Add costs for transit fares (bus, rail, airfare), vehicle insurance, parking, and vehicle annual licensing/registration/property taxes

[robbieorvis]

Currently, the EPS does not assign any costs to the consumers of public transit (passenger HDV, passenger rail) or air travel, only to the equipment owners. However, to correctly capture changes in household spending from policies, we need to add in these costs.

The BTS has data on revenue from fares for public transit, ridership, and average fares, which can be used together to estimate the total spending per passenger mile and then allocated to each type of travel mode. See these three sources:

https://www.bts.gov/content/average-passenger-fares-current-dollars https://www.transit.dot.gov/ntd/data-product/2019-data-tables https://www.bts.gov/content/average-passenger-revenue-passenger-mile

[robbieorvis]

Full email thread for background:

Hi Robbie,

The trips themselves are tracked in passenger HDV and passenger rail modes. The vehicle and fuel costs are charged to the vehicle owners (private companies and government). However, bus, rail, ship, and air fares paid by riders are not currently tracked.

The ownership shares of each vehicle type are already in the model, so we know how we would allocate fare revenues. We would just need to calculate the fare amounts. This would involve coming up with a multiplier to convert passenger-miles by each of these four modes (bus, passenger rail, passenger ferry, and passenger aircraft) to fare in dollars.

The U.S. model treats passenger ships as recreational boats (which have no fare), but other countries treat that subscript as passenger ferries. So other countries would need a multiplier covering passenger ships.

We do have a policy that reduces air travel, so we also need a multiplier for aircraft, even though we are shifting away from aircraft rather than to aircraft. Consumers save the cost of airfare, which could be large savings.

If you can provide the necessary multipliers in an Excel file, I can do the cash flow programming in Vensim.

Can you open a ticket for this on GitHub eps-us repo?

Jeff

On May 20, 2021 11:13 AM, Robbie Orvis robbie@energyinnovation.org wrote: Hey Jeff,

Quick question for you: are the costs of “consuming” transportation options covered in the model for things that require fares and tickets? For example, if we shift a lot of people out of cars and into buses, are we accounting for the fares riders have to pay to ride the bus? Same thing for air and rail travel. I’m asking because if we look at policy scenarios with heavy mode shifting, I want to be sure the costs of alternative passenger mode options are being covered appropriately so that we don’t overstate changes in household spending.

[jrissman]

Completed in aa488e4.

Note that this changes mode shifting from a negative-cost policy to a positive-cost policy on the cost curves. I think this is an improvement because it always was odd that mode shifting to transit was the most cost-saving policy in the cost curve often. But it seems pretty expensive now. I feel pretty confident with the data I added for cost per passenger-mile because it comes straight from BTS National Transportation Statistics, but you might want to review it anyway, in variable trans/FpUCD. In particular, one thing you could change is the extrapolation method. I estimated fares (after inflation) would continue to increase at the historical rate they have in the past, but you could instead hold inflation-adjusted fares constant if you think that produces more realistic future cost projections.

The other thing you could consider is increasing the share of trips that are eliminated (rather than shifted to bus or rail) in trans/RTMF because in the post-pandemic world, it's become more obvious that people can work from home and meet remotely, so the pre-pandemic estimates of the potential to completely eliminate trips feel like potentially an under-estimate to me now.

I'm going to leave this one open because I think other EPS team members should review the extrapolation method in FpUCD and the mode shift recipients in RTMF. You can test by using a policy package that contains only the "passenger LDV mode shifting" policy set to the maximum of 26%, and then view the Cost Curve to get the cost per ton abated. Assigning @mkmahajan and @robbieorvis for data review.

[robbieorvis]

I’m looking into this now. Programming wise all looks good. A few things come to mind related to the outputs we are seeing.

I found a better datasource for bus fares, because it specifically breaks out these fares and distance. It’s from the American Public Transit Agency. See her: https://www.apta.com/research-technical-resources/transit-statistics/public-transportation-fact-book/. I modified the input data to use this source for buses.

Generally though I’ve realized that we are missing some important data on car costs that likely exacerbates the difference here.

For one, we don’t have insurance costs, which is probably between $500 and $1000 a year per car. Over a 15 year lifetime, that would dramatically increase the cost of car ownership. For example, the NPV of lifetime ownership for a gasoline car (13 year lifetime) is about $45,000. Adding $500-$1000/year in insurance fees would increase that $6500 to $13000, which could add nearly a third at the upper end of that range. We also don’t include taxes, which could add another 5-10%, and we don’t include vehicle registration fees, which can total several hundred dollars per year (like in CA), and potentially $5000 or so over a vehicle lifetime. Finally, we don’t have parking fees, which one study (https://inrix.com/blog/cod/has) at $1,600/year on average (I think this is way too high and only covers cities, which is primarily where you have to pay to park. I’d be nationally this is closer to maybe $100, but for now we’ll include it). If we were to include all these costs, it would pretty dramatically increase the cost of car ownership and make mode shifting look more favorable.

