Chicago Fed Insights

Charging Ahead: How Might the Used Car Market Help Increase Consumer Adoption of Electric Vehicles?

November 17, 2022

The motor vehicle industry is facing a momentous paradigm shift away from cars and trucks powered by the internal combustion engine (ICE) toward those powered by electricity. There are a number of environmental policy reasons—notably, the goal to reduce greenhouse gas (GHG) emissions in the transportation sector—helping to drive this technological transformation in the automotive sector. To date, U.S. policymakers have generally focused on raising the share of electric vehicle (EV) sales among new vehicle sales.1 Currently, the official goal in the United States—as set by the Biden administration in August 2021—is for half of new vehicle sales in 2030 to be made up of sales of battery electric vehicles (BEVs),2 plug-in hybrid electric vehicles (PHEVs),3 and fuel cell electric vehicles (FCEVs). Yet, economic factors will very much influence the diffusion of EV technology across the country. For instance, at $66,000, the average price of new EVs in 2021 was $20,000 higher than the average price of all new cars.

In this blog post, we examine how the used vehicle market might serve as the mechanism by which BEVs become more popular and prevalent. To address this question, we expand on a previous blog post that utilized vehicle registration data to explore the used BEV market. In the United States, the overall used car market has historically been two-and-a-half to three times larger than the overall new car market, according to data from the U.S. Transportation Department’s Bureau of Transportation Statistics. So, the used car market can play an important role in the wider adoption of the new automotive propulsion technology.

Data

To study how the used car market has helped popularize BEVs thus far (and could help in the future), we look at over a decade’s worth of vehicle registration data. We utilize data from two sources: 1) the AutoCount database from Experian Automotive and 2) the Wards Intelligence Data Center. The AutoCount database comprises the universe of nonfleet4 vehicle registrations across the entire United States, sourced from state-level department of motor vehicles (DMV) title and registration data. These data are updated monthly. We use data from AutoCount on the make (e.g., Chevrolet), model (e.g., Blazer), model year, odometer reading, registration date (month and year), and new or used indicator of all vehicle registrations from model year (MY) 2010 through MY2021.5 We refer to a make-model of a specific model year (e.g., the 2021 Chevrolet Blazer) as a product for the remainder of this blog post. We start with data for MY2010 because annual BEV sales had broken 5,000 units in MY2011 and we wanted to begin tracking the registration data shortly before that milestone was hit.6 From Wards Intelligence we add information on the vehicle powertrain (gasoline, electric, hybrid, etc.) to the registration data.

Analysis

There are 248 million light vehicle registrations in the 11 model years of our data sample. Those registrations can be matched to 3,078 unique products. Because we cannot observe individual vehicles (by vehicle titles or vehicle identification numbers) over time, we track the changes to all product registrations on a monthly basis. In order to keep track of the aging for each product, we set its age to (month) 1 when its very first new registration appears in the data set; with each passing month since that point, the product’s age goes up by one month. This approach allows us to compare all observed products by their ages. Distinguishing a product’s age—by tracking the number of months since the first new registration—results in 222,724 observations (given that an observation is a product-age).

A new vehicle becomes a used one when it is registered as used. To measure the rate at which new vehicles become used ones, we determine a ratio: For each product we divide its cumulative used registrations observed through month t by all its new registrations ever recorded in our data sample.7 We call this the used vehicle prevalence ratio, as it measures the transition rate from new to used vehicles. Figure 1 shows the weighted average of this measure for all 3,078 products, anchored by age as measured in months from the first new registration of each product. The weights are determined by the number of new registrations per product. This ratio starts out as zero, as only new vehicles are registered initially. Over time, new vehicles turn into used ones—first at an accelerating rate, but then at a decelerating rate. At age 120 months (ten years) the used prevalence ratio settles at just below 0.8.

1. Used vehicle prevalence ratio, by age, model years 2010–21

Figure 1 is a line chart that plots the rate at which new vehicles turn into used ones in our data sample. For each product, we divide its cumulative used registrations observed through month t by all its new registrations ever recorded in the sample. We take a weighted average across all the products at each age (in months)—with the weights determined by the number of new registrations per product.
Notes: See the text for the definition of the used vehicle prevalence ratio for all the products (the make-models of specific model years) in our sample. Age is measured by the number of months since the first unit of a product was registered as new.
Source: Authors’ calculations based on data from AutoCount.

Next, we distinguish the transition rates from new to used vehicles according to the products’ powertrain type (see figure 2). In our analysis of the vehicle registration data, we identify four mutually exclusive categories of powertrain type for the sample’s products: internal combustion engine; hybrid, including plug-in hybrid (see note 3); mixed (which refers to make-models available with two or more powertrain options that cannot be clearly delineated in the data);8 and pure battery electric. We find that the transition rate from new to used vehicles for products with battery electric powertrains is behaving very differently from the transition rates for the products in the other three categories.

2. Used vehicle prevalence ratio, by age and powertrain category, model years 2010–21

Figure 2 is a line chart that plots the rates at which new vehicles turn into used ones in our data sample. It distinguishes vehicles and products by four different types of powertrains: internal combustion engine; hybrid, including plug-in hybrid; mixed (which refers to make-models available with two or more powertrain options that cannot be clearly delineated in the data); and pure battery electric. For each product, we divide its cumulative used registrations observed through month t by all its new registrations ever recorded in the sample. We take a weighted average of these values across the products in each powertrain category at each age (in months)—with the weights determined by the number of new registrations per product.
Notes: See the text for the definition of the used vehicle prevalence ratio, as well as for details on the four powertrain categories, for the products (the make-models of specific model years) in our sample. Age is measured by the number of months since the first unit of a product was registered as new.
Sources: Authors’ calculations based on data from AutoCount and Wards Intelligence.

