WORKING PAPER 2019-01

When does electrifying shared mobility make economic sense?

Authors: Nikita Pavlenko, Peter Slowik, Nic Lutsey Date: January 2019 Keywords: electric vehicle; ride-hailing; total cost of operation; payback period

Introduction The use of shared and electric vehicles analysis, we also assess the importance together offers a promising opportu- of driver access to home charging on Over the past several years, the nity to accelerate the benefits of each. electric vehicle operating costs. We reach and use of shared vehicles has Given the high annual miles traveled track the shift in per-mile operating expanded significantly throughout the of vehicles used in shared fleets, TNC costs and the associated payback United States, particularly in large met- drivers have an opportunity for greater period for BEVs relative to conven- ropolitan areas. Examples of shared annual fuel savings from hybrid and tional and hybrid vehicles under a vehicle fleet applications include taxis, battery-electric vehicles (BEVs) than variety of use cases. carsharing, and ride-hailing. Use of private drivers. This, in turn, is likely to ride-hailing fleets, often referred to mean lower per-mile operating costs, as transportation network companies and shorter payback periods, depend- Methodology (TNCs), is especially on the rise. ing on the exact vehicle and energy This study develops a TCO approach to TNC fleets account for a major prices. Although TNC vehicles are con- evaluate the relative costs of purchas- share of trips in major cities—and nected with many broader questions ing and operating vehicles for TNC, this sector is pro-jected to both related to congestion and transit use, taxi, and carsharing applications. We intensify and expand its geographic their electrification offers an oppor- evaluate costs over a five-year owner- scope (Sperling, Brown, & tunity to eliminate TNC vehicles’ local ship period, consistent with typical D’Agostino, 2018; Transportation emissions. Excitement about such con- ownership duration of both Kelley Research Board, 2018). vergence opportunities has led to the Blue Book and Edmunds TCO assess- policy investigation into the viability ments. The costs incorporated in the Largely a separate trend, the of new regulations to increase zero- TCO analysis include upfront vehicle deploy-ment of electric vehicles emission vehicle adoption in fleets cost and taxes, discounted annual has accel-erated in many of the (California Air Resources Board, 2018). energy costs (i.e., electricity charging same urban areas experiencing or gasoline fueling), and maintenance growth in shared mobility. The This report assesses the timing of cost- costs. We apply a 5% discount rate on increased adoption of electric effectively electrifying shared mobility future expenses beyond the purchase vehicles results from the con-fluence fleets in U.S. cities, with a focus on year. We also incorporate an estima- of local level policies, financial ride-hailing. We develop a total cost tion of the opportunity cost of public incentives, charging of operation (TCO) metric for conven- charging for electric vehicles in ride- infrastructure, and awareness tional, hybrid, and electric vehicles hailing operations. campaigns (Slowik & Lutsey, in eight U.S. cities. We incorporate 2018). Improvements in battery regional variation in incentives, taxes, The vehicle types assessed in this technology have continued to and energy costs and apply vehicle study include a reference conventional decrease the price of electric technology improvements to assess vehicle, a reference hybrid vehicle, vehicles and increase their the changing purchase and operat- and three reference BEVs (150-mile, driving range. However, electric ing costs through 2025. Within the 200-mile, and 250-mile electric range), vehicles remain several thousand dollars more expen-sive than their conventional counter-parts in 2018.

© INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION, 2019 WWW.THEICCT.ORG WHEN DOES ELECTRIFYING SHARED MOBILITY MAKE ECONOMIC SENSE?

each with separate cost and effi- in high-annual-mileage fleets. Fuel cell our analysis includes efficiency ciency specifications. We assess the vehicles are in a nascent stage and face improvements over time for all vehicle shifting upfront and operating costs greater barriers with model availabil- types, including reducing test-cycle of BEVs, relative to conventional and ity, refueling infrastructure, and higher and real-world per-mile CO2 and fuel hybrid vehicles, from 2018 to 2025. upfront costs (see, e.g., Isenstadt & consumption by 3% per year for con- This allows us to estimate the payback Lutsey, 2017). Beyond the informa- ventional vehicles, and 1% per year for period—the number of years of use tion in this section, we include further hybrid and electric vehicles, through required to offset the difference in details in tables in the appendix. 2025. Taking into account those upfront costs—over time, factoring annual efficiency improvements, we in regional variation, shifts in energy Battery costs are expected to decrease estimate that the fuel economy of significantly from 2018 through 2025, costs, and technological improvement. conventional and hybrid vehicles will largely due to economies of scale from Applicable near-term state purchasing improve from approximately 30 and higher-volume battery manufactur- incentives are included but generally 53 miles per gallon in 2018, respec- ing. We base our battery cost reduc- are phased out around 2019–2021 and tively, to approximately 37 and 57 tions on the best available bottom- are excluded for 2022–2025. miles per gallon by 2025. up material and production battery manufacturing studies, vehicle engi- VEHICLE SPECIFICATIONS neering teardown analysis, and direct SHARED APPLICATIONS Upfront purchase costs are one of automaker statements about battery To assess the use-phase costs incurred the largest factors in the TCO of each costs (Ahmed, Nelson, Susarla, & Dees, by vehicle use, this analysis highlights vehicle. In 2018, BEV purchase costs 2018; Anderman, 2016; Anderman, four different shared applications 2018; Berckmans et al., 2017; Davies, are typically a few thousand dollars to assess the relationship between C., 2017; Lienert & White, 2017; UBS, more than comparable hybrid vehicles, vehicle use, TCO, and payback period. 2017). Based on this information, we which in turn are typically a few The daily driving distances for ride- assume per-kilowatt-hour battery pack thousand dollars more than compara- hailing drivers are inferred from costs decline at a compounded 5.5% ble gasoline models. The largest deter- Uber’s 2015 driver roadmap and an per year from $205 in 2017 to $130 in minant of electric vehicle purchase assumption of 20 miles driven per 2025. Also, based on UBS (2017), we prices relative to conventional vehicles hour of work, split out across full- assume non-battery electric vehicle is the cost of battery packs, which, time drivers (40 hours per week) and components (e.g., motor, power elec- based on an industry survey, were part-time drivers (18 hours per week) tronics) will decrease by 21% from 2017 approximately $205 per kilowatt hour (Benenson Strategy Group, 2015). to 2025. (kWh) in 2017, down from $1,000 in Maven Gig drivers log an average of 2010 (UBS, 2017; Chediak, 2018). Our For conventional gasoline vehicles, approximately 135 miles per day, with reference vehicle is a car with utility we assume additional technology approximately 10% of drivers going and size specifications that approxi- costs to comply with efficiency stan- beyond the vehicle’s range on a given mately match models like the Nissan dards. Between 2018 and 2025, the day. Taxi services, which often operate Sentra, Toyota Corolla, Chevrolet upfront purchase costs of conven- fleets with drivers sharing multiple Cruze, and Toyota Prius. For BEVs, our tional gasoline models are assumed to vehicles over several daily shifts, fre- utility dimensions roughly match the increase by 0.5% per year, meaning the quently have even higher daily driving Nissan Leaf and Chevrolet Bolt, but we 2018 reference conventional vehicle distances (Metro Transportation consider varying electric battery pack sees an increase in price from $19,000 Denver, 2018; New York City Taxi & size and electric range, as discussed to $19,700 due to its additional effi- Limousine Commission, 2014, 2016). further below. We do not assess ciency technology, based on infor- Carsharing fleets tend to be used for plug-in hybrid electric vehicles or fuel- mation from the U.S. Environmental short, intra-urban trips; data for this cell electric vehicles here. Compared Protection Agency (EPA, 2016) and assumption were taken from car2go to BEVs, plug-in hybrids are chal- Lutsey et al. (2017). For hybrid costs, in five major North American cities lenged by relatively high fueling and we assume a decrease in costs from (Martin & Shaheen, 2016). We rec- maintenance costs, and higher upfront $23,000 in 2018 to $22,000 in 2025 ognize there is broad variation and costs when battery costs drop in or approximately 0.6% per year, uncertainty in these applications, future years, and they often operate based on EPA (2016) and German and we focus on average applica- similar to non-plug-in hybrid models (2015). Based on the same sources, tions because we are more interested

