1  

  

HOT LANE POLICIES AND THEIR IMPLICATIONS 1 

2 

By 3  4 

Rahul Goel 5 

PhD Scholar, 6 

Indian Institute of Technology, 7 

New Delhi, India 8 

Email: rahulatiitd@gmail.com 9 

10 

And 11 

12 

Mark W. Burris, Ph.D., P.E. 13 

E.B. Snead I Associate Professor 14 

Department of Civil Engineering 15 

Texas A&M University 16 

CE/TTI Building Room 301G 17 

3136 TAMU 18 

College Station, TX 77843-3136 19 

Corresponding author: mburris@tamu.edu 20 

Ph: 979-845-9875 21 

Fax: 979-845-6004 22 

23 

24 

Paper Submitted for Publication in the Springer Journal: 25 

Transportation 26 

 27 

28 

 29 

 30 

   31 

2  

  

 1 

Abstract 2 

This research examined the major changes in a corridor due to high occupancy/toll (HOT) lane 3 

implementation. This was accomplished by comparing the impacts of HOT lanes on three pairs 4 

of HOT lanes with similar design and operational characteristics. These pairwise comparisons 5 

of similar HOT lanes reduced the impact of exogenous factors and removed the issue of 6 

comparing HOT lanes that were so dissimilar it would be impossible to isolate the reasons for 7 

difference in results from the lanes. 8 

With strict registration requirements for free high occupancy vehicle (HOV) 3+ travel on the I-95 9 

Express Lanes (ELs) in Miami there were indications that some carpoolers switched to lower 10 

occupancy modes. Tolled access for HOV2s on I-95 and the SR-91 ELs near Los Angeles 11 

resulted in lower usage of those ELs by the HOV2s as compared to most HOV lanes where 12 

HOV2 access is free. 13 

On the SR167 (Seattle) and I-25 (Denver) HOT lanes, exogenous factors like the price of gas 14 

and the economic recession seemed to be the primary influence on the usage of those HOT 15 

lanes. In both cases, carpool usage increased along with the price of gas. On I-25, the 16 

increasing unemployment rate coincided with a decrease in toll paying travelers. On SR 167 17 

there were also indications of mode shifts among the transit, carpool and toll paying SOVs due 18 

to the fluctuating price of gas. 19 

20 

Key Words: Managed Lanes, Policy, Express Lanes, HOT Lanes. 21 

22 

3  

  

1.0 High Occupancy/Toll (HOT) Lanes Introduction 1 

HOT lanes are gaining interest throughout the United States as a strategy for meeting multiple 2 

performance objectives in congested urban freeway corridors. Over the past 15 years the 3 

United States has witnessed ten High-Occupancy Vehicle (HOV) lanes being converted to HOT 4 

lanes. HOT lanes allow lower occupancy vehicles (generally, single occupant vehicles) to 5 

access the carpool lanes by paying a fee while higher occupant vehicles continue to access the 6 

lanes for free. In December 1995, the nation’s first HOT lane opened. This project was the 91 7 

Express Lanes (ELs – note that many of these HOT lanes are called Express Lanes), where 8 

carpools with three or more passengers could use the lanes for free (except during the period 9 

from 1998 to 2003 when HOV3+ travelers were required to pay half the SOV toll). This project 10 

was followed by nine HOT lanes converted from HOV lanes across the United States. These 11 

projects included (with year of conversion in parentheses) I-15 Express Lanes in San Diego, 12 

California (1996), Katy (1998) and Northwest (1999) Freeways in Houston, Texas, I-394 (2005) 13 

and I-35W (2009), in Minneapolis, Minnesota, I-25 in Denver, Colorado (2006), I-15 in Salt Lake 14 

City, Utah (2006), and more recently, SR-167 in Seattle, Washington (2008), and I-95 in Miami, 15 

Florida (2008). As of September 2009 there were ten fully operational HOT lanes. 16 

17 

HOT lanes can provide benefits in reducing travel time, offering travelers viable options to 18 

congestion, improving freeway efficiency, and raising revenue to offset implementation and 19 

operating costs. However, the extent to which these benefits can be attained depends on the 20 

usage of those lanes. 21 

22 

With ten fully operational HOT lanes already in place, there is an opportunity to learn how 23 

different factors and policies impact the mode choice of travelers and ultimately the usage of 24 

HOT lanes. Additionally, the ways in which different HOT lane policies affect HOT lane 25 

operations also provides guidance for future HOT lane projects. This research examines six of 26 

4  

  

the ten HOT lanes. The most recent I-35W HOT lanes in Minneapolis were excluded due to 1 

lack of data. The two Houston HOT lanes and Utah’s HOT lane were excluded because of their 2 

significant differences from other lanes. For example, US290 in Houston does not allow SOVs 3 

and charges HOV2s a flat $2 fee during the morning peak period. These characteristics make it 4 

difficult to compare US290 to other HOT lanes. 5 

6 

The initial research effort focused on examining and comparing all six of these HOT lanes. 7 

However, it was clear that the differences between the lanes were just too great to draw many 8 

solid conclusions when comparing how different aspects of all the lanes impacted their 9 

performance. However, there are three pairs of similar HOT lanes and comparing these 10 

facilities in a pairwise fashion lead to interesting conclusions. 11 

12 

2.0 Comparison of I-95 and SR-91 Express Lanes 13 

2.1 Description of the I-95 Express Lanes, Miami 14 

The 95 Express Lanes are located on I-95 near Miami and extend from SR‐112 to the Golden 15 

Glades Interchange (GGI) area just north of NW 151st Street. The implementation of these ELs 16 

included multiple design-and- policy changes, with one of the most significant being the change 17 

from a single HOV lane per direction to two express lanes per direction. . The project created 18 

an additional express lane in each direction by narrowing the inside median width and reducing 19 

travel lanes from 12 feet to 11 feet. Additionally, the existing HOV lane buffer was reduced to 20 

one foot of separation between the general purpose lanes (GPLs) and the ELs and flexible 21 

delineators have been installed within the buffer area. The resulting cross section allows for two 22 

