USPS Study: EV Economics Depend On Smart-Grid Revenue

Spread the love

John Petersen

On August 28th, the Office of the Inspector General of the U.S. Postal Service published the results of a feasibility study titled, “Electrification of Delivery Vehicles.” While the feasibility study reaches a foregone conclusion and recommends the purchase of a 3,000 unit demonstration fleet, I was surprised by the high level of Federal subsidies the Inspector General thought necessary to bring EVs within Postal Service capital investment policies. I was even more surprised by the conclusion that the tipping point in the economic analysis was revenue from ancillary vehicle to grid, or V2G, services.

The Postal Service operates a fleet of 219,000 vehicles, including 146,000 delivery vehicles. The feasibility study focused on the long-life vehicles, or LLVs, that have been a part of the American landscape since the late 80’s.

LLV Right.jpgLLV Left.jpg

The current version of the LLV is built on a GM truck chassis, costs the Postal Service about $19,000 and gets about 10 miles per gallon; which isn’t bad for the kind of low-speed stop-start driving on a typical mail route. The average LLV is driven about 18 miles a day and roughly 96% of the LLV fleet drives less than 40 miles a day. The vast majority of LLVs are parked at Postal Service facilities from 5 p.m. till 8 a.m.

The proposal evaluated in the Postal Service feasibility study would replace the internal combustion engine and drive train with an electric drive and 20 kWh of lithium-ion batteries of unspecified chemistry. The projected cost of a 3,000-unit fleet of electric LLVs, or E-LLVs, is $120 million, or $40,000 per unit. The projected cost of associated charging station infrastructure and training is $16.75 million.

The most striking aspect of the Inspector General’s report is the fact that it was written from the perspective of an EV buyer, rather than an EV seller. After years of reading up-beat promotional materials that talk about ten-year battery lives and seven- to ten-year payback periods, it was refreshing to see a more skeptical buyer’s analysis that:

  • Assumed the battery pack would have to be replaced after five years;
  • Assumed a fifty percent reduction in repair and maintenance costs;
  • Assumed a stable correlation between gasoline and electricity prices;
  • Required internal returns of thirty percent per year like you see in most businesses;
  • Required payback periods of less than three years like you see in most businesses;
  • Concluded that substantial Federal subsidies were essential; and
  • Concluded that ancillary revenues from V2G services were essential.

The Inspector General’s report was not overly kind to E-LLVs, but then I’ve never expected undue kindness from fleet buyers who are invariably constrained by capital spending policies that require a return on investment, as opposed to a return of investment. The good news for EV developers is that the Inspector General was able to put together a plan that worked for the Postal Service. The bad news is the plan will be difficult for other fleet users to duplicate because the feasibility study assumes that:

  • The Postal Service will get grants for 74% of the cost difference between a standard LLV and an E-LLV;
  • The Postal Service will save roughly $1,300 per vehicle year from reduced fueling costs;
  • The Postal Service will save roughly $1,500 per vehicle year from reduced maintenance; and
  • The Postal Service will earn roughly $2,300 per vehicle year from V2G services.

The Inspector General’s report analyzed four possible scenarios. In the basic scenario of no grants and no V2G revenue, the E-LLVs were a poor investment that had a negative return over ten years. In two middle of the road scenarios that included (a) grants without V2G revenues and (b) V2G revenues without grants, the payback periods were in the five-year range and internal rates of return were 15% to 20%. In a best-case scenario that included both grants and V2G revenue, the payback period was under two years and the internal rate of return was over 60%. Since the Postal Services has influential friends in high places, I think it’s a safe bet that they’ll be able to negotiate the details of a best case project.

The only thing that concerns me about the strategy the Postal Service has adopted for its E-LLV demonstration fleet is the long-term stability of V2G revenue. The E-LLV fleet will be on the road every day from 8 a.m. to 5 p.m., the precise period when demands on the power grid are greatest. So while the proposed fleet of 3,000 E-LLVs will have the theoretical ability to provide 45 MW of frequency regulation services, it will only be able to provide frequency regulation services when demand for those services is relatively low. While I’ve not been able to find any detailed estimates of the national demand for frequency regulation services during off-peak hours, I have to assume that the aggregate demand for frequency regulation is smaller than demand for other grid-based storage systems. I also have to assume that V2G services will compete directly with alternatives like the flywheel systems that Beacon Power (BCON) is developing which will be available 24/7.