2 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION WORKING PAPER 2019-01 WHEN DOES ELECTRIFYING SHARED MOBILITY MAKE ECONOMIC SENSE?

in the broader overall implications Table 1. Summary of assumed average driving patterns for selected shared applications. than seeking out the implications for Private Ride-hailing Ride-hailing early-adopting drivers and fleets with vehicle driver driver higher annual travel activity. owner (full-time) (part-time) Carsharing Taxi driver Annual The four shared applications, along driving 15,000 40,000 18,000 8,000 70,000 with comparable approximate (miles) numbers for private vehicle owner- Days driving 365 280 150 320 280 ship as a reference, are presented in per year Table 1. The table shows the annual Daily miles 41 143 120 25 250 number of miles traveled for each of the shared applications, along with the Table 2. Percentage of annual miles on original daily charge based on electric vehicle assumed days per year that the vehicle range for each shared application. is used. The daily miles traveled is an estimate of the average number of Private Ride-hailing Ride-hailing Vehicle vehicle driver driver miles driven each day that a vehicle is technology owner (full-time) (part-time) Car-sharing Taxi driver used, which is used below to assess a BEV-150 mile 100% 70% 81% 100% 40% given shared application’s reliance on BEV-200 mile 100% 95% 99% 100% 60% fast charging versus overnight home charging. In Table 1, the two shared BEV-250 mile 100% 100% 100% 100% 80% applications with the highest vehicle BEV-250 mile - 0%–50% - - - use are full-time ride-hailing drivers (limited overnight charging) and taxi services. The annual miles traveled for other applications, such for each vehicle technology across vehicles, primarily because electric as part-time ride-hailing or carshar- various shared applications. For each motors have far fewer parts than ing, are more similar to private vehicle BEV technology that we assess, we internal combustion engines, do ownership, although their day-to-day estimate the share of that vehicle’s not require engine fluids, and have driving patterns may be very different. electricity charging that is expected less wear and tear on brakes due to The daily miles traveled in each of the to be overnight at home, based on regenerative braking. Hybrid-electric fleet use cases assessed here inform the assumption that drivers will seek vehicles generally incur ongoing main- our assessment of the relative share out DC fast charging whenever their tenance costs somewhere between of overnight home charging relative to remaining battery range falls below 50 conventional vehicles and BEVs. Our public fast charging. Using a weighted miles on a given day’s drive. Generally, maintenance cost assumptions are random distribution of miles driven shared applications with more annual based on the UBS (2017) electric car per day over the annual operation for miles traveled, such as full-time ride- teardown analysis. We assume mainte- each of the fleet use cases above, we hailing and taxi services, have an nance costs of $0.061 per mile, $0.037 assess the number of days per year increased reliance on public DC fast per mile, and $0.026 per mile for the that drivers travel beyond the capacity charging, particularly when driving conventional vehicle, hybrid vehicle, for the vehicle technology in question shorter-range BEVs. To evaluate and BEV, respectively. (i.e., BEV-150, BEV-200, BEV-250) the potential upper bound for BEV costs, we also evaluate the TCO for and assume that the remaining daily REGIONAL VARIATION travel distance is powered via direct shared applications where drivers are current (DC) fast charging. Using the unable to charge their vehicles over- To assess the impact of regional varia- weighted random distribution, we night regularly and thus use public DC tion on the TCO and payback periods assume on days that drivers with 50 fast charging for 50%–100% of their for BEVs, we analyze eight U.S. cities miles of range or fewer remaining on charging needs. to include a wide range of geographic their batteries will opt to recharge that coverage, BEV purchasing incentives, day at a DC fast charging station. and energy costs in our analysis. MAINTENANCE COST These cities are Austin, Chicago, Table 2 shows this study’s esti- BEVs generally incur lower main- Denver, Los Angeles, New York, mates of overnight home charging tenance costs than conventional Portland, San Francisco, and Seattle.

WORKING PAPER 2019-01 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION 3 WHEN DOES ELECTRIFYING SHARED MOBILITY MAKE ECONOMIC SENSE?

Along with covering a range of dif- prices from the Energy Information 30% decrease in vehicle range, increas- ferent geographic regions, climates, Administration (EIA, 2018a), averaged ing drivers’ use of DC fast charging energy prices, and other factors, for the last year. We incorporate EIA’s on those days relative to home over- these markets are also relatively high assessment of city-specific gasoline night charging. To quantify this impact electric vehicle uptake markets with prices where applicable, using EIA’s on electricity consumption over the a proliferation of shared mobility prices for nearby cities for Austin and course of a year, we then weight the applications (Slowik & Lutsey, 2018; Portland, the two cities in our analysis annual home versus DC fast charging Shaheen, Martin, Hoffman-Stapleton, not contained in EIA’s dataset. For the for that vehicle to account for the share & Slowik, 2018). cities assessed here, gasoline can be of cold weather days for that city. as much as 20% cheaper (e.g., Austin) Regional policy incentives in this or 20% more expensive (e.g., Los ELECTRICITY PRICES AND analysis include battery electric Angeles) than the national average. CHARGING COSTS vehicle rebates such as those available To project the cost of gasoline from in Denver and California. These rebates 2018 to 2025, we use data from the For our analysis of fueling costs for amounted to a reduction of the upfront International Energy Agency’s World BEVs, we incorporate both regional purchase costs by $5,000 in Denver Energy Outlook (IEA, 2015), estimat- variation in electricity prices as well and $2,500 in both of the California ing an approximate 4% growth rate as additional costs attributable to fast cities in this analysis (Colorado Energy that increases the price of gasoline to charging. We show the current prices Office, n.d.; California Air Resources nearly $3.90 per gallon by 2025. available to residents in select cities Board, n.d.). State rebates typically based on the largest utility in the met- worth $2,000 and $2,500 are also Regional climate variation also may ropolitan area in Table 3. Most utili- available in New York and Oregon, affect the cost of BEV operation, as ties offer either flat rates, for which all respectively. The state of Washington electric vehicles can suffer from effi- electricity is priced the same regard- has a sales tax exemption for BEVs ciency losses in cold weather. Cold less of the energy quantity or time of that applies up to the first $32,000 of weather operation increases the day consumed, or tiered rates, which vehicle purchase costs (Washington internal resistance within the battery increase based on the number of kWh State Department of Licensing, 2018). and lowers its effective capacity; used in a month. Tiered rates shown The clean vehicle rebates are assumed furthermore, cold weather driving below are from the second-cheapest to phase out at various points by 2022, also necessitates heating the vehicle tier of prices available. In place of either whereas the Washington state tax interior, requiring additional energy tiered or flat rates, some utilities also exemption phases out after 2018. For (U.S. Department of Energy [DOE], offer time-of-use (TOU) rates wherein most of the cities in the analysis, sales n.d.a; Union of Concerned Scientists, the lower cost of energy delivery, typi- taxes ranged from 8% to 10% of vehicle 2018). Together, these effects can cally overnight, is reflected in lower purchase price. In contrast, Oregon’s reduce BEV range significantly; recent rates to consumers. As indicated, the 0% sales tax reduces the upfront costs data suggest that BEV range may tiered rates in Los Angeles and San for new vehicles by several thousand decrease by at least 25%, with some Francisco make their electricity prices dollars regardless of vehicle technol- studies noting range decreases of higher than many of the others, but ogy, uniformly reducing the relative around 45% (see, e.g., Salisbury, 2016; the TOU rates there are more similar TCO compared to other cities across DOE, n.d.a; Yuksel & Michalek, 2015). to rates elsewhere. Some utilities vehicle categories. The $7,500 federal require separate meters to access income tax credit is not included here. To assess the impact of climatic varia- electric vehicle rates, but we ignore This tax credit applies uniformly across tion on BEV TCO, we use an assess- those installation costs for this analysis. the cities in our study, and it is set ment from the National Climatic Data For areas that offer multiple electricity to expire at different times for differ- Center of the average number of days options, we adopt the lowest available ent automakers (e.g., perhaps within per year with an average temperature residential charging cost, which is the 2019 for Tesla and General Motors, and of 45 °F or lower for each of the cities TOU option when it is available. We later for other automakers depending studied. We then revise the charging apply a uniform increase in electricity on when they reach 200,000 electric behavior assumptions in Table 2 to price to 2025 for all cities based on vehicle sales). account for the share of annual driving the projected increase in electricity days for each shared application that prices generally, as modeled by the To incorporate regional variation in occur in cold weather, by city. For each U.S. Energy Information Administration vehicle fueling costs into the analysis, cold weather day, we assume that (2018b). Some of the utilities in the we use average monthly retail gasoline drivers operate their vehicles with a listed regions plan to offer TOU rates