ELs , a one foot buffer (and no median shoulder), and four GPLs in both the northbound and 23 

southbound direction (FDOT, 2009).Note that this research only examined Northbound (Phase 24 

1A) travel as the Southbound ELs opened only recently in January 2010. Compared to the 25 

5  

  

previous HOV lanes which could be accessed from anywhere, the ELs have no intermediate 1 

access points and can be accessed only at the two ends of the lanes (Florida DOT, 2009).The 2 

project also enhances and expands Bus Rapid Transit service. The EL operation also included 3 

increasing the carpool eligibility to HOV3+ from HOV2+ in the previous HOV lanes. Also, 4 

carpools and vanpools have to be registered to be able to use the HOT lanes for free. The 5 

eligibility criteria for a registered 3+ carpool includes: a) participants must live within a 3-mile 6 

radius of each other b) participants must work within a one mile radius of each other and c) 7 

participants must have a start/end work time within 30 minutes of each other (95 Express 8 

website). 9 

10 

2.2 Impact on Carpooling on I-95 11 

In the year following the opening of the ELs there was a 4.6 percent increase in the person 12 

throughput on the whole corridor. This is similar to the growth in traffic from 2006 to 2008 (Cain, 13 

2009). This likely indicates that the 256 percent increase in SOVs in the ELs was mostly due to 14 

the mode shift within the corridor and not due to an overall increase in travelers (see Table 1). 15 

The overall decrease in the number of HOV2s likely indicates that these carpools either shifted 16 

mode to SOV (an overall 33 percent increase in SOVs) or they shifted to higher occupancy 17 

modes (overall 9.6 percent increase in HOV3s). Overall, this resulted in a decrease in 18 

carpool/transit mode share from 50.8 percent to 48.6 percent of travelers. 19 

20 

To further investigate these mode shifts, data from a May 2009 survey was analyzed (see Table 21 

2). Over 200,000 travelers in the Miami-Dade area were contacting regarding the survey and 22 

9156 travelers participated in the survey. Table 2 contains only responses from those who 23 

indicated they used the ELs. For this analysis, the ‘usual mode’ of respondents was defined by 24 

6  

  

the survey question: “How do you usually get to and from work/school? Select the mode you 1 

use most often (3 or more days per week).” 2 

Table 1: Person Throughput by Vehicle Type in I-95 Express Lanes (Northbound; PM 3 Peak Period- 4 to 6 PM) 4 

Vehicle Type Express Lanes Total Facility (GPLs + ELs) Total Person Volume per Peak

Period Total Person Volume per Peak

Period 2008 2009 %

Increase 2008 2009 %

Increased SOV 1061 3778 256.1% 9141 12206 33.5% HOV2 3040 1899 -37.5% 10437 8181 -21.6% HOV3 477 171 -64.2% 2335 2558 9.6% Transit 810 821 1.4% 810 821 1.4% Total 5387 6669 23.8% 22723 23766 4.6%

Source: Cain, 2009. Note: All transit buses use the ELs. 5 

6 

Table 2: Mode Shifts on I-95 by EL Users 7 

Usual Mode of Travel (Number of respondents)

SOV (n=2558)

HOV2 (n=158)

HOV3+ (n=120)

Current Mode

used to Access Express Lanes

Toll Paying (SOV/HOV2/Unregistered HOV3+)

94% 92% 61%

Registered HOV3+ 0% 3% 33%

Hybrid 5% 6% 1%

Vanpool 0% 0% 16%

Prior Mode

on I-95

Carpool in HOV 6% 71% 81%

Hybrid in HOV 4% 4% 2%

Not Carpool lanes 86% 24% 18%

Note: The percentages for current mode do not total 100 percent since the respondents could 8 select more than one mode if they frequently used different modes to access the ELs. 9  10 

Prior to implementation of the ELs, HOV2s had unrestricted, toll-free access to the HOV lane 11 

and nearly a quarter of them still did not use the lanes (see Table 2). This may have been due 12 

to the relatively congested conditions on the lane at that time. With the ELs, HOVs now have to 13 

pay a toll and can only enter the lanes at their beginning – but congestion was greatly reduced. 14 

7  

  

Since almost all HOV2s (92 percent) now use the Express Lanes as paying users it is a clear 1 

indication of the attractiveness of the improved travel conditions on the Express Lanes. 2 

3 

For those who mentioned HOV3+ as their current usual mode, 81 percent of them previously 4 

used the HOV lanes as carpools. However, with the ELs, only 33 percent are registered HOV3+ 5 

users while 61 percent are toll paying SOV/HOV2/unregistered HOV3+. These data, combined 6 

with the substaintial decrease in HOV3+s in the ELs during the peak period (Table 1), indicates 7 

a substantial amount of those travelers have shifted to either SOV or HOV2. This shift of the 8 

HOV3+ to toll paying vehicles can likely be attributed to the strict registration requirement for the 9 

carpool registration. 10 

8  

  

Table 3. Characteristics of Different HOT lanes (Pair-Wise Comparison) 1 

HOT Lane Number of Lanes

Separation Intermediate Access points

Free Travel

Eligibility

Impacts

SR 91X, Los Angeles

2 HOT lanes and 4 GPLs per direction

Painter buffer and pylons

None HOV3+ -HOV2+ traffic volumes in the corridor remained relatively stable - HOV3+ traffic volumes increased by 40 percent in PM peak after the Express Lanes started (but this is starting from a small baseline). -40 percent of the survey respondents mentioned driving comfort and safety in the Express Lanes as the reasons to use the Express lanes other than travel time savings.

I-95, Miami 2 HOT lanes and 4 GPLs per direction

Double White Stripes with breakable poles

None HOV3+ registered carpools

-For those who mentioned HOV3+ as their usual mode, 81 percent of them previously used the HOV lanes as carpools. However, with the opening of the Express Lanes, only 33 percent are registered HOV3+ users while 61 percent are toll paying LOVs (SOV/HOV2). - Daily transit ridership increased 30% due to travel time improvement onthe Express Lanes. This was despite a bus fare increase, a decrease in gas prices, reduced service and an economic recession.

