Value of Time, Value of Time with respect to Tolls, Value of Tollway

7 posts / 0 new
Last post
Claire Bozic
Value of Time, Value of Time with respect to Tolls, Value of Tollway

I'm working on a project that includes testing tolls on currently free expressways.  Our region has some tollways already, and we include tolls in generalized cost.  We've calibrated a "perception factor" for the current tollways, and it results in the correct number of tollway users but a much higher value of time than a rule of thumb value of time based on a portion of wages.  We explain this by 1) electronic toll collection reduces the perception of paying tolls, and 2) some users, like trucks, don't actually pay the toll but rather pass it along.  Also, using the tollway rather than local roads provides a bundle of benefits in addition to travel time savings including better pavement condition, better incident management, better plowing in winter, fewer or no flooded locations during heavy rains and lower chances of a crash.  We are working with an economic impact consultant who thinks our testing results of a tolled system (congestion priced with managed lanes) shows more tolls collected then benefits to drivers based on a traditional value of time and minutes of time saved.  The tollway operator observed that drivers in our region are insensitive to toll changes, but the toll changes have been in a narrow range.   Hearing people's thoughts on the subjects might help us figure out how to reconcile this.  

Pilo Willumsen

Hi Clare,

It has been observed (and therefore likely to be true) that electronic payment reduces the perceived cost of paying a toll. It is also true that drivers value the better quality of a road and its reliability and are willing to pay for them if a toll is required (but we should incorporate this effect on the Value of Time also on other good roads even if not tolled; user pay for the quality, not the fact they are tolled). It is also true that some drivers of trucks and cars pass on the cost of the toll to their company, spouse or parents (but even then the willingness to pay is not unlimited). It is also true that for longer trips willingness to pay may be a function of vehicle occupancy (and the goods transported) and that for urgent trips willingness to pay to save time can increase dramatically. You can model all these observed effects using the Value of Time as dependent on the type of payment, type of road and who actually bears the cost of the toll. Or you can identify coefficients when multiplied by the time spent on each stretch of road (some tolled) gives you a well-calibrated and balanced model. So, solution 1: do not change the VOT, introduce some coefficients on time spent on roads.

But there may be a better solution. You mention that you have calibrated a “perception factor” that results in a higher VOT than expected. You did not mention whether you assumed a distribution of VOTs and used a multi-class assignment or a single value and the treacherous “toll diversion curve’, something I avoid whenever I can. The interpretation of the VOT is slightly different in each case. If you are using a single VOT (not the best idea) then the VOT you have found is actually the lowest VOT of the drivers that pay a toll. Many drivers have a higher VOT and benefit the consumer surplus; those with a lower VOT (that day, at that time, for that trip purpose and with so many people in the vehicle) decide not to pay the toll and incur the time penalty. If the capture rate of the tolled facility is below 50% then the average VOT is lower than the one that explain your current observations. In Cost Benefit Analysis it is customary in many places to use a single equity VOT, independent of income levels and related but different from behavioural VOT (that calibrates the model), for evaluation. This reduces the trend to build roads for the rich only. So, solution 2: equity and behavioural VOT are usually not the same and that observed at existing toll roads is likely to be an overestimation of the average as it represents those with a higher willingness to pay. You may even be able to show that considering a distribution of VOTs the value you estimated represent the top X% of the driving population.

Rob Bain

Hi Claire

I’m preparing a new research report specifically examining toll price elasticities (and affordability – ie. in the context of budget constraints). The report is currently at the 90% completion stage. The research demonstrates just how easily the true price response to tolls can be masked (dominated) by other factors – chief amongst which is the quantity and quality of available alternatives.

I can load the report temporary onto my website for downloading if it might be of interest to you (or, indeed, others). I’d just ask you to bear in mind that it is still a work-in-progress; albeit that the heavy lifting has been completed.

Just let me know.