4 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION WORKING PAPER 2019-01 WHEN DOES ELECTRIFYING SHARED MOBILITY MAKE ECONOMIC SENSE?

by 2025, but these rate choices are not Table 3. Residential electricity prices ($/kWh) in select cities in 2018 and 2025. reflected in the 2025 price shown. Even though TOU pricing is always lower 2018 2025 than flat or tiered rates in the same area, City Time of use Flat / tier Time of use Flat / tier some flat or tiered rates are lower than Austin $0.14 $0.15 TOU rates when looking across regions. Chicago $0.10 $0.11 Many factors drive these prices, but Denver $0.08 $0.10 $0.08 $0.10 TOU rates are more reflective of the Los Angeles $0.13 $0.35 $0.14 $0.37 generation cost in an area (Nicholas, New York $0.09 $0.18 $0.09 $0.19 2018; Ryan & McKenzie, 2016). Portland $0.04 $0.12 $0.04 $0.12 The future cost of fast charging is San Francisco $0.13 $0.43 $0.14 $0.46 more difficult to estimate because it Seattle $0.13 $0.14 is highly sensitive to utilization. Some companies try to cover systemwide Table 4. Fast charging electricity prices ($/kWh) in select cities. costs by providing a uniform price for a given meter whether it has high uti- 2018 lization or low utilization. Two such City Low case Central case High case 2025 companies providing service nation- Austin $0.27 $0.43 $0.59 $0.28 ally with unified pricing are EVgo and Chicago $0.27 $0.43 $0.59 $0.28 Blink. EVgo charges by the minute whereas Blink charges by the kWh. Denver $0.32 $0.45 $0.59 $0.33 EVgo’s member price is $0.15–$0.21 Los Angeles $0.23 $0.41 $0.59 $0.24 per minute, which translates into a cost New York $0.29 $0.44 $0.59 $0.30 per kWh of $0.24–$0.34 by assuming Portland $0.27 $0.38 $0.49 $0.28 an average power delivery rate over a charging session of 40 kW. This San Francisco $0.23 $0.41 $0.59 $0.24 reflects the fact that vehicles charging Seattle $0.27 $0.38 $0.49 $0.28 at a 50kW charger do not draw the maximum power over a charging The cost of fast charging may decrease hybrid and $0.09–$0.13 per mile for session. Blink’s rates are by the kWh in the future as utilization increases the conventional vehicle, based on our and range from $0.49–$0.59 depend- and the number of chargers per site assumptions shown in the appendix. ing on location. Where no chargers increases; however, even with these per from these two companies exist in kWh cost decreases, some researchers the cities shown, the higher price of OPPORTUNITY COSTS OF estimate that there needs to be at least $0.59/kWh was assumed. FUELING a 40% markup on the cost of electric- Depending on the distance of a daily ity to make the business case for fast For this analysis, we assume a “central drive and the battery capacity of chargers (Serradilla, Wardle, Blythe, & case” based on average market prices given BEV, a driver may need to stop Gibbon, 2017). Fast charging energy for 2018. For future years, we assume and refuel before the day’s drive is can cost 2 to 3 times that of residen- that fast charging electricity prices complete. Although DC fast chargers tial energy for a host site (Nicholas & approximately match the “low case” offer substantially faster charging than scenario, largely as a result of greater Hall, 2018), and this remains generally Level 2 chargers, even fast chargers charger utilization and reduced costs true in our assumed electricity rates may take 30 minutes or longer to of non-hardware components such as in our analysis here through 2025. To complete a charge—substantially more site preparation, grid upgrades, and give context to the 2025 fast charging time than refueling a conventional installation costs (see, e.g., Nicholas electricity costs relative to driving on vehicle at a gas station. To quantify & Hall, 2018). For future years, we gasoline-fueled vehicles, the range of the impact of the time lost charging also apply a general electricity price $0.24–$0.33 per kilowatt hour trans- that otherwise could have been spent increase as reflected by the EIA annual lates to per-mile costs of about $0.08– driving, we assess the opportunity energy outlook (EIA, 2018b). We show $0.09 per mile for the BEV, compared costs associated with vehicle fueling these prices in Table 4. to about $0.06–$0.08 per mile for the for full-time, ride-hailing drivers.

WORKING PAPER 2019-01 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION 5 WHEN DOES ELECTRIFYING SHARED MOBILITY MAKE ECONOMIC SENSE?