I-15, San Diego (until May 2008)

2 Reversible HOT Lanes and 4 GPLs per direction

Concrete Barrier

None HOV2+ - Total weekday usage of Express Lanes increased from 7685 (pre-Express Lanes) to 15,000 in three years after the Express lanes started — Congestion on the GPLs was alleviated by directing the increasing traffic on corridor to HOT lanes

I-25, Denver 2 Reversible HOT Lanes and 4 GPLs per direction

Concrete Barrier

None HOV2+ -Express lane users indicated more use of transit due to the Express Lanes. - Little change in carpooling after the lanes opened - Large increase in transit ridership seen with an increase in gas price (2008) and vice versa (2009).

9  

  

I-394, Minneapolis (Diamond Section)

1 HOT Lane and 2 GPLs (most of the time) per direction

Double White Stripes

6 in each direction

HOV2+ — Bus ridership increased after the HOT lanes, significantly higher than in the control corridor. — User surveys indicated no negative impact on carpooling due to the Express Lanes. — Carpoolers decreased in peak periods. Attributed to the increase in operational hours

SR 167, Seattle

1 HOT Lane and 2 GPLs per direction

Double White Stripes

6 Northbound and 4 Southbound

HOV2+ -Carpool usage decreased by more than 25 percent in one year with an increase in bus ridership. Increasing gas prices seem to be the primary reason. - A 42% increase in carpooling. More carpools than on I-394. Both possibly due to greater GPL congestion.

1 

10  

  

2.3 Reasons for Using the I-95 Express Lanes 1 

Respondents who have used the ELs were asked their reasons for using the lanes. The 2 

respondents were grouped by the modes they used to access ELs: carpool, toll paying users 3 

(SOVs/HOV2/unregistered HOV3+) and Hybrids. More than three-quarters of the respondents 4 

in each group indicated the time-savings/free-flowing traffic on the ELs as the reasons for using 5 

them. For example, prior to the ELs the 7.5 mile trip from downtown Miami to the Golden Glades 6 

Interchange took 25 minutes on the HOV lane during the peak period. Now, on the ELs, this trip 7 

only takes 8 minutes. Apart from these reasons, the absence of trucks and a perception of 8 

safety in the lanes were also noted as important reasons for using the ELs. 9 

10 

11 

2.4 Travelers not Using the I-95 Express Lanes 12 

Among the respondents who mentioned they used the HOV lanes previously but no longer use 13 

the ELs, almost three-quarters of them carpooled (occupancy not known) on the HOV lanes 14 

(while the rest of them used hybrid vehicles or motorcycles). When carpoolers were asked their 15 

reasons for not using the ELs, almost 10 percent mentioned not qualifying as a 3+ carpool, 16 

more than 50 percent mentioned not willing to pay a toll (which would also indicate not 17 

qualifying as a 3+ carpool), while more than one-third mentioned no access to get in to or out of 18 

the ELs. Therefore, among the changes made in the design and operations of the ELs, access 19 

and toll were the two major reasons for carpoolers to not use the ELs. 20 

21 

2.5 Similarities between SR 91 Express Lanes and I-95 Express Lanes 22 

There are many similarities between the SR 91 Express Lanes (December 1995 through 23 

January 1998, the period when HOV3+ could access the ELs for free) and I-95 Express Lanes. 24 

Similarities were found in both operations as well as design: 25 

11  

  

two ELs and four GPLs in each direction, 1 

HOV3+ requirement for free carpool access to the ELs, 2 

no intermediate access to the ELs, and 3 

flexible pylons separating the GPLs and the ELs with no median shoulder (see Table 3). 4 

Both corridors had high congestion levels before the ELs. Around the time of Express lane 5 

implementation, average daily traffic (ADT) on the two corridors was 200,000 vehicles per day 6 

(vpd) on SR 91 and 290,000 vpd on I-95. Major differences were that there was no HOV lane on 7 

the SR 91 corridor and the SR 91 express bus route 149 did not use the ELs. Despite these two 8 

differences, the many similarities between the two ELs provide an opportunity to compare the 9 

different impacts that EL implementation had in the respective corridors. As the SR 91 Express 10 

Lanes were built by a private company one of their key objectives was profit. Thus HOV2s had 11 

to pay a toll. On the I-95 Express Lanes the objective was to encourage ridesharing (but 12 

excluding fampools) and transit. This resulted in the requirement that HOV2s pay and HOV 3+ 13 

had to register. The objectives were different but the resulting policies were similar. 14 

15 

2.6 Impact on HOV3+ Travelers 16 

After one year of operation of the SR 91 Express Lanes, there was a 40 percent increase in 17 

HOV3+ vehicles in the 2-hour peak period (496 to 725) on the SR 91 corridor (Sullivan, 1998) 18 

while there was a 10 percent increase (778 to 852) of HOV3+ vehicles on the I-95 corridor 19 

(Cain, 2009). However, the large difference in the percentage increase may be largely due to 20 

the low number of baseline HOV3+ vehicles on the SR 91 corridor since, on SR 91, there were 21 

previously no HOV lanes and less incentive for carpooling (Sullivan, 1998). Also, the increase in 22 

HOV3+ on SR 91 can be attributed primarily to the implementation of ELs providing much 23 

needed additional capacity. In case of I-95 there was a 64 percent reduction in HOV3+ vehicles 24 

in the ELs and an increase in HOV3+ vehicles in the GPLs. 25 

12  

  

Next, the usage of the ELs by the toll exempt vehicles in the two corridors after one year of 1 

operation (December 1996 for SR 91 and December 2009 for I-95) were compared. One way, 2 

weekday volume of HOV3+ on SR 91 Express Lanes (2500 vehicles) was approximately equal 3 

the total toll exempt trips (including Hybrids, Buses, motorcycles) in one week (approximately 4 