Kind regards,

Rob
E: info@robbain.com

From: cbozic=cmap.illinois.gov@mg.tmip.org [mailto:cbozic=cmap.illinois.gov@mg.tmip.org] On Behalf Of Claire Bozic
Sent: 22 August 2018 06:00
To: TMIP
Subject: [TMIP] Value of Time, Value of Time with respect to Tolls, Value of Tollway

I'm working on a project that includes testing tolls on currently free expressways. Our region has some tollways already, and we include tolls in generalized cost. We've calibrated a "perception factor" for the current tollways, and it results in the correct number of tollway users but a much higher value of time than a rule of thumb value of time based on a portion of wages. We explain this by 1) electronic toll collection reduces the perception of paying tolls, and 2) some users, like trucks, don't actually pay the toll but rather pass it along. Also, using the tollway rather than local roads provides a bundle of benefits in addition to travel time savings including better pavement condition, better incident management, better plowing in winter, fewer or no flooded locations during heavy rains and lower chances of a crash. We are working with an economic impact consultant who thinks our testing results of a tolled system (congestion priced with managed lanes) shows more tolls collected then benefits to drivers based on a traditional value of time and minutes of time saved. The tollway operator observed that drivers in our region are insensitive to toll changes, but the toll changes have been in a narrow range. Hearing people's thoughts on the subjects might help us figure out how to reconcile this.
--
Full post: https://tmip.org/content/value-time-value-time-respect-tolls-value-tollway
Manage my subscriptions: https://tmip.org/mailinglist
Stop emails for this post: https://tmip.org/mailinglist/unsubscribe/12884

Claire Bozic

Rob- I was hoping I’d hear from you! Thank you, you were the person I was thinking of when I posted that question. It would be of interest to me, and I’ll keep its work in progress nature in mind. -Claire

 

Rob Bain

Thanks Claire. I’ve uploaded the Price Elasticity research report temporary onto my website (for download). See here (first paper at the top):

http://www.robbain.com/articlesandpapers.htm

Small parts are missing (eg. the foreword) but the content body is all there.

Kind regards,

Rob

From: cbozic=cmap.illinois.gov@mg.tmip.org [mailto:cbozic=cmap.illinois.gov@mg.tmip.org] On Behalf Of Claire Bozic
Sent: 22 August 2018 10:43
To: TMIP
Subject: Re: [TMIP] Value of Time, Value of Time with respect to Tolls, Value of Tollway

Rob- I was hoping I’d hear from you! Thank you, you were the person I was thinking of when I posted that question. It would be of interest to me, and I’ll keep its work in progress nature in mind. -Claire

--
Full post: https://tmip.org/content/value-time-value-time-respect-tolls-value-tollway
Manage my subscriptions: https://tmip.org/mailinglist
Stop emails for this post: https://tmip.org/mailinglist/unsubscribe/12884

vajucywym

Clare, Rob, and Luis:

Picking up on the point that it is the distribution of VOT that is critical and must be taken into account, you might find the brief note at https://www.caliper.com/pdfs/caliper-a-note-on-toll-road-assignment-meth... of interest.

In addition, I might comment that I do not believe that static aggregate traffic assignment methods are suitable for modeling managed lanes.  Even when tolls do not change, fluctuating traffic and levels of service will influence toll road utilization.

We have had good results using a microsimulation-based DTA to model managed lanes, as it makes the choice to pay a toll sensitive to the expected level of service on the tolled facility.

Best regards,
Howard

Howard Slavin, Ph.D.
President
Caliper Corporation
617-527-4700 ext. 301

Frank_Milthorpe

Hi all

I agree with comments about the need for more than a single VOT, either through segmentation (eg high, middle, low income) or distributed values.

Also some studies have found a positive asc for toll roads. For example Bates and Richardson 2011 available at:
  http://www.icmconference.org.uk/index.php/icmc/ICMC2011/paper/view/208

Directly from the above paper.

"In this context, it is widely believed that the choice of tolled routes does not depend purely on the time saving, but also on the quality of the journey time, and possibly other features of the tolled route. In choice modelling terms, this could be represented by an “Alternative Specific Constant” (ASC) in the utility equation. While earlier experience of introducing tolled options in an SP choice context was that the ASC for the tolled road was negatively perceived, a number of more recent studies, in the context of experience of existing tolled roads, have found a positive ASC ."

For existing facilities there are a number of approaches which can be adopted to ensure that the volumes in the model match the observed number of users. However, this is more problematic when producing forecasts for yet to opened facilities.

Cheers

Frank

Frank Milthorpe
Manager Major Projects Demand Forecasting
Roads and Maritime Services
Sydney, Australia