Data on the hourly wages of ride-hail- Table 5. Vehicle fueling and charging behavior for full-time ride-hailing applications, by ing drivers is scarce and can be highly vehicle technology. uncertain, with substantial variation Average duration depending on TNC service, time of (Minutes per event) day, and region. We assess the time Charging or fueling stops per day 2018 2025 spent charging for all three BEV tech- BEV-150 milea 1.2 32 9 nologies relative to time spent fueling a conventional and hybrid vehicles. We BEV-200 mile 0.5 36 11 assume that the hourly earnings before BEV-250 milea 0.2 47 15 BEV-250 mile (no expenses for a median driver in 2017 0.5 47 15 are $22.90, based on data collected by overnight charging) the New York City Taxi and Limousine Hybrid 0.2 5 5 Commission from a variety of ride- Conventional 0.4 5 5 hailing services operating within New a Electric vehicles assumed to have overnight home charging York City (Parrot & Reich, 2018). This estimate includes the elapsed time with the expected average peak power BEV-150, we estimate that a driver will between a driver’s first trip of the day of 150 kW, although some will have as spend approximately 175 hours on DC and the end of that driver’s last trip. high as 350 kW. For conventional and fast charging per year in 2018, decreas- Based on our assessment of vehicle hybrid vehicles, the number of refuel- ing to approximately 50 hours per charging behavior, which is summa- ing stops is estimated as a function of year by 2025. In contrast, a BEV-250 rized in Table 2, we estimate that full- the assumed 40,000 miles of travel requires an average of 9 hours of fast time ride-hailing drivers are able to for a full-time, ride-hailing driver and charging per year in 2025, comparable power between 70% and 100% of their those vehicles’ tank sizes and fuel with the 8.5 hours required for refu- average travel from home overnight economies (U.S. Department of Energy, eling a conventional vehicle. This is charging, depending on BEV range n.d.b). We assume that each gasoline due to the smaller number of DC fast from 150 to 250 miles. The share of refueling stop takes approximately five charging events needed for longer- home charging varies further, accord- minutes (GasBuddy, 2017). Beyond range BEVs, as well as the anticipa- ing to weather conditions, because simply fueling up, conventional vehicle tion of next-generation high-powered cold weather decreases BEV battery drivers—particularly full-time ride- charging stations. capacity and necessitates more hailing drivers—may use the opportu- frequent public DC fast charging. nity to purchase food or use facilities, The table also shows the case where a Averaged across the eight cities in potentially extending a given stop by BEV-250 driver lacks access to over- our analysis, this suggests a home several more minutes. However, for the night home charging. In such a case, charging rate of between 63% and 97% purposes of this analysis, we focus only the high average daily travel would of annual travel, depending on BEV on the time spent fueling. necessitate public fast charging more range and regional climate. Over the than twice as often as the BEV-250 course of a year, this averages out to Table 5 summarizes our assumptions driver with home charging. As identi- between 0.2 and 1.2 DC fast charging of vehicle fueling behavior, illustrating fied above but not quantified here, stops per day per vehicle. the assumed number of stops per day there is evidence that conventional that a given vehicle technology would vehicle drivers may pair fueling with Electric vehicle charging time is a require for a full-time ride-hailing breaks to eat or use facilities, extend- function of both the vehicle’s battery application. The charging stops and ing the duration of stops and thereby capacity and the speed of the charger. duration exclude overnight residential minimizing the relative gap in oppor- Based on our assumption that drivers charging. The minutes per stop entries tunity costs between electric and con- will seek out DC fast charging when illustrate our estimate of the duration ventional vehicle refueling. We note below 50 miles of range, we assume of fueling for each type of vehicle and that the figure values are averages that the charging time is determined are used to estimate the total hours over a year of driving; many drivers by the number of kWh needed to bring of fueling necessary for the assumed could ultimately charge multiple times the vehicle from 50 miles of range 280 days of operation over the course on higher-driving days, and not at all remaining to 80% state of charge. We of a year. From 2018 to 2025, the time on many other days. The no-overnight- assume an average transfer rate of 37 spent charging electric vehicles drops charging case is analyzed in greater kW in 2018 that increases through 2025 by approximately two-thirds. For a detail below.

6 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION WORKING PAPER 2019-01 WHEN DOES ELECTRIFYING SHARED MOBILITY MAKE ECONOMIC SENSE?

Results Vehicle cost Taxes Maintenance Fuel Average cost per mile $60,000 $1.20 TOTAL COST OF OPERATION $50,000 $1.00 In this section, we use the parameters $0.80 and assumptions discussed above $40,000 to develop a TCO for the five vehicle $30,000 $0.60 categories across a variety of uses. $20,000 $0.40 We first describe the results of the TCO assessment, measured as dollars $10,000 $0.20 spent per 5 years of ownership, broken Average cost per mile $0 $0.00 out by driver profile. In Figure 1, we Five-year total cost of operation Private Ride-hailing Ride-hailing Carsharing Taxi assess the 5-year TCO for a BEV-250 vehicle driver driver across five different potential shared owner (full-time) (part-time) applications, using 2018 data for the cost projections averaged across the Figure 1. Comparison of 5-year TCO across shared applications for a 250-mile range battery electric vehicle in 2018. eight cities in the analysis. The total cost includes upfront purchase cost, fueling, ongoing maintenance, taxes, Vehicle cost Taxes Maintenance Fuel Opportunity cost Average cost of charging per mile and applicable state-level financial incentives. Although the purchase $60,000 $0.30 cost and upfront costs are fixed across $50,000 $0.25 the five cases, the maintenance and fueling costs vary significantly accord- $40,000 $0.20 ing to the annual miles traveled and $30,000 $0.15 fast charging behavior attributed to $20,000 $0.10 the different shared applications. $10,000 $0.05 Average cost per mile The two shared applications with the $0 $0.00 highest annual miles traveled, full-time BEV BEV BEV Hybrid Conventional Five-year total cost of operation ride-hailing drivers and taxi operators, (150-mile) (200-mile) (250-mile) (53 mpg) (30 mpg) have the highest overall TCO. However, Comparison of 5-year TCO across vehicle technologies for full-time ride-hailing these two uses spread out the TCO Figure 2. drivers in 2018. costs over more total miles, thus result- ing in per-mile costs of approximately vehicles (see Table 1), we focus on full- availability of consistent fast charging, $0.20 per mile, the lowest of the five time ride-hailing drivers in the analsyis we focus on full-time, ride-hailing applications. Private vehicle owners, below, as the lower per-mile cost drivers in the analysis below. carsharing fleets, and part-time ride- benefits of electric vehicles in high- To assess the relative benefits of elec- hailing drivers have much higher mileage applications is a key question trification, we next evaluate the TCO per-mile costs, ranging from $0.44 of this research. For taxis, on the other to $0.93 per mile, depending on the for the five vehicle technology types hand, we identify the potential for the annual miles traveled and the number for full-time, ride-hailing drivers. Figure lowest per-mile cost of any shared of days driven. This suggests that 2 incorporates the same cost com- application in the analysis, largely frequent, high-volume drivers have the ponents in 2018 as Figure 1, adding due to the high volumes of vehicle most to gain from electrification. in the opportunity costs associated use, wherein the car is shared across with vehicle refueling relative to the We find that carsharing and part-time drivers for shifts. The high annual gross pay for driving in a TNC fleet. The ride-hailing have some of the highest miles traveled for taxis incurs greater figure illustrates the TCO for full-time per-mile costs in the analysis, largely fast-charging costs for those drivers, ride-hailing drivers in 2018. For full- because the upfront cost of the BEV as they would generally require at time ride-hailing divers, we estimate purchase is spread out over far fewer least one DC fast charge per day. Due per-mile costs of $0.21 to $0.31 per total miles traveled. Given the much to uncertainty about the ability of mile. Here, the hybrid offers the lower annual miles traveled of carshare charging between driver shifts and the lowest per-mile costs, followed by the

WORKING PAPER 2019-01 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION 7 WHEN DOES ELECTRIFYING SHARED MOBILITY MAKE ECONOMIC SENSE?

