, and profitability.(43) Reliability plays a key role in meeting these performance standards.

Office of Operations