Overall, I believe the Postal Service proposal to deploy a fleet of 3,000 E-LLVs presents an unparalleled opportunity to provide a reliable real-world testing laboratory for ideas that have not yet been reduced to practice. The Postal Service has long promised “neither snow, nor rain, nor heat, nor gloom of night, nor the winds of change, nor a nation challenged, will stay us from the swift completion of our appointed rounds.” Since one of the biggest challenges facing America is the efficient use of energy and the prevention of waste, I can’t imagine a better organization to lead the way.

In a perfect world, the Postal Service would break its planned E-LLV fleet into as many as a half-dozen subgroups that would each use a different battery chemistry from a different vendor. The willing industry participants I can identify off the top of my head include Altair Nanotechnologies (ALTI), Ener1 (HEV), Johnson Controls (JCI), Valence Technologies (VLNC), and A123 Systems (IPO pending). With proper monitoring, the amount and relative uniformity of the data generated in the first few years of testing for both EV and V2G applications could be priceless.

As a side note, I’m pleased to announce that I’ve accepted an invitation to appear as a luncheon speaker at the Electrical Energy Storage Applications and Technologies conference in Seattle on October 4th through 7th. While bloggers like me frequently get invited to speak at investment conferences, EESAT is in an entirely different animal. It’s a biennial international technical conference co-sponsored by the DOE, Sandia National Laboratories and the Electricity Storage Association. The agenda currently inclu
des technical presentations from the U.S. and eleven foreign countries. EESAT is not appropriate for investors, but it’s a must for companies that are active in the energy storage sector and for institutional investors who need to better understand why storage is important and where the growth opportunities lie.


John L. Petersen, Esq. is a U.S. lawyer based in Switzerland who works as a partner in the law firm of Fefer Petersen & Cie and represents North American, European and Asian clients, principally in the energy and alternative energy sectors. His international practice is limited to corporate securities and small company finance, where he focuses on guiding small growth-oriented companies through the corporate finance process, beginning with seed stage private placements, continuing through growth stage private financing and concluding with a reverse merger or public offering. Mr. Petersen is a 1979 graduate of the Notre Dame Law School and a 1976 graduate of Arizona State University. He was admitted to the Texas Bar Association in 1980 and licensed to practice as a CPA in 1981.


  1. John:
    About two weeks ago, I got a chance to meet with officals from one of the highest-profile EV/PHEV charging infrastructure companies in the US.
    When I asked them about V2G, they claimed that the main barrier wasn’t with the charging infrastructure (at least not with their technology) or the grid, but rather with the vehicles.
    First gen EVs/PHEVs, they claimed, are not equipped to “talk” with the grid and therefore have their storage capacity optimized remotely based on grid operator needs.
    What are your thoughts on this? Did the report address this issue?

  2. Charles,
    The feasibility study does not get into a great deal of detail respecting the technical capabilities of the E-LLVs. As I noted in the article, even the battery chemistry was left open ended. The study does speak to lengthy discussions between the Postal Service and the PJM ISO, so I would imagine the technical requirements have been discussed. But there’s no way to know until more data becomes available.