conventional vehicle for our base year Vehicle cost Taxes Maintenance Fuel Opportunity cost Average cost of 2018. The BEV-150 has the highest of charging per mile costs of any vehicle in this assessment $60,000 $0.30 because of the considerably higher opportunity costs associated with $50,000 $0.25 frequent fast charging and the higher $40,000 $0.20 recharging costs that are incurred from $30,000 $0.15 more frequent DC fast charging. $20,000 $0.10 As technology improves and energy $10,000 $0.05

costs shift over time, the relative costs Average cost per mile incurred by ride-hailing drivers for $0 $0.00 BEV BEV BEV Hybrid Conventional different vehicle technologies shift in Five-year total cost of operation (150-mile) (200-mile) (250-mile) (57 mpg) (37 mpg) turn. The costs associated with fast charging decrease over this time Figure 3. Comparison of 5-year TCO across vehicle technologies for full-time ride-hailing period as a result of greater charger drivers in 2025. utilization, reduced costs of non-hard- ware components—such as site prepa- BEV (150-mile electric range) Conventional ration, grid upgrades, installation—and BEV (200-mile electric range) Hybrid faster charging speeds, thus reducing BEV (250-mile electric range) energy and opportunity costs (see, $0.32 e.g., Nicholas & Hall, 2018). $0.30 Figure 3 illustrates the TCO for all five vehicle technologies in 2025. As $0.28 shown, BEV-250s have the lowest $0.26 per-mile cost, at $0.20 per mile, while the BEV-200 and BEV-150 have costs $0.24 of $0.22 and $0.25 per mile, respec- tively. The conventional and hybrid $0.22 vehicles see costs of $0.25 and $0.21 Average cost per mile $0.20 per mile, respectively. $0.18 Figure 4 illustrates the change in 2018 2019 2020 2021 2022 2023 2024 2025 per-mile costs for full-time ride-hailing drivers from 2018 to 2025 based on a Figure 4. Five-year TCO average cost per mile for full-time ride-hailing drivers across 5-year TCO. From 2018 to 2025, the vehicle categories, 2018 to 2025. cost of driving hybrids and conven- respective per-mile costs continue with hybrids, and the BEV-250 ulti- tional vehicles increases by 3% and 4%, to move in opposite directions after mately has the lowest per-mile costs. respectively, due to both increasing 2018, improving the value proposi- gasoline prices and additional technol- Again, we note that the federal tax tion for BEVs over time. The TCO of ogy necessary to increase efficiency credits of $7,500 per BEV are excluded the hybrid also increases over the toward the prevailing vehicle stan- here because their expected expiration time period assessed, but the break- dards. At the same time, the per-mile varies by automaker in upcoming years even point relative to the BEV-250 cost of BEVs declines by 18% to 21%, (i.e., perhaps for Tesla and General mostly due to battery price decreases, occurs around 2023, as a result of the Motors within 2019, and others there- as well as from increased efficiency and lower upfront costs of hybrids and after). The results from Figure 4 would reduced opportunity costs from faster their high fuel economy relative to be quite different for 2018–2021 if the charging rates. Although the TCO for conventional vehicles. By 2025, the tax credit for each BEV was included. the BEV-250 and conventional vehicle BEV-150 has lower operating costs With inclusion of the federal tax credit are similar in 2018 after factoring in than conventional vehicles, while the the BEV-250 already would be at cost available state-level incentives, their BEV-200 becomes cost-competitive parity with hybrid vehicles in 2018,

8 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION WORKING PAPER 2019-01 WHEN DOES ELECTRIFYING SHARED MOBILITY MAKE ECONOMIC SENSE?

instead of in 2023 as shown in the BEV (250-mile), 0% home charging Conventional figure. With inclusion of the federal BEV (250-mile), 50% home charging Hybrid tax credit, the BEV-150 would reach BEV (250-mile), 100% home charging cost-per-mile parity with the hybrid by $0.32 2020, whereas the figure shows it only reaches cost parity with conventional $0.30 vehicles in 2024. $0.28 Our assessment of the 5-year TCO $0.26 illustrates that BEVs will become com- petitive with hybrid and conventional $0.24 vehicles over the next several years. Figure 4 suggests that long-range $0.22

BEVs already in 2018 are competi- Average cost per mile tive with conventional vehicles in full- $0.20 time ride-hailing applications, under $0.18 the assumptions outlined above even 2018 2019 2020 2021 2022 2023 2024 2025 without the federal tax credit. Mid- and shorter-range BEVs reach cost com- Figure 5. Five-year TCO average cost per mile for full-time ride-hailing drivers driving a petitiveness with conventional vehicles BEV-250, with 0%, 50%, and 100% home charging electricity rates, 2018 to 2025. approximately 2 to 5 years after long- With this we are illustrating the case driving. Figure 6 illustrates the impact range BEVs. However, these results of drivers without reliable options for of lower and higher annual driving for hold true only in cases where there is nightly home charging, who are there- drivers, from a low of 30,000 miles access to consistent home charging. fore more reliant on higher cost DC per year, to a high of 60,000 miles per In cases where overnight home fast charging. For drivers using high year, relative to the central assump- charging is unavailable, such as for rates of fast charging, we see sub- tion of 40,000 miles per year. For a driver occupying a multi-family stantially higher per-mile costs—as the low driving case, the BEV-250 housing unit without a dedicated much as 25% higher for drivers relying per-mile operating costs in 2025 parking space, the lower electricity entirely on DC fast charging. This increase by around 21%, from $0.20 costs for home charging would be moves the breakeven point between to $0.24 per mile, and it retains its replaced with higher per-kWh costs BEV-250s and conventional vehicles relative cost advantage over conven- for public fast charging. Fast charging significantly later—to 2022 for drivers tional and hybrid vehicles. For higher can cost from 3 to 9 times as much as with 50% home charging, and to 2024 annual driving, the per-mile TCO of overnight home charging, depending for drivers who are unable to charge the BEV-250 is reduced by 13%, down on location (see Table 3 and Table 4) at home. This suggests that reliable to $0.17 per mile. However, because in 2018, although we assume that this access to cheaper home charging of the increased fast charging and price differential will decrease over makes a substantial difference in the opportunity costs for the BEV-250 at time as the cost of DC fast charging cost-competitiveness of BEVs. For greater miles traveled, the comparable declines, as outlined above. ride-hailing drivers living in multi-unit hybrid sees a greater cost reduction dwellings without private garages, for in the high-annual-driving case. As Figure 5 illustrates the impact of example, it would be critical to ensure shown, at 60,000 miles per year the greater reliance on fast charging for access to comparable home charging BEV-250 and hybrid are even in cost full-time ride-hailing drivers. In this options for BEVs to be a viable option. in 2025. In other words, whereas our figure, we consider the BEV-250 full- central case had BEV-250 first beating time ride-hailing driver with access to Because there is considerable varia- the hybrid on costs in 2023, as shown home charging every night—providing tion in the actual annual miles traveled in Figure 4, the high-annual-driving 100% home charging, as in the scenario among full-time, ride-hailing drivers, case has the BEV-250 beating the above—and we add cases for 50% we evaluate whether the BEV-250 hybrid on cost two years later, in 2025. home charging, for situations where retains its relative cost advantage over However, we provide the caveat that charging is available half of the nights conventional and hybrid vehicles in this analysis does not have sufficient of the year, and 0% home charging. 2025 for different amounts of annual data to analyze how maintenance