2000 trips) on the I-95 Express Lanes. 5 

6 

There are different factors which could be responsible for the different HOV3+ usage trends on 7 

the two corridors: 8 

Induced traffic: Over the first year of EL operations on SR 91, there was an increase of 9 

14 percent (28000 vehicles) in the ADT of the corridor. Sixty percent of that increase 10 

was new traffic (traffic above the long term growth trend and not from parallel streets). 11 

According to one estimate (Sullivan, 1998), during the PM peak period, this induced 12 

traffic contributed to a 39 percent increase in the pre-Express level of HOV3+. 13 

Conversely, over the first six months of EL operations on I-95, northbound traffic on I-95 14 

(ELs and GPLs) decreased by almost 3 percent. 15 

New incentives for carpooling: Unlike on the I-95 corridor, there were no HOV lanes on 16 

the SR 91 corridor. With ELs, SR 91 travelers had an additional incentive for forming a 17 

3+ person carpool. This may explain the high percentage increase in HOV3+ volume on 18 

the SR 91 corridor. 19 

Registration Requirement: On the SR 91 Express Lanes, registration is required by all 20 

users (toll and non-toll) with no extra eligibility requirements for carpool members other 21 

than the number of people in the vehicle. However, on the I-95 Express Lanes, there are 22 

strict requirements that must be met to register as a carpool. The effect of strict 23 

guidelines on carpooling can be seen in the results from the I-95 survey. Sixty-six 24 

13  

  

percent of those whose typical mode is HOV3+ mentioned using the ELs as toll users 1 

while before the ELs more than 80 percent of them used the HOV lanes as carpools. 2 

EL availability: SR 91 Express Lanes started operating in both directions from opening 3 

day. The I-95 Express Lanes began only in one direction (northbound, going away from 4 

downtown Miami). In the southbound direction, the HOV lanes were still congested 5 

during peak periods. Therefore, for the two-way commuters, that might not have been 6 

enough incentive to register as a carpool and use the lanes or switch to transit. 7 

Access Convenience: SR 91 survey responses included almost no complaints about a 8 

lack of intermediate access points. This was not surprising given the commuter patterns 9 

of many people using the facility. However, in a fall 1996 telephone survey of SR 91 10 

commuters it was found that 76 percent of the HOV3+ travelers used the ELs for their 11 

trip while 24 percent used the GPLs. Mastako et al. (1998) surmised the access 12 

inconvenience was the reason for not all the HOV3+ using the ELs (even though they 13 

could use the lanes for free). In the I-95 survey, almost one-third of the respondents who 14 

previously carpooled in the HOV lanes but do not use the new ELs mentioned access 15 

inconvenience as one of the reasons. This difference in the perception of access 16 

inconvenience between the two ELs may be because I-95 carpoolers were already used 17 

to unrestricted access to the HOV lanes while SR 91 travelers had no HOV lanes and 18 

hence, the SR 91 Express Lanes was a fresh start for them. It is also possibly because 19 

of the difference in the O-D patterns of commuters on the two corridors. 20 

21 

2.7 Impact on HOV2 Travel 22 

Since both SR 91 and I-95 ELs treat HOV2s as toll paying vehicles similar to SOVs, their impact 23 

on HOV2s can be examined and compared. After the SR 91 Express Lanes opened, there was 24 

a significant increase in HOV3+ vehicles as well as SOVs in the corridor. However, only a small 25 

14  

  

increase in the number of HOV2s was observed. Only 30 percent of HOV2s used the SR 91 1 

Express Lanes as compared to 60 percent in conventional HOV lanes in study areas during the 2 

peak 2 hours (Sullivan, 1998). Therefore, “opening of the SR 91 Express Lanes did relatively 3 

little to encourage or discourage HOV-2 ridesharing” (Sullivan, 1998). 4 

5 

Only 23 percent of all HOV2s in the corridor use the I-95 Express Lanes during the 2 hour 6 

afternoon peak period. Therefore, the percentage of HOV2s using the ELs on both corridors is 7 

clearly less than those in conventional HOV lanes (approximately 60 percent). This may be 8 

because the only incentive left for HOV2 as compared to SOV mode on ELs is that they could 9 

split the toll. The low percentage of HOV2s using the HOT lanes may also be due to the lack of 10 

intermediate access points. 11 

12 

2.8 Impact on Transit 13 

In late 1996, bus riders on the SR 91 corridor were asked if the toll lanes had changed their 14 

travel in any way. Only 14 percent answered positively and 60 percent said no. Conversely, in a 15 

May 2009 survey, bus riders along I-95 mentioned service reliability (55 percent) and travel time 16 

(75 percent) were better with the new ELs. This is because the express bus on the SR 91 17 

corridor does not use the ELs while the express bus on I-95 does. With the implementation of 18 

the I-95 Express Lanes (going from one HOV lane to two ELs) the travel time of buses from 19 

downtown Miami to the Golden Glades Interchange during the afternoon peak period was 20 

reduced from 25 minutes to 8 minutes. However, there was no evidence of SR 91 Express 21 

Lanes changing bus operations in any way. After the ELs opened there was no change in 22 

ridership for the express bus on SR 91 but there was a 30 percent increase in average daily 23 

ridership for the express bus on I-95 comparing January through March in 2008 versus 2009 24 

(Cain, 2009). It is interesting to note that the bus ridership of the 95 Express Lanes increased in 25 

15  

  

spite of the reduced fiscal advantage of using transit (increase in bus fare and decrease in the 1 

price of gas) as well as reduced service and economic recession. 2 

3 

3.0 Comparison of I-15 and I-25/US 36 Express Lanes 4 

3.1 Description of I-25/US 36 Express Lanes, Denver 5 

The I-25/US 36 (hereafter referred to as I-25 for simplicity) Express Lanes began operations in 6 