Some quick back of the envelope math then:

NPV of Lifetime Vehicle Cost for Gas Car in EPS in 2020: $44,485 Estimated Annual Insurance Costs (undiscounted): $1194 (source: https://www.aaa.com/autorepair/articles/average-annual-cost-of-new-vehicle-ownership); Lifetime @13 years = $15,522 Estimated Vehicle Licensing, Taxes, and Registration: $753 per year Tax Percent (source: https://www.aaa.com/autorepair/articles/average-annual-cost-of-new-vehicle-ownership.): Lifetime @13 years = $9,789 Estimated Parking Fees: $1,600 per year; Lifetime @13 years = $20,800

Total Cost of Car Ownership: $90,596, almost exactly double the lifetime vehicle cost.

Let’s compare cost per mile:

Average lifetime = 13 years; average annual distance = 11353, average loading = 1.67; total passenger miles driven = 246,473

44485/246473=$0.18/passenger-mile 90596/246473=$0.37/passenger-mile

Data in model on bus fare (this has my updated data from APTA): $0.29/passenger-mile

So, with the updated APTA data (which I think is better than BTS data, because it’s so detailed) and accounting for other costs, the narrative is flipped. There are of course insurance and registration fees spread out for buses and other modes too, but they are spread out over many, many more passenger-miles and likely to be negligible. Plus, important to remember that total public transit revenue is a mix of fare revenue and government funding, the latter of which can often represent a very large share (same data source estimates only 30% of revenue for entire transit industry comes from fares).

I’m going to commit the data update now, and then we should discuss if/how we want to proceed with the missing costs, since they are such a large part of the picture.

---

Robbie Orvis Senior Director of Energy Policy Design +1 415-799-2171 1225 Eye Street, N.W. Suite 904 Washington, D.C. 20005 www.energyinnovation.orghttp://www.energyinnovation.org/ @.***

From: Jeff Rissman @.> Sent: Tuesday, November 9, 2021 1:48 AM To: Energy-Innovation/eps-us @.> Cc: Robbie Orvis @.>; Assign @.> Subject: Re: [Energy-Innovation/eps-us] Add in public transit costs in terms of fares and trips (#161)

Completed in aa488e4https://github.com/Energy-Innovation/eps-us/commit/aa488e494a7c7c2085fd93d75e660ea580c64a70.

Note that this changes mode shifting from a negative-cost policy to a positive-cost policy on the cost curves. I think this is an improvement because it always was odd that mode shifting to transit was the most cost-saving policy in the cost curve often. But it seems pretty expensive now. I feel pretty confident with the data I added for cost per passenger-mile because it comes straight from BTS National Transportation Statistics, but you might want to review it anyway, in variable trans/FpUCD. In particular, one thing you could change is the extrapolation method. I estimated fares (after inflation) would continue to increase at the historical rate they have in the past, but you could instead hold inflation-adjusted fares constant if you think that produces more realistic future cost projections.

The other thing you could consider is increasing the share of trips that are eliminated (rather than shifted to bus or rail) in trans/RTMF because in the post-pandemic world, it's become more obvious that people can work from home and meet remotely, so the pre-pandemic estimates of the potential to completely eliminate trips feel like potentially an under-estimate to me now.

I'm going to leave this one open because I think other EPS team members should review the extrapolation method in FpUCD and the mode shift recipients in RTMF. You can test by using a policy package that contains only the "passenger LDV mode shifting" policy set to the maximum of 26%, and then view the Cost Curve to get the cost per ton abated. Assigning @mkmahajanhttps://github.com/mkmahajan and @robbieorvishttps://github.com/robbieorvis for data review.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHubhttps://github.com/Energy-Innovation/eps-us/issues/161#issuecomment-963862279, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AK5N6SMVZ3L3MA7XTKFCOYTULC7ZVANCNFSM45JFM6DQ. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

[jrissman]

It seems worthwhile to put in the insurance, vehicle registration, and parking costs for passenger LDVs.