WORKING PAPER 2019-01 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION 9 WHEN DOES ELECTRIFYING SHARED MOBILITY MAKE ECONOMIC SENSE?

needs are likely to increase for all the $0.30 $0.29 vehicle types over time in such heavy $0.25 $0.25 annual driving cases. $0.25 $0.24

$0.21 $0.22 $0.20 PAYBACK PERIOD $0.20 This section expands upon the TCO $0.17 $0.17 analysis to estimate the payback $0.15 period for BEVs relative to conven- tional vehicles, accounting for ongoing $0.10 technological improvement and Average cost per mile regional cost variation. Similar to the $0.05 approach previously described, for 2018 through 2025 we evaluate the $0.00 difference between the upfront, fixed Conventional Hybrid BEV Conventional Hybrid BEV Conventional Hybrid BEV (250 mile) (250 mile) (250 mile) costs of the BEV-250 and the con- 30,000 miles per year 40,000 miles per year 60,000 miles per year ventional vehicle as well as the shift in variable annual costs, updating to Figure 6. Five-year TCO average cost per mile for full-time ride-hailing drivers driving a account for shifts in efficiency and BEV-250 relative to conventional and hybrid vehicles, across several different estimates energy prices each year. of annual miles traveled in 2025.

Figure 7 illustrates the city-specific Austin Chicago Denver Los Angeles New York City decline in payback period we observe Portland San Francisco Seattle Average over the time period. Here, the 7 payback period shown on the vertical 6 axis reflects the number of years of operation necessary for a full-time, 5 ride-hailing driver to recoup the addi- tional cost for acquiring a BEV-250 4 relative to a conventional vehicle in a given year. For example, a BEV-250 3 in 2020 would require about 3.5 years of driving to recoup the costs 2 if purchased in Denver (green line), Payback period (years) but nearly 5 years if purchased in 1 Austin (brown line). Abrupt changes and upticks in the payback period 0 are attributable to the anticipated 2018 2019 2020 2021 2022 2023 2024 2025 expiration of fiscal incentives, such Figure 7. Payback period for a 250-mile battery electric vehicle relative to a conventional as the phaseout of Colorado’s credit vehicle for a full-time, ride-hailing driver, 2018–2025. through 2021. For state incentive programs that currently do not have a Figure 7 illustrates that on average 4 years and 2 years, respectively, by defined phaseout or expiration date, (black hashed line) across the cities 2025. we assume incentives to be available assessed here, the average payback through 2020. As stated earlier, this period for a BEV-250 from 2018 to Although not depicted in the figure, our analysis excludes federal tax credits. 2025 declines from approximately 5 results indicate that electric vehicles If $7,500 federal tax credits are years to just under 2 years, with sub- are considerably more economically included, the average payback period stantial variation among cities, par- attractive in ride-hailing operations drops to 3 years in 2018, down from ticularly during the initial years. The than in use by private owners. This is 5 years in the figure, and to less than decline in payback period is similar for due largely to their much higher annual half a year in 2025, down from 2 years the BEV-150 and BEV-200, with the activity that is typically 40,000 miles in the figure. payback period for both declining to compared to a more typical 15,000

10 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION WORKING PAPER 2019-01 WHEN DOES ELECTRIFYING SHARED MOBILITY MAKE ECONOMIC SENSE?

miles for an average private vehicle. to conventional vehicles and hybrids. compelling than for typical private With the fuel savings accruing 2 to 3 On the opposite end of the spectrum, vehicle owners. times faster, on average, a full-time cities with low TOU rates, such as Los ride-hailing driver of a BEV-250 sees Angeles, Portland, and San Francisco, Even without purchasing incentives, a payback period that is less than half consistently generate short payback BEVs will become the most economi- that of a private owner of the same periods for electric ride-hailing opera- cally attractive technology for ride- electric vehicle. tion across all comparisons. hailing operations in the 2023–2025 time frame. Our central case scenario The payback period for the BEV-250 Conclusion has electric vehicles becoming the relative to hybrids takes longer to most economically attractive option decline than when compared to con- Our assessment investigates the for ride-hailing drivers by 2023. For ventional vehicles. For the hybrid conditions under which electrifying especially high-annual-driving ride- comparison in particular, region- shared mobility makes economic hailing vehicles, hybrids retain their specific factors contribute to long sense in the United States. Although advantage until 2025, due to electric payback periods in the early years— the case already can be made in vehicles’ greater fast charging needs. for example, high electricity prices or many situations in 2018, this analysis Electric vehicles beat conventional and low gas prices—or both together can is focused on the broader trends for hybrid vehicles on economic grounds narrow the operating costs in some average shared-use vehicle cases. in this time frame for several reasons. cities, such as Austin, and reduce the We close by offering four main con- Foremost is the expected decline in relative per-mile benefits of BEVs in clusions and several final reflections battery pack costs from around $200 those cities. This effect decreases as related to future work. Although we per kilowatt hour in 2018 to $130 by the upfront price and operating costs analyzed several different shared 2025. Of course if battery costs decline of BEVs decline over time. electric vehicle cases, we ultimately more quickly than this, electric ride- focused our analysis on ride-hailing hailing could proliferate even more The regional variation among payback applications, due to their especially quickly. In addition, greater access to periods is primarily attributable to compelling case for electrification. overnight charging as well as technol- differences in average temperatures, ogy advancements that increase fast Based on underlying economics, ride- incentive availability, and energy costs charging speeds greatly reduce driver hailing vehicles are ripe for electrifi- across the cities. Cold weather cities, downtime and associated opportu- cation. Because of their greater annual such as Chicago and New York, require nity costs. Average ride-hailing drivers mileage, typical full-time ride-hailing greater reliance on public DC fast across diverse cities—including Austin, drivers have fuel savings that accrue 2 charging during the winter months, Chicago, Denver, Los Angeles, New to 3 times faster when they buy more increasing the payback periods for York, Portland, San Francisco, and fuel-efficient vehicles. Without incen- drivers. Cities with the most attrac- Seattle—would see payback periods tives, hybrids are the most attractive tive payback periods generally have for plug-in electric vehicles versus technology in 2018 based on lowest warmer climates, lower electric- conventional vehicles drop from 4–7 ity costs, and incentives for electric total cost of operation. Battery electric years in 2018 to 2–3 years by 2023. vehicle drivers. In Denver, Los Angeles, vehicles in 2018 beat hybrids in ride- and San Francisco, upfront electric hailing applications, provided that Access to affordable charging will vehicle rebates result in a reduction in state-level incentives and the $7,500 be critical to unlocking the economic the payback periods by several years federal tax credit are included, and benefits of electric ride-hailing. over the 2018 through 2021 period, when home overnight charging is avail- Electric vehicles become more attrac- elevating these cities to the shortest able. The payback on electric vehicles tive than conventional and hybrid payback periods among the analyzed in ride-hailing applications is achieved vehicles most quickly in cases where cities. The phaseout of rebates after at least twice as quickly as for private drivers have access to overnight 2021 brings these cities closer to the electric vehicle owners, which is again residential charging. Greater reliance eight-city average. Energy cost dif- attributable to their greater annual on public fast charging increases ferentials are another important factor mileage. In addition, because ride-hail- operating costs by about 25% due in regional variation. Austin has low ing and taxi fleets approach vehicles to higher energy costs from fast gasoline prices and relatively high elec- from a commercial perspective, this charging and also greater opportu- tricity prices, negatively influencing makes economic metrics regarding nity costs from time spent charging. the value proposition of BEVs relative fuel savings and payback period more For electrification to become a viable