June, 2006 and extend between downtown Denver and US 36. The ELs are 7 miles long with 7 

two reversible lanes separated from the GPLs by a barrier. SOVs are allowed to use the ELs for 8 

a fee while HOV2+ and motorcycles can use the lanes for free. In May 2008, 2000 permits were 9 

issued to Hybrids for free access to the lanes (Colorado DOT website 2011). The SOV fee 10 

varies by the time of day in an attempt to control the demand for the facility. The goals of the 11 

lane were to optimize HOV lane use, provide a new transportation option, and for toll revenues 12 

to cover expenses. 13 

14 

3.2 Impact on Transit along I-25 15 

There are two primary transit routes, Route B and Route 120X, and approximately 10 secondary 16 

transit routes that use the I-25 Express Lanes. The implementation of the ELs had no impact on 17 

travel time as the travel time of transit buses was generally maintained at the pre-ELs level 18 

(Colorado DOT, 2007). 19 

20 

The total ridership of the 12 primary and secondary routes from 2005 (pre-ELs) through 2009 for 21 

the months September and October along with the average price of gas (EIA website) over the 22 

same two months was examined (see Table 3). Comparing the transit ridership soon after the 23 

implementation of the lanes (September – October 2006) to the ridership levels the year before, 24 

there was a very small change (0.4 percent increase). Transit ridership in subsequent years 25 

16  

  

changed much more dramatically. However, this was likely due to the fluctuations in the price of 1 

gas rather than the new express lanes (APTA, 2011). 2 

However, according to survey results it appears that the ELs did have some impact on transit 3 

use. In August and September of 2008, a survey was conducted in which respondents included 4 

those who recently used the ELs. There were indications from the survey that EL 5 

implementation encouraged some additional use of transit. Approximately 12 percent of the EL 6 

users also agreed to the statement “Because of the Express Lanes I use public transit more 7 

often than I otherwise would.” This additional use of transit due to the ELs may help explain the 8 

one anomaly in Table 4. For most results in Table 4, a decrease in the price of gas corresponds 9 

to a decrease in transit use (and vice-versa). However, just after the ELs opened (September 10 

and October of 2006) the price of gas was almost 10 percent lower than the previous year, yet 11 

transit ridership on the EL was almost unchanged. It is possible the new transit ridership due to 12 

the ELs could have offset the expected decrease in use of transit due to the drop in the price of 13 

gas. However, this is only a possibility – it could have been that transit ridership did not change 14 

for a number of other reasons. 15 

16 

Table 4: Average Weekday Ridership for Selected Denver Transit Routes 17 

(Regional Transportation District, unpublished data, Routes using I-25) 18 

Year Gas Prices (cents per

gallon)

Change Transit Ridership

Change

September- October 2005

283.0 13982

September- October 2006

256.2 -9.5% 14034 0.4%

September- October 2007

273.1 6.6% 15148 7.9%

September- October 2008

374.2 37.0% 16926 11.7%

September- October 2009

235.0 -37.2% 14633 -13.5%

Source: Jeff Dunning, Regional Transport District 19 

17  

  

3.3 Impact on Carpooling 1 

In the August 2008 survey it was found that 54 percent of the respondents, whose only mode of 2 

access to ELs was drive alone, mentioned they drive alone (irrespective of lane) for 4 days or 3 

more in a week. Almost all of them (98.3 percent) mentioned they did not carpool at all in the 4 

last 7 days. Therefore, this group of travelers consistently travels as an SOV. 5 

Also, 17 percent of the carpoolers in the ELs mentioned that they carpool in order to use the 6 

ELs. This indicates that ELs have been able to encourage carpooling among some travelers. 7 

A graph of the average weekday usage of ELs by toll users, HOV users and total users was 8 

compiled from monthly EL performance reports (see Figure 1). Along with the usage data, the 9 

average gas price is also plotted for each month for the Denver area. This data was obtained 10 

from Energy Information Administration (EIA) website. 11 

As shown in Figure 1, HOV usage tends to increase when the price of gas increases, and vice-12 

versa. This makes sense as travelers attempt to share travel costs (such as gas) as those 13 

prices rise. The total number of HOV users has been fluctuating between 6000 and 8000 14 

vehicles per day. Toll usage shows an increasing trend from the opening of the lanes until early 15 

2008 when it starts decreasing. 16 

18  

  

1 

Figure 1: Usage of ELs and Gas Prices (June 2006 Through July 2009) 2 

3 

The unemployment rate for the Denver area may partially explain the decreasing number of toll 4 

users. The unemployment rate was fairly stable from 2005 until early 2008, after which there 5 

was a steady increase in unemployment. Interestingly, the decreasing number of toll paying 6 

users starts around the same time. This is not to say each lost toll paying customer lost their 7 

job- but the worsening economic problems likely pushed some travelers off the tolled lanes. 8 

9 

3.4 Similarities between I-15, San Diego Express Lanes and I-25 Express Lanes 10 

Until May 2008, the I-15 Express Lanes in San Diego extended 8 miles in the median of I-15, 11 

from SR 163 to SR 56/Ted Williams Parkway. It was a two-lane, reversible facility with access 12 

available only at its two ends. SOVs paid a fee to use the lanes. HOVs of two or more 13 

19  

  

occupants (carpools, vanpools, and buses), motorcycles, designated hybrid vehicles and two-1 

axle trucks were permitted to use the I-15 Express Lanes for free. There are many similarities 2 

between the I-15 Express Lanes in San Diego (from March 1998 through May 2008) and the I-3 

25 Express Lanes near Denver. Both have two reversible lanes with concrete barrier separation 4 

from the GPLs, no intermediate access points and both the ELs allow HOV2+ carpools to travel 5 

for free. Both were converted to ELs with no design changes from the previous HOV lanes. The 6 

lengths are also comparable (8 miles and 7 miles respectively) and both the corridors have 4 7 

GPLs in each direction. Therefore, the following section explores these two lanes, their 8 

similarities and what can be learned from their differences. 9 

10 

3.5 Usage of the I-15 and I-25 Express Lanes 11 

During the year 1999, when the I-25 Express Lanes were planned, the ADT of I-25 was 12 

approximately 200,000 vpd. On I-15, the ADT exceeded 250,000 vpd in 1996 when the Express 13 