I can try to code it in a way that lets you put these things in for any vehicle type, but if you don't have data on (say) the insurance costs paid by public transit agencies to insure their buses, I think that's probably okay. As you note, public transit agencies receive a lot of their funding from things other than fares, and we're not trying to model their internal financials - we're only trying to get the right cost impacts from mode shifting. Insofar as transit agencies are funded by fares (say, 30%, or whatever), the insurance must already be priced into the fares to a similar degree. So this is mainly of issue for passenger LDVs, where the owner is paying the insurance, registration, and parking fees, and isn't just passing them through to customers or getting them covered by government funding.

I'll try to add these other costs soon so we can do a public 3.3.1 release soon.

[robbieorvis]

Agreed – great!

---

Robbie Orvis Senior Director of Energy Policy Design +1 415-799-2171 1225 Eye Street, N.W. Suite 904 Washington, D.C. 20005 www.energyinnovation.orghttp://www.energyinnovation.org/ @.***

From: Jeff Rissman @.> Sent: Monday, December 6, 2021 2:43 PM To: Energy-Innovation/eps-us @.> Cc: Robbie Orvis @.>; Mention @.> Subject: Re: [Energy-Innovation/eps-us] Add in public transit costs in terms of fares and trips (#161)

It seems worthwhile to put in the insurance, vehicle registration, and parking costs for passenger LDVs.

I can try to code it in a way that lets you put these things in for any vehicle type, but if you don't have data on (say) the insurance costs paid by public transit agencies to insure their buses, I think that's probably okay. As you note, public transit agencies receive a lot of their funding from things other than fares, and we're not trying to model their internal financials - we're only trying to get the right cost impacts from mode shifting. Insofar as transit agencies are funded by fares (say, 30%, or whatever), the insurance must already be priced into the fares to a similar degree. So this is mainly of issue for passenger LDVs, where the owner is paying the insurance, registration, and parking fees, and isn't just passing them through to customers or getting them covered by government funding.

I'll try to add these other costs soon so we can do a public 3.3.1 release soon.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/Energy-Innovation/eps-us/issues/161#issuecomment-987122640, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AK5N6SL63WAJZ56DGRM6MRDUPUG5DANCNFSM45JFM6DQ. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

[jrissman]

I've added insurance costs and parking costs to the model. I've also updated the aircraft to recipient mode shifting distribution to favor more virtual meetings and less use of passenger rail, given the poor state of passenger rail in the U.S. (For Europe, China, and Japan, you might want to adjust this back to favor passenger rail a bit more.)

I've also updated the projection of changes in future fares to assume they follow inflation (i.e. they increase about as fast as other goods' prices increase). This past 12 months, transportation services have actually increased in cost more slowly than inflation (see BLS data here, with a 3.9% increase in transportation services compared to a 6.8% increase for all items). Therefore, our previous projection that transportation services would increase in cost faster than inflation in the future seems likely to be wrong. Overall inflation is an average, so half the economy has to increase in cost more slowly than the overall inflation rate, mathematically speaking.

All of these changes help to fix the problem with the mode shifting policy looking too expensive. It now is more reasonable, though still probably slightly too expensive.

I haven't added the vehicle registration/license fees yet because I disbelieve the AAA source Robbie linked to. I'm going to look for a better source.

[jrissman]

Okay, I've been looking into the topic of taxation of cars more deeply. I think the AAA value struck me as high because some states have an annual property tax on cars, while other states do not have vehicle property taxes. In some states (such as California), the taxable value of a vehicle declines as the car depreciates, but other states do things differently. This article says that in Connecticut, the tax is levied on 70% of the vehicle's retail value and is a tax rate of 2.6%, and doesn't appear to decline with depreciation. They say this works out to an annual tax of $650/car in Connecticut, which is about right, given the average new car sales price in the U.S. in early 2020 (before pandemic-related inflation) was $37,876 (per Kelley Blue Book).

This article says that 23 states don't charge annual vehicle property taxes, and of the ones that do, the average per vehicle ranges from $25 in Louisiana to $1,011 in Virginia.

So the AAA number of $753 looks more reasonable in this light, particularly considering it is not just property tax, but also includes the vehicle registration / licensing fees, which exist in all 50 states.

I think I've satisfied myself that the AAA value (1) is probably a reasonable average for the U.S., and (2) doesn't include sales taxes from the initial purchase of the car, which makes it a suitable value to use in this part of the model.

[jrissman]

Okay, this set of features is finally done and appears to be working well. Adding the costs of fares, insurance, parking, and licensing/registration/property tax helps to capture on-road transportation-related cash flows in much better detail. It also greatly improves the accuracy of the cost-effectiveness calculation for the mode shifting policy (i.e. in the Cost Curve diagram). Closing this issue.