WORKING PAPER 2019-01 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION 11 WHEN DOES ELECTRIFYING SHARED MOBILITY MAKE ECONOMIC SENSE?

mainstream option for ride-hailing 2020s. Even when electric vehicles in high-mileage fleet applications are drivers, greater access to lower-cost reach hybrid ownership cost, com- clear, more work is needed to better overnight charging infrastructure at panies will need to deploy charging understand the policy toolkit avail- homes, multi-unit dwellings, at the networks, with policy support, to able at state, city, and utility levels that curb, and at public locations will be meet the specific charging needs of can accelerate the transition to much important. This is an important area growing ride-hailing fleets. It also is broader applications for shared electric to be addressed in future policies, for likely that ride-hailing drivers have mobility. This work identifies expand- example in utility charging plans and awareness and education barriers ing access to affordable residential utility rate structures. similar to those of private drivers, so and public fast charging as critical to there is a role for companies to play ensuring an attractive electric vehicle Electrification of ride-hailing fleets is in educating and steering their drivers value proposition for ride-hailing not likely to occur naturally, so new toward electric vehicles with financ- drivers in U.S. cities. Analysis like this policy and company efforts will be ing and purchasing guidance. The will improve greatly with better resolu- . Despite the results suggesting key supply of electric vehicles is limited tion into diverse annual patterns of ride- a compelling case to shift to plug- in most markets, but based on the hailing drivers, if more detailed data ins, such a shift will be neither quick opportunity identified here, ride- become available over time. Despite nor likely if driven by market forces hailing company demand could start the uncertainties, it appears to be clear alone. Even with the positive econom- to dictate their ideal electric vehicle that, with the right mix of policies and ics, ride-hailing fleet electric uptake is model types and attributes. largely in city-specific pilots. Hybrids company leadership, shared fleets can are more economically attractive, and Although the near- and long-term be at the forefront of the transition to this likely will persist until the early economic benefits of electric vehicles electric vehicles.

12 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION WORKING PAPER 2019-01 WHEN DOES ELECTRIFYING SHARED MOBILITY MAKE ECONOMIC SENSE?

References Gasbuddy. (2017). Foot traffic report for the fuel & convenience store industry. Retrieved from https://business.gasbuddy.com/ Ahmed, S., Nelson, P., Susarla, N., Dees, D. (2018, March). wp-content/uploads/2017/05/FootTrafficReportQ1.pdf Automotive battery cost using BatPaC. Presented at the IEA Workshop on Batteries for Electric Mobility, Paris, France. German, J. (2015). Hybrid vehicles: Trends in technology develop- Retrieved from https://www.iea.org/media/Workshops/2018/ ment and cost reduction. Retrieved from the International Session2ShabbirAhmedANL.pdf Council on Clean Transportation, https://www.theicct.org/ publications/hybrid-vehicles-trends-technology-development- Anderman, M. (2016). The Tesla battery report: Tesla Motors: and-cost-reduction Battery technology, analysis of the Gigafactory and Model 3, and the automakers’ perspectives. Retrieved from Total Battery International Energy Agency. (2015). World Energy Outlook. Consulting website: http://www.totalbatteryconsulting.com/ Retrieved from https://www.iea.org/newsroom/news/2015/ industry-reports/Tesla-report/Extract-from-the-Tesla-Battery- november/world-energy-outlook-2015.html Report.pdf Isenstadt, A., & Lutsey, N. (2017). Developing hydrogen fueling Anderman, M. (2018, February). Progress in EV-battery cell cost infrastructure for fuel cell vehicles: A status update. Retrieved and performance: How far and how fast? Presented at the from the ZEV Alliance website: http://www.zevalliance.org/ SAE Hybrid Vehicle Symposium. Retrieved from http://www. hydrogen-infrastructure-status/ pvsheridan.com/SAE-2018/2.30%20pm%20-%20Menahem%20 Lienert, P., & White, J. (2017, January 9). GM races to build Anderman,%20Total%20Battery%20Consulting.pdf a formula for profitable electric cars. Reuters. Retrieved Benenson Strategy Group. (2015). Uber—The driver roadmap: from https://www.reuters.com/article/us-gm-electric-insight/ Where Uber driver-partners have been, and where they’re gm-races-to-build-a-formula-for-profitable-electric-cars- going. Retrieved from https://newsroom.uber.com/wp-content/ idUSKBN1EY0GG uploads/2015/01/BSG_Uber_Report.pdf Lutsey, N., Meszler, D., Isenstadt, A., German, J., & Miller, J. (2017). Berckmans, G., Messagie, M., Smekens, J., Omar, N., Vanhaverbeke, Efficiency technology and cost assessment for U.S. 2025– L., & Van Mierlo, J. (2017). Cost projection of state of the 2030 light-duty vehicles. Retrieved from the International art lithium-ion batteries for electric vehicles up to 2030. Council on Clean Transportation, http://www.theicct.org/ Energies 2017, 10(9), 1314. Retrieved from http://www.mdpi. US-2030-technology-cost-assessment com/1996-1073/10/9/1314 Martin, E., & Shaheen, S. (2016). Impacts of Car2Go on vehicle California Air Resources Board. (2018, August 1). [Letter from ownership, modal shift, vehicle miles traveled, and greenhouse Governor Edmund G. Brown Jr. to California Air Resources gas emissions: An analysis of five North American cities. Board]. Retrieved from https://www.arb.ca.gov/msprog/zero_ Retrieved from Innovative Mobility Research, Transportation emission_fleet_letter_080118.pdf Sustainability Research Center, UC Berkeley, https://www.epa. gov/sites/production/files/2017-06/documents/05312017- California Air Resources Board. (n.d.). Plug-in vehicle resource martin.pdf center. Retrieved from: https://www.driveclean.ca.gov/pev/ Costs/Vehicles.php Metro Transportation Denver. (n.d.). Metro Taxi—Sustainable. Retrieved from http://www.metrotransportationdenver.com/ Chediak, M. (2017, December 5). The latest bull case for electric about/hybrid-taxis/ cars: The cheapest batteries ever. Bloomberg. Retrieved from https://www.bloomberg.com/news/articles/2017-12-05/ New York City Taxi & Limousine Commission. (2014). Taxicab latest-bull-case-for-electric-cars-the-cheapest-batteries-ever factbook. Retrieved from http://www.nyc.gov/html/tlc/down- loads/pdf/2014_tlc_factbook.pdf Colorado Energy Office. (n.d.). Alt fuel vehicle tax credits. Retrieved from https://www.colorado.gov/pacific/energyoffice/ New York City Taxi & Limousine Commission. (2016). TLC alt-fuel-vehicle-tax-credits factbook. Retrieved from http://www.nyc.gov/html/tlc/down- loads/pdf/2016_tlc_factbook.pdf Davies, C. (2017). VW I.D. EV boast: We’ll hugely undercut Tesla’s Model 3 says exec. SlashGear. Retrieved from https://www. Nicholas, M. (2018). Ensuring driving on electricity is cheaper than slashgear.com/vw-i-d-ev-boast-well-hugely-undercut-teslas- driving on gasoline. Retrieved from the International Council model-3-says-exec-17491688/ on Clean Transportation, https://www.theicct.org/publications/ ensuring-driving-on-electricity-cheaper-than-gasoline Energy Information Administration. (2018a). Weekly retail gasoline and diesel prices. Retrieved from https://www.eia.gov/dnav/ Nicholas, M. & Hall, D. (2018). Lessons learned on early electric pet/pet_pri_gnd_dcus_nus_m.htm vehicle fast-charging deployments. Retrieved from the International Council on Clean Transportation, https://www. Energy Information Administration. (2018b) Annual energy theicct.org/publications/fast-charging-lessons-learned outlook 2018. https://www.eia.gov/outlooks/aeo/

WORKING PAPER 2019-01 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION 13 WHEN DOES ELECTRIFYING SHARED MOBILITY MAKE ECONOMIC SENSE?