  3. Congratulations on your latest speaking engagement.
    I support the concept of using the USPO, and other delivery fleets, to support “greener” vehicles. But, I too read the USPO report. And, I too think it was missing detail. More important, I feel it was a sorely superficial analysis.
    For example, what have the new LLVs done regarding “lightweighting”? Would a vehicle based in the proposed IDEA van from Bright be a better alternative?
    As another example, the vehicles they are considering are targetted for general duty across the range of mileage, weather and terrain. That approach inherently drives the battery size larger. If vehicles are categorized by projected energy requirement based on these types of considerations, initial fleet purchases could target delivery routes that would require smaller batteries.
    I’m not sure I would field test products from a half dozen different suppliers. While it would “spread the wealth”, it would sub-optimize production efficiency. I’d suggest two at the most.
    I would however consider evaluating a micro-hybrid approach. For many delivery routes, particularly in the US Southern tier of States, the weather is generally temperate, the terrain is relatively flat and there are many “short to mid-length” routes. A microhybrid with start-stop and regen may be just right. Obviously that would dramatically reduce the battery pack size. It would also probably enable using less expensive battery chemistries. Alternatively, for the very short routes, a pure BEV may be “best”.
    My point is that, by attempting to optimize fleet cost using a “one size fits all” approach, the USPO may actually be substantially increasing costs. A better approach may be to standardize on one body style and two chassis & drivetrain cominations, a microhybrid or BEV and a PHEV. And, the PHEV should probably have the ability to support two or three battery sizes.
    I agree with your observation regarding V2G. In this area, the report provided no information that would support their goals. Although there is a need for grid support in the early PM (5-9), the vehicle batteries would be somewhat discharged by that time, unless they substantially oversized the batteries or set the vehicle to maintain them at a high charge level. Neither is a good option. There is also a need for grid stabilization in the early AM (6-9). But if the batteries are used then, the vehicles begin traveling with depleted batteries. Even changing the delivery time doesn’t help much. I don’t see how V2G would work to the extent that they forecast for this application.
    Finally, I saw no consideratin given to the “fuel option premium”. As a “tie-in” to a recent post, there is a benefit to the USPO to de-couple from the gas supply. Given their delivery credo, “neither rain nor snow…”, it would seem to be valuable to add, “nor lack of gasoline”.

  4. Very interesting article John.
    It seems like the USPS will be missing an opportunity to increase the cost effectiveness of it’s EV fleet by fine-tuning the battery pack to the route. Since routes are of different length, it would seem to make the most sense to only put in a battey pack that can go a few miles more than the expected route length, perhaps having dedicated EV-10, EV-20, and EV-40 versions. Due to the much lower cost of the EV-10’s these may not need so many grants to make them economic on those routes.
    As a fleet operator, they might also be able to benefit strongly from battery swapping. I believe a lot of USPS routes involve returning to the station several times a day. If they then swap batteries, they will need less overall range, and the batteries charging at the station will be available for frequency regulation and other grid services during the day.
    Perhaps some of this was in the report, but you didn’t mention it?

  5. Interesting comments and it looks like a refreshingly honest assessment by the USPS on EV. I wonder about relying on long term V2G revenue streams, as it seems to me that grid-dedicated energy storage will outcompete V2G applications.
    However, with respect to the USPS typical operating hours, often high demand for grid regulation service is in the early morning and late evening periods when electric load is increasing/decreasing rapidly – roughly around 6 AM and 10 PM depending on location and time of year. So typical vehicle operating times of 8-5 PM would not forgo all of the good opportunities for providing ancillary services.
    Tom Konrad’s ideas for tuning battery size to routes and battery-swapping are intriguing, and may squeeze a little more value out of a EV fleet.
    The USPS’s proposal to serve as “a national laboratory for EV” is a good one. There is much to be learned and as they describe in their report, they are well positioned to test EV in a variety of locations/conditions.

  6. The big benefit of vehicle uniformity that really can’t be overstated is comparability of operating data under a wide variety of conditions. When you disperse a reasonable sample of identical vehicles around the country and run them under similar conditions, you develop a tremendous amount of data by studying the variables. Since the IG report specifically recommended holding off on larger deployment decisions until EV technology matured, the best we can hope for right now is a large carefully controlled demonstration project.

  7. Tom, advanced lead acid might well do the job for significantly less, but it’s pretty clear that the decision has already been made to test lithium and there are some decisions that it’s just not worth fighting. This is one of them.
    Ted, I was surprised to see that regulation values were highest in the morning and evening because it seems a bit counterintuitive, but when you consider ramping it makes a world of sense. Thanks for teaching me. There’s never a need to apologize for disagreeing with me. That’s why I love these comment sections.