Lanes started. Therefore, congestion along I-15 was considerably worse. One objective of the 14 

lanes was to relieve GPL congestion (along with increase in HOV lane use and to help fund 15 

transit) (Supernak et al. 2002). 16 

The I-25 and the I-15 Express Lanes started with a similar number of carpools; 7,680 carpools 17 

and 8,050 carpools per weekday, respectively. Over a period of three years the carpool usage 18 

on I-15 Express Lanes increased to 11,400 per weekday (December 1996 through November 19 

1999) (San Diego Association of Governments, 1999) while it changed very little on I-25 (June 20 

2006 through July 2009). 21 

During the same period, daily toll users in both ELs reached almost the same number; 3500 vpd 22 

on I-15 and 3400 vpd on I-25. Therefore, the ELs differ in terms of encouraging their carpool 23 

usage, but not tolled SOV travel. 24 

20  

  

One key difference between the two ELs was in their objectives. I-15 includes the objective of 1 

optimizing the use of HOV lanes as well as to reduce congestion on I-15 by providing a toll 2 

option to SOVs while I-25 has just the objective of optimizing the use of HOV lanes. Also, I-15 3 

has dynamic pricing while I-25 has fixed variable pricing with a minimum peak hour toll decided 4 

by the bus fare ($3.25). Thus, on I-25, the decreasing trend of toll users might be partially due to 5 

the high fixed toll value during the peak period which might have discouraged some travelers 6 

from continuing the use of the ELs. 7 

In terms of supporting services, both ELs have park-and-ride lots. The difference in the available 8 

capacities of the park-and-ride lots might influence the difference in the increase in carpooling in 9 

the two ELs. However, there are three park and ride lots in the vicinity of the I-25 HOT lane with 10 

over 2400 spaces operating at less than two-thirds full. Therefore, two factors have likely 11 

played a role in the much larger use of carpooling on I-15: gas prices and congestion. However, 12 

no conclusions can be made because of lack of information in this aspect. 13 

14 

4.0 Comparison of I-394 and SR 167 HOT Lanes 15 

4.1 Description of the I-394 HOT lanes, Minneapolis 16 

The I-394 Express Lanes run 11 miles between downtown Minneapolis and the western 17 

suburbs. The project allows solo drivers to pay to use the ELs while carpoolers, buses, and 18 

motorcyclists may use the lanes free of charge. Tolls are set by dynamic pricing ensuring 19 

continued free flow in the lanes at about 50 to 55 mph. Along with maintaining free flow on the 20 

lanes, objectives of the lanes included maximizing HOV lane usage and funding highway and 21 

transit improvements. The ELs consist of two sections. The per-trip fee depends on where users 22 

enter and exit the MnPASS ELs. The variable, per-trip fee is always charged for SOV use in the 23 

eastern 3 miles, from Trunk Highway (TH) 100 to I-94. This section consists of 2 reversible 24 

21  

  

lanes, barrier separated from the GPLs. The western 8 mile of the ELs are separated by paint 1 

stripes. This section west of TH 100 is called the diamond lane section and fees are only 2 

charged in the peak direction during rush hours (6 AM to 10 AM eastbound and 2 PM to 7 PM 3 

westbound). 4 

5 

4.2 Impact on I-394 Transit Ridership 6 

Between July and September 2005, shortly after the MnPASS lanes opened, transit ridership 7 

along I-394 increased by more than 13 percent over the ridership from the same period in 2004, 8 

before MnPASS was available. In comparison, transit ridership in the I-35W corridor increased 9 

only 1.4 percent. Chum and Burris (2008) indicated some possibility of bus ridership being 10 

positively affected by the adaptation of HOV to HOT lanes. However, there was no specific 11 

study to determine the effect that having a new SOV toll option had on existing transit users and 12 

it was unknown how many former transit users switched to the SOV toll mode. 13 

Both Turnbull (2008) and Johns et al. (2006) observed that the transit usage remained relatively 14 

constant during the later years after the MnPASS implementation, due in part, to the park and 15 

ride lots generally operating at or near capacity. 16 

4.3 Description of the SR 167 HOT Lanes, Seattle 17 

The SR 167 HOT lanes began operation in May of 2008. They consist of a single HOT lane 18 

running in each direction of SR 167 for approximately nine miles between Renton and Auburn. 19 

SOVs are allowed to use the lanes for a toll, while two-person carpools, vanpools, transit and 20 

motorcycles can use the lanes for free. 21 

Previously, carpoolers could enter and leave the lane anywhere along its entire length and the 22 

lane was separated by a single white line from the GPLs. As a HOT lane, motorists can enter 23 

22  

  

and exit the HOT lanes only at designated access points providing more predictable entry and 1 

exit maneuvers. The access point design is based on I-394 diamond lane section of the HOT 2 

lane project in operation in Minnesota (http://www.wsdot.wa.gov/projects/sr167/hotlanes/). Other 3 

than at entry/exit points, a two-foot buffer with two solid white stripes separate the HOT 4 

lanes from the GPLs. A 10 foot inside shoulder is also available. The objectives of the lane were 5 

to test the HOT lane concept and the publics’ reaction along with collecting data on its use. 6 

4.4 Impact on SR 167 Transit Ridership 7 

There were seven buses using the HOV lane during the peak period in the peak direction. With 8 

HOT lanes the only service change in bus operations was the slight modification of two of those 9 

routes (routes 564 and 565). Transit officials fine-tuned the route alignments, directing buses to 10 

enter SR 167 at SR 516 instead of 84th Avenue. The adjustment allowed the buses to better 11 

use the HOT lanes’ access zones (Washington State DOT, 2009). 12 

Bus ridership increased by 17 percent in the corridor and by 23 percent in the region from 2007 13 

to 2008 likely due to increased gas prices (Puget Sound Trends, 2009). Ridership then 14 

decreased in 2009.The decrease in 2009 was attributed to a reduction in gas prices, economic 15 

recession and regional job loss. Conversely, comparing the months May through November for 16 