Parrot, J. A., & Reich, M. (2018). An earnings standard for New Transportation Research Board. (2018). Broadening understand- York City’s app-based drivers: Economic analysis and policy ing of the interplay among public transit, shared mobility, and assessment. (Report for the New York City Taxi and Limousine personal automobiles (TCRP Research Report 195). Retrieved Commission). Retrieved from https://static1.squarespace. from http://onlinepubs.trb.org/onlinepubs/tcrp/tcrp_rpt_195.pdf com/static/53ee4f0be4b015b9c3690d84/t/5b3a3aaa0e2e7 U.S. Department of Energy. (n.d.a). Maximizing electric 2ca74079142/1530542764109/Parrott-Reich+NYC+App+Driver cars’ range in extreme temperatures. Retrieved from s+TLC+Jul+2018jul1.pdf Office of Energy Efficiency and Renewable Energy Ryan, N., & McKenzie, L. (2016). Utilities’ role in transport electri- website: https://www.energy.gov/eere/electricvehicles/ fication: Capturing benefits for all ratepayers. Public Utilities maximizing-electric-cars-range-extreme-temperatures Fortnightly. Retrieved from https://www.fortnightly.com/ U.S. Department of Energy. (n.d.b). Side-by-side comparison of fortnightly/2016/04/utilities-role-transport-electrification- 2017 Toyota Prius, 2017 Nissan Sentra, 2017 Toyota Corolla, and capturing-benefits-all-ratepayers 2017 Chevrolet Cruze. Retrieved from https://www.fuelecon- Salisbury, S. (2016). Cold weather on-road testing of a 2015 Nissan omy.gov/feg/Find.do?action=sbs&id=38061&id=38252&id=381 Leaf. Retrieved from Idaho National Laboratory website: 53&id=37907 https://avt.inl.gov/sites/default/files/pdf/fsev/2015LeafColdWe U.S. Environmental Protection Agency. (2016). Proposed deter- atherTestJune2016.pdf mination on the appropriateness of the model year 2022- Serradilla, J., Wardle, J., Blythe, P., & Gibbon, J. (2017). An 2025 light-duty vehicle greenhouse gas emissions standards evidence-based approach for investment in rapid-charging under the midterm evaluation. Retrieved from https:// infrastructure. Energy Policy, 106, 514–524. Retrieved www.epa.gov/regulations-emissions-vehicles-and-engines/ from https://www.sciencedirect.com/science/article/pii/ midterm-evaluation-light-duty-vehicle-greenhouse-gas#final S030142151730232X UBS. (2017). UBS evidence lab electric car teardown – Disruption Shaheen, S., Martin, E., Hoffman-Stapleton, M., & Slowik, P. (2018). ahead? Retrieved from https://neo.ubs.com/shared/ Understanding how cities can link smart mobility priorities d1ZTxnvF2k/ through data. Retrieved from Transportation Sustainability Union of Concerned Scientists.(2018). Do electric vehicles work in Research Center, UC Berkeley, https://escholarship.org/uc/ cold temperatures? Retrieved from https://www.ucsusa.org/ item/7303t6sw clean-vehicles/electric-vehicles/cold-weather-electric-cars- Slowik, P., & Lutsey, N. (2018). The continued transition to electric batteries#.W37wIJNKjR vehicles in U.S. cities. Retrieved from the International Council Washington State Department of Licensing. (2018). Alternative on Clean Transportation, https://www.theicct.org/publications/ fuel vehicles and plug-in hybrids Washington state tax exemp- continued-EV-transition-us-cities-2018 tions. Retrieved from https://www.dol.wa.gov/vehicleregistra- Sperling, D., Brown, A., D’Agostino, M. (2018). Could ride-hailing tion/altfuelexemptions.html improve public transportation instead of undercutting it? Yuksel,T., & Michalek, J. (2015). Effects of regional temperature on GreenBiz. Retrieved from https://www.greenbiz.com/article/ electric vehicle efficiency, range, and emissions in the United could-ride-hailing-improve-public-transportation-instead- States. Retrieved from Carnegie Mellon University, https:// undercutting-it www.cmu.edu/me/ddl/publications/2015-EST-Yuksel-Michalek- EV-Weather.pdf

14 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION WORKING PAPER 2019-01 WHEN DOES ELECTRIFYING SHARED MOBILITY MAKE ECONOMIC SENSE?

Appendix

Table A1. Summary of regional assumptions for cost of operation calculations

Residential electricity Price DC fast electricity price ($/ FUEL PRICE ($/GAL) ($/kWh) kWh) Number of cold weather Market 2018 2025 2018 2025 2018 2025 days per year Austin $2.37 $3.15 $0.14 $0.15 $0.43 $0.28 17 Chicago $2.82 $3.75 $0.10 $0.11 $0.43 $0.28 139 Denver $2.54 $3.38 $0.08 $0.08 $0.45 $0.33 130 Los Angeles $3.53 $4.69 $0.13 $0.14 $0.41 $0.24 0 New York $2.79 $3.71 $0.09 $0.09 $0.44 $0.30 115 Portland $3.01 $4.00 $0.04 $0.04 $0.38 $0.28 102 San Francisco $3.48 $4.62 $0.13 $0.14 $0.41 $0.24 1 Seattle $3.16 $4.20 $0.13 $0.14 $0.38 $0.28 107

Note: Numbers in the table are rounded.

Table A2. Summary of vehicle assumptions for cost of operation calculations

CONVENTIONAL Hybrid BEV-150 BEV-200 BEV-250 Metric 2018 2025 2018 2025 2018 2025 2018 2025 2018 2025 Real world FE (mpg) 29.9 37.0 52.9 56.7 Electricity use (kWh/mile) 0.28 0.26 0.28 0.26 0.28 0.26 Real world range (miles) 150 150 200 200 250 250 Real world range (kilometers) 241 241 322 322 402 402 Battery pack usable capacity 37 34.5 49.3 46.0 61.6 57.5 (kWh) Maintenance cost (cents/mile) 6.1 6.1 3.7 3.7 2.6 2.6 2.6 2.6 2.6 2.6 Vehicle cost $19,000 $19,675 $23,000 $22,051 $29,336 $23,266 $31,987 $24,728 $34,638 $26,189

Note: Numbers in the table are rounded.

WORKING PAPER 2019-01 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION 15