BEVs transition to the used market more slowly

Figure 2 shows that BEV products do transition from being new to used vehicles at a much slower rate than any of the other types of products. For example, at 60 months (five years) of a product’s age, 55% of ICE vehicles have changed status from being new to used. For BEVs the transition happens much more slowly: Only 14% of BEV products are registered as used at the same age marker of 60 months.

What can explain the significantly longer ownership spells of new BEV products? There are likely several factors. One might be a relatively lower willingness on the part of BEV owners to sell or trade in their vehicles because of their strong personal commitment to the new, cleaner technology. Another might be the fact that improvements to BEVs can be gained simply through software downloads and installations, possibly obviating the need to buy a newer version of the same model. And yet another reason could be lower usage of BEVs (compared with ICE vehicles) by their owners because of “range anxiety,” or the fear of running out of battery power while driving. Lower relative usage of a BEV due to range anxiety reduces its wear and tear and might lengthen the ownership spell. In the next section, we explore this last factor for why BEVs are taking longer to transition to used status.

Exploring one possible explanation for the lack of used BEVs

We can investigate the relevance of vehicle usage in the decision to hold on to a vehicle, given that our data set allows us to observe a vehicle’s odometer reading every time it gets registered. Figure 3 shows the average odometer readings for used vehicles within each of the four powertrain categories at every age (in months) of the products. One can clearly see that used BEVs show significantly lower mileage than all other types of vehicles. For example, at age 120 months (ten years) an average used BEV has been driven for about 50,000 miles—less than half the mileage of an average used vehicle in any of the other three powertrain categories. Our finding of low BEV usage by their owners is consistent with a recent study using a sample of BEVs registered in California.

3. Average odometer readings of used vehicles, by age and powertrain category, model years 2010–21

Figure 3 is a line chart that plots the average odometer readings of used vehicles in our sample. The figure distinguishes vehicles and products by four different types of powertrains: internal combustion engine; hybrid, including plug-in hybrid; mixed (which refers to products available with two or more different powertrain options that cannot be clearly delineated in the data); and pure battery electric. We take an average of the odometer readings for the vehicles within each of the four powertrain categories at each age (in months) of the products. An odometer reading is observed when a vehicle’s registration changes.
Notes: See the text for details on the four powertrain categories for the products (the make-models of specific model years) in our sample. Age is measured by the number of months since the first unit of a product was registered as new.
Sources: Authors’ calculations based on data from AutoCount and Wards Intelligence.

The difference in usage between BEVs and other types of vehicles is rather striking. The observed lower usage of BEVs is consistent with a slower transition rate to used vehicle status. As a result, the used vehicle channel has not nearly been as relevant for popularizing BEVs as it could be. This insight suggests that if the usage deficit for BEVs could be addressed—for instance, by building out the network of charging stations—the resulting higher usage of BEVs could also lead to more transactions in the used vehicle market and, thereby, boost the dispersion of this new technology.

We will conduct further investigations with these vehicle registration data, so watch out for additional posts reporting our results.


Notes

1 One caveat is that the Inflation Reduction Act, which was passed in August 2022, does incentivize purchases of used electric vehicles through tax credits.

2 More details on BEVs—which are also referred to as all-electric vehicles (AEVs)—are provided by Diaz (2020, p. 11). In 2022 the share of BEV sales among new light vehicle (car and light truck) sales in the United States reached 5.2% (year-to-date as of November 2022), according to our calculations based on data from Wards Intelligence.

3 PHEVs are primarily distinguished from other hybrid electric vehicles (HEVs) by the fact that PHEVs can be charged from an external power source and can run without gasoline. For more details on both HEVs and PHEVs, see Diaz (2020, pp. 9–10).

4 Fleet vehicles are those owned by an organization (e.g., a rental car company or governmental agency), not an individual (such vehicles are considered nonfleet). Further details on fleet vehicles are available online.

5 The data we use in this blog post were imported from the AutoCount database on September 23, 2021.

6 General Motors’ EV1 was only available for lease on a very select basis from 1996 through 1999. Tesla started selling its Roadster BEV in MY2008, but the number of sales for that model was extremely small, amounting to 1,329 units over five years, according to data from Wards Intelligence.

7 For example, say that for a specific make-model only 100 new units were sold of a specific model year. Thirty-six months (three years) after the first unit of that product was sold, 40 of the 100 cars are now registered as used. The product’s used prevalence ratio at that point is 0.4 (which is derived by dividing 40 by 100). Notably, given the nature of our data set and our approach to analyzing it, when the same vehicle is registered as used more than once, we would treat it as if it were two or more different vehicles registered as used. Given this limitation, the used prevalence ratio for a particular product could exceed one.

8 For example, the 2015 Ford Fusion was available as a gasoline-powered (ICE) vehicle, as well as a hybrid vehicle (HEV and PHEV). The registration data do not allow us to distinguish between the varieties (with different powertrains) of the same product, such as the 2015 Ford Fusion.


The views expressed in this post are our own and do not reflect those of the Federal Reserve Bank of Chicago or the Federal Reserve System.

Subscribe

Register to receive email alerts when new issues are published.

Subscription Signup

Your request has been submitted. Please tell us more about yourself.

Subscription More Info
Find Publications By:
Find Publications By:
Publication Date
to

Find or Reset
Having trouble accessing something on this page? Please send us an email and we will get back to you as quickly as we can.

Federal Reserve Bank of Chicago, 230 South LaSalle Street, Chicago, Illinois 60604-1413, USA. Tel. (312) 322-5322

Copyright © 2023. All rights reserved.

Please review our Privacy Policy | Legal Notices