  8. I just received a reader comment by e-mail that said:
    I hope the postal service has back-up ICE vehicles for their proposed Electric fleet.
    I don’t believe the proposed vehicles will have a long enough range with Stop and Go driving conditions.
    Put a couple vehicles in cold climates and the battery range will drop dramatically, just from the cold weather. Add in a heater fir driver comfort and I think you will be looking at a 3 to 4 hour effective use span. .Information I obtained from an electric car development engineer says lithium batteries like a temperature range of 65F to 85F and life on either side of those parameters is poor at best.
    Tom K.
    I agree with Tom that some or all of these factors may be problems, but the only way we’ll know for sure is to build the fleet and test it throughout the country to find out what works and what doesn’t. Any way we slice it, a 3,000 vehicle test beats the heck out of a decision to convert 146,000.

  9. Well, if V2G is such a money maker, than why doesn’t doesn’t a grid operator purchase a set of li-ion batteries and regulate the grid, and make millions of dollars? Why not skip the EV step?

  10. Point taken regarding the desirability of large scale testing. That being said, there is nothing in the USPO report to indicate that such a detailed, multivariate test will be structured. Indeed, the few options considered in the report would argue that the analysis will likely be similarly simplistic. Further, since the report seems to give substantial weight to V2G sale of electricity, and since most ISOs do not yet have appropriate tariffs, and since the report specifically notes discussions with the PJM ISO, it seems likely the test fleet will be targeted, at least predominately, at the PJM area. Given the harsh winters in PJM, and the associated impact upon battery performance, I am not sure this is the best choice for initial testing. Further still, based on the number of charging stations assumed in the report, a limited number of test sites appears likely.
    It should also be noted that large sample sizes are not necessarily critical to statistically valid analysis. If one assumes a binomial distribution for the population being tested (yes, I know that is not necessarily a valid assumption), a sample size as small as about 30 could be adequate.
    Finally, given the precarious financial condition of the USPO, why not focus initially upon gathering benefits and test results from applications with the highest probability of good payback? The test can always be expanded subsequently as knowledge and technology improve. This approach has the added public relations benefit of emphasizing initial successes. In contrast, initial testing across a wide range of applications will almost certainly generate mixed results that could be open to “misinterpretation”.

  11. Anon, apparently the standard protocol is to have system operators (which typically don’t own assets) buy regulation services. Beacon Power is building a flywheel based system to provide those services and both Altair Nanotechnologies and A123 Systems have sold demonstration units to AES. The whole idea of V2G is “if we’re going to buy the batteries and connect them to the grid, we may as well try to optimize the value.” It’s a great argument if you’re talking about thousands of vehicles but the theory gets much murkier when you talk about hundreds of thousands.
    Mike, most of the ISOs are wrestling with the issue of storage tariffs. While established tariffs are not common today, it’s a fair bet that they will long before the USPO can buy 3,000 E-LLVs and put them into service. I agree that 3,000 units would be statistical overkill, which is why I would prefer to see 5 or 6 different configurations with a dozen geographically dispersed work sites per configuration. This could be a great study if properly designed and monitored. It could also be a boondoggle. At the end of the day the IG’s report said “if the DOE pays to build the beasts we’ll drive them.” That does not seem like poor management to me.

  12. Fred, the lithium-ion developers are not a personal favorite class for me but it seems that Electrovaya has managed to avoid the bloated stock valuations that make some of the others unattractive. I also try to avoid foreign companies because of currency and other differences. That being said, I think I may have to give Electrovaya a closer look.

  13. John: I am in the project financing business and must say that the bussiness DO NOT require 30% IRRs. That is extremely high where as 15-20% is more common. In addition, “Pay Back Time” is a bad metric for something that you must have like a truck to deliver the mail or a house. After all, what is the “Pay Back” time on the gasoline operated truck or on your home mortage. NPV is a much more useful metric. gw


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.