2008 and 2009, the average daily tolled trips on the HOT lanes increased by 46 percent. 17 

Therefore, the relationship between the HOT lane usage (toll as well as carpool) and gas prices 18 

was examined further. Tuesday through Thursday HOT lane usage from May 2008 through April 19 

2009 (compiled from the performance reports sent by Kevin C. Beireis, WSDOT - Toll Division) 20 

and monthly average gas prices in Seattle (EIA website) are compared in Figure 2. Carpool 21 

usage on the HOT lanes and gas prices followed remarkably similar trends. On the other hand, 22 

tolled vehicles generally increased in number as the price of gas declined. 23 

23  

  

This strongly implies that the reduction in the price of gas resulted in carpools breaking up and 1 

transit riders leaving the bus (as noted above). While the carpools can shift to SOVs on HOT 2 

lanes and/ or GPLs, bus riders can shift to these two modes as well as carpool on the HOT 3 

lanes. Other shifts, such as from a carpool to transit, are unlikely to occur due to a decrease in 4 

the price of gas. 5 

6 

7 

Figure 2: Usage of HOT lanes by HOV Vehicles and Tolled Vehicles and Gas Prices for 8 May 2008 Through April 2009 9 

10 

The above findings provide some insight into the impact of the price of gas on the mode choice 11 

of travelers. An increase in the price of gas makes carpooling more attractive by travelers 12 

splitting the cost of auto expenses and possibly saving the toll if the traveler was previously an 13 

SOV on the HOT lanes. Therefore, the usage of HOT lanes, to some extent, is dependent on 14 

exogenous factors like the price of gas which are generally not included as one of the elements 15 

in the planning of HOT lanes. It also implies that the performance of HOT lanes should be 16 

May‐08

Jul‐08

Sep‐08Oct‐08

Nov‐08

Dec‐08

Apr‐09

0

50

100

150

200

250

300

350

400

450

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

20000

Apr‐08

Jun‐08

Jul‐08

Sep‐08

Oct‐08

Dec‐08

Feb‐09

Mar‐09

May‐09

Cents

Number of Vehicles

Month

Carpool Vehicles

Tolled Vehicles

Gasoline Price per gallon

24  

  

judged in light of the gas prices in the respective locations and it becomes even more crucial 1 

when the two HOT lanes are compared based on their usage. 2 

3 

4.5 SR 167 HOT Lanes and I-394 Diamond Section 4 

There are many similarities between the SR 167 HOT lanes and the diamond lane section of the 5 

I-394 Express Lanes. Both are concurrent, one lane each direction alongside 2 GPLs in each 6 

direction (as an exception, I-394 has one auxiliary lane in westbound direction in addition to the 7 

2 GPLs). The HOT lanes are separated from the GPLs using double white lines. Also, both HOT 8 

lanes have dynamic pricing with designated intermediate access points. One difference was 9 

congestion levels on the two corridors at the time of HOT lane implementation. The ADT was 10 

120,000 vpd on SR 167 in 2008 and 148,200vpdon I-394 in 2005. Before the SR 167 HOT 11 

lanes, the speed in the GPLs commonly dropped to 35 mph during the peak period. However, 12 

before the I-394 Express Lanes, travel speeds in the GPLs were generally above 50 mph. 13 

The many similarities between the two HOT lanes help to avoid many exogenous factors and 14 

provide an opportunity to compare different impacts of the two HOT lanes 15 

16 

4.6 Usage of the HOT lanes 17 

When the SR 167 HOT lanes started operation in May 2008, the price of gas reached its peak 18 

($4). When I-394 HOT lanes started operation in May 2005, price of gas was $1.95 a gallon and 19 

reached over $3.00 a gallon by July 2006 (Cambridge Systematics, 2006). Therefore, the price 20 

of gas could have had a significant impact on usage of the HOT lanes and comparisons 21 

between the lanes need to be cognizant of this. 22 

23 

25  

  

I-394 usage data from January through April of 2007 (Minnesota DOT, 2008) was compared to 1 

SR 167 usage data from January to April 2009 (carpool data from WSDOT) (see Table 5). The 2 

price of gas in the two cities was similar at these times and both lanes had been open for at 3 

least 6 months. This may help to minimize the impact of the price of gas. 4 

5 

6 

Table 5: SR 167 and I-394 Traffic Volume Data 7 

Road Time of Day

HOV Lane

GPLs Total Toll (SOV)

No Toll (carpools, HOV, transit)

SR 167

AM Peak (7 to 8 AM)

250 (6%)

690(18%)

3200 (76%)

4140 (100%)

PM Peak (5 to 6 PM)

155 (4%)

760(21%)

3000 (75%)

3915 (100%)

I-394

AM Peak (7 to 8 AM)

360a (7%)

744(13%)

3921 (80%)

5025 (100%)

PM Peak (5 to 6 PM)

310b (5%)

474(6%)

5572c (89%)

6356 (100%)

a = counted at the Louisiana section of the diamond lanes. 8 b = counted at the Winnetak Avenue section of the diamond lanes 9 c = one additional auxiliary lane 10  11 

The one large difference was that HOVs comprise a larger percentage of travelers on SR 167. 12 

Since the GPLs on SR 167 were more congested than on I-394, carpoolers may have more 13 

incentives to carpool on SR 167 than I-394. 14 

15 

5.0 Conclusions 16 

This research examined: 17 

the different factors which led to the development of six HOT lanes 18 

the changes made in the corridor due to HOT lane implementation 19 

the different impacts of the HOT lanes on travel and mode choice in the corridor 20 

21 

26  

  

These issues were examined by comparing three pairs of HOT lanes with similar design and 1 

operational characteristics. By comparing the impacts of the HOT lanes on two similar corridors 2 

the impact of many of the exogenous factors are minimized. 3 

4 

With the strict registration requirement for HOV3+ to travel for free on the I-95 ELs, there were 5 

indications of some carpools breaking up. Tolled access for HOV2s on the I-95 ELs as well as 6 

the SR 91 ELs resulted in lower usage of both of these ELs by HOV2s (fewer than 30 percent of 7 

the total corridor HOV2s use the ELs) as compared to a conventional HOV lane (60 percent) 8 

where HOV2 access is free. As for transit, the Express bus does not use the SR 91 ELs and, 9 

not surprisingly, there was almost no change in its ridership after the SR 91 ELs were 10 

implemented. Conversely, on I-95, the Express bus uses the ELs and travel time of buses 11 

decreased by 17 minutes due to ELs implementation. Additionally, the daily ridership on the I-95 12 

express bus increased by 30 percent. 13 

14 

On the SR167 and I-25 ELs, exogenous factors such as the price of gas and the economic 15 

recession seemed to influence the usage of the ELs. On both the ELs, carpool usage increased 16 

along with the price of gas. On I-25, the increasing unemployment rate coincided with a 17 

decreasing number of toll paying travelers. On SR 167 there were also indications of mode 18 

shifts among the transit, carpool and toll paying SOVs due to fluctuating gas prices. With 19 

declining gas prices, the transit and carpool usage decreased while the number of toll paying 20 

users increased. 21 

22 

Acknowledgements 23 

The authors express their sincere thanks to the project sponsor, the Federal Highway 24 

Administration, and to the Contracting Officer Technical Manager (COTM), John Halkias, for his 25 

27  

  

guidance and support. The authors would also like to thank the many people who supplied 1 

information regarding the HOT lanes around the U.S.—either through their research reports or 2 

via direct contacts. The authors would like to acknowledge Stacey Stegman, Colorado DOT for 3 

providing the I-25 survey data. Additionally, Dr. Katie Turnbull and Dr. David Ellis of the Texas 4 

Transportation Institute provided valuable insight and comments on this research effort. An 5 

earlier version of this paper was presented at the Transportation Research Board annual 6 

Meeting in January 2011. 7 

8 

References 9 

95 Express website.<http://www.95express.com/> (accessed on March 2,2010) 10 

American Public Transportation Association (2011) Potential Impact of Gasoline Price Increases 11 

on U.S. Public Transportation Ridership, 2011 -2012, Washington, D.C., 12 

http://www.apta.com/resources/reportsandpublications/Documents/APTA_Effect_of_Gas_Price13 

_Increase_2011.pdf (Accessed on September 29, 2011) 14 

15 

Cain, A (2009) Miami Urban Partnership Agreement (UPA) project, Phase 1A - transit 16 

evaluation. Prepared by the Center for Urban Transportation Research, University of South 17 

Florida, Tampa, Florida. Prepared for Federal Transit Administration (FTA). 18 

19 

Cambridge Systematics Inc.(2006) I-394 MnPASS technical evaluation: final report. Prepared 20 

for Minnesota Department of Transportation, Oakland, California. 21 

22 

Colorado DOT (2007). Summary of I-25 Express Lanes. Denver, Colorado (February 2007) 23 

24 

28  

  

Colorado DOT website (2011) < http://www.coloradodot.info/programs/hybrids > accessed 1 

October 2011. 2 

3 

Chum, G.L. and Burris, M.W. (2008) Potential mode shift from transit to SOV on a HOT lane, 4 

Transportation Research Record: Journal of the Transportation Research Board, No. 2072, 5 

Transportation Research Board of the National Academies, Washington, D.C., 10-19. 6 

7 

Congressional Budget Office (2009) Using Pricing to Reduce Traffic Congestion 8 

<http://www.cbo.gov/ftpdocs/97xx/doc9750/AppendixA.8.2.shtml> (accessed on February 9 

27, 2010) 10 

11 

EIA website<http://www.eia.doe.gov/> 12 

13 

Florida DOT (2009) Final Report- 95 express midyear report project status for urban 14 

partnership agreement. Prepared by the Florida Department of Transportation, Miami, 15 

Florida. 16 

17 

iCommute website <http://www.icommutesd.com/Carpool.aspx>( Accessed on March 13, 18 

2010). 19 

20 

Johns, B. Thompson, N. Sallman, D. Zmud, J. Munnich, L. and Sheehan, S. (2006) I-394 21 MnPASS – a new choice for commuters: how is it working? Rethinking Transportation, Hubert 22 Humphrey Institute for Public Affairs, University of Minnesota, Minneapolis, Minnesota. 23 

24 

29  

  

Mastako, K. A. Rillet, L. R. and Sullivan, E. C. (1998)Commuter behavior on California state 1 

route 91 after introducing variable-toll express lanes, Transportation Research Board 1649, 2 

TRB, National Research Council, Washington, D.C, 47-54. 3 

4 

Minnesota DOT, (2008) I-394 HOV report including 2008- 1st quarter January – March. 5 

Prepared by Regional Transport Management Centre. 6 

7 

Puget Sound Trends (2009) http://psrc.org/data/trends Accessed March 2010. 8 

San Diego Association of Governments (1999) Report to the California legislature: San 9 

Diego’s Interstate 15 congestion pricing & transit development demonstration program as 10 

required by Section 149.1(g) of the Street & Highways code. San Diego, California. 11 

Sullivan, E. (1998) Evaluating the impacts of the SR91 variable toll express lane facility. Final 12 

report for State of California Department of Transportation, Sacramento. 13 

14 

Supernak, J. Golob, J. Golob, T.F. Kaschade, C. Kazimi, C. Schreffler, E. and Steffey, D. 15 

(2002) San Diego’s Interstate 15 congestion pricing project: traffic-related issues, 16 

Transportation Research Record: Journal of the Transportation Research Board, No. 1812, 17 

Transportation Research Board of the National Academies, Washington, D.C., 43-51. 18 

19 

Turnbull, K.F. (2008) High occupancy toll lanes and public transportation. Journal of the 20 

Transportation Research Board 2065, TRB, National Research Council, Washington, D.C, 21 

36-40. 22 

23 

Washington State DOT (2009) SR 167 HOT Lane Pilot Project, First Annual performance 24 

Report (May 2008- April 2009), Olympia, Washington 25 

30  

  

1