The Washington PostDemocracy Dies in Darkness

Warren’s false claim that ‘auto dealer markups cost consumers $26 billion a year’

Sen. Elizabeth Warren (D-Mass.) delivered a speech on the "Unfinished Business of Financial Reform." (Video: Senator Elizabeth Warren/YouTube)

“One study estimates that these auto dealer markups cost consumers $26 billion a year. Auto dealers got a specific exemption from CFPB [Consumer Financial Protection Bureau] oversight, and it is no coincidence that auto loans are now the most troubled consumer financial product. Congress should give the CFPB the authority it needs to supervise car loans – and keep that $26 billion a year in the pockets of consumers where it belongs.”

— Sen. Elizabeth Warren (D-Mass.), speech at the Levy Institute, April 15, 2015

Sen. Warren has long objected to Congress’s decision to exempt automobile dealers from the Consumer Financial Protection Bureau, which she helped set up as an adviser to the Obama administration. In a recent speech, she indicated that she will push Congress to eliminate this exemption and allow the CFPB to supervise loans made by car dealers.

We take no position on whether this is necessary — or whether auto dealers should be barred from arranging loans — but we did become curious about a statistic that Warren has frequently cited — that “auto dealer markups cost consumers $26 billion a year.” She referenced a 2011 report released by the Center for Responsible Lending (CRL), which describes itself as a nonpartisan, nonprofit organization that fights “predatory lending practices.”

How accurate is this $26 billion figure and is Warren describing it correctly?

The Facts

Buying a car often involves two negotiations – the price of the car and then interest rate for the financing. (There can also be a negotiation over the value of a trade-in.) Consumers can come into a dealer with a pre-approved annual percentage rate (APR) from a lending institution, such a credit union, or they can get financing from the car dealer. The dealer typically will send an application to various lenders to get quotes.

Whether someone would get a better rate from the auto dealer is open to question, since dealers expect to get paid for their work in arranging the loan. That compensation is known as the “dealer reserve.” But you can haggle over the rate from the auto dealer. “Try to negotiate the lowest APR with the dealer, just as you would negotiate the best price for the vehicle,” the Federal Trade Commission advises.

The dealer has the option of adjusting the APR (to below a bank’s APR) by reducing its compensation in order to facilitate the sale. Dealers also may tout low or even zero percent financing offers for certain model cars. At the same time, that dealer discretion may result in higher rates for people with poor credit (or who do not negotiate well). Consumer advocates and academic studies also say there is evidence of racial disparities in rates that are offered. We are not defending auto sales tactics, but wanted to examine how this $26 billion figure was reached.

The CRL report is based on 2009 data, in the depths of the financial crisis, and has not been updated. Although the report says that the “total rate markup volume” was $25.8 billion, it does not fully explain how that figure was calculated. A chart in the report says the key information about the dealer reserve was derived from an annual survey published by the National Automotive Finance Association, which is a membership organization of lenders who specialize in “subprime” loans to people with poor credit.

That’s right. This is data on the subprime market, a fact that CRL does not disclose in its report. CRL took the data and applied it to the entire auto loan market, even though subprime loans were only one-fifth of the auto-loan market in 2009. Moreover, just 25 of the 175 companies that were in the subprime auto loan market at the time participated in the survey. Four respondents were deemed “large” lenders with more than 80,000 loans outstanding, while 10 respondents had fewer than 10,000 accounts. (In fact, only one company that answered the survey is one of the top 20 lenders.)

“The [NAF] report has never purported to represent the entire auto financing industry,” said Jack Tracey, the group’s executive director. “The 2010 survey reported data from only 25 companies, all in the non-prime financing space, and it would be incorrect to extrapolate such a small sample size to the entire auto financing market.”

Christopher Kukla, CRL senior vice president, acknowledged that the data used in the report is incomplete. “One thing I will say is that the amount of publicly available data is minimal,” he said. “We would be the first to admit that it is not a perfect data set.”

But there’s a bigger problem. The numbers in the CRL report do not match up with the numbers in the NAF survey. The CRL report says the average markup per used car loan was $780, but the NAF survey says it was $280. For new cars, CRL says $494 but the NAF survey lists $477. CRL says the average reserve was $714 but the NAF survey says $330.

We won’t bore you with the math. But calculations based on the data for auto-financed sales in the CRL report and the stated NAF survey estimates for dealer reserves yield a figure of $11.76 billion, or less than half the figure in the report. That may seem like a big number, but remember, it’s based on data that is a subset of a subset of the overall auto loan market, so it’s practically meaningless.

Even more curious, the CRL report asserts, without explanation, that the markup volume increased 24 percent from 2007 to 2009. (Kukla says the 2007 figure was based on a different data set, from the Consumer Bankers Association.) But the NAF survey shows that from 2007 to 2009, the average dealer reserves declined 24.9 percent for new vehicles and 23.3 percent for used vehicles.

We should note that these numbers come from 15 of the respondents, since that’s how few consistently responded to the survey over three years. Kukla claims CRL relied on an average of data derived from all 25 respondents, from the survey’s appendix, “which we believe gives a fuller picture of the industry.”

It’s puzzling how the numbers could change so much with the addition of a few more respondents, especially because all of the large lenders were included in the sample of 15. It is possible a couple of small (under 10,000 accounts) lenders, with hard-to-believe reserves above $2,000, skewed the numbers.

Moreover, none of the NAF report survey questions regarding dealer reserve have responses from all 25 participating companies, as only 15 arranged loans through an auto dealer. Even then, 10 said the dealer reserve was based on a flat fee, not a percentage of the finance charge — which is practice advocated by critics such as Warren.

As an example of how fuzzy estimates of “dealer markups” are, the Consumer Federation of America in 2004 issued a report in which the dealer reserve totaled “hundreds of millions and as much as one billion dollars annually.” Yet somehow, in the CRL report, the number climbed more than 25 times in the space of seven years.

Yet even if the $26 billion number were credible — which it is not — Warren incorrectly characterizes it. She claimed that the study showed “auto dealer markups cost consumers $26 billion a year.” But Kukla acknowledged that this figure includes compensation for dealers who arranged the loans for car buyers.

“We are careful to say that, yes, this included compensation that is likely to be fairly gotten,” he said. “It is fair for dealers to get compensation for the work they do.” (Nevertheless, the group’s Web site includes a calculator, derived from the report, that purports to show the “windfall profits” earned by auto dealers.)

In 1977, the Federal Reserve Bank rejected a proposal that would have required auto dealers, as part of the Truth in Lending Act, to disclose how much of the finance charge they receive for arranging a loan.  “In most instances, the portion of the finance charge which represents the dealer’s participation is not an amount which the consumer could save by obtaining a direct loan from a lending institution,” the board concluded.

A Warren spokeswoman declined to provide an on-the-record response.

The Pinocchio Test

As regular readers know, we hold politicians accountable for the accuracy of sources they cite. There are enough warning flags in the CRL report — such as its lack of transparency on how the $26 billion number was calculated — that would have given a close reader pause. The National Automobile Dealers Association in 2012 submitted to the FTC a lengthy rebuttal of the report that also raised serious questions about the research.

But besides citing a faulty number, Warren misleadingly says it represents “auto dealer markups.” The group that produced the report said that figure includes reasonable compensation owed to car dealers. She earns Four Pinocchios. (Update: CRL posted a response to this column, claiming “errors and omissions,” but failed to rebut the key findings of our report or explain why it failed to disclose that it applied a subsection of subprime data to the entire auto loan market.)

Four Pinocchios

(About our rating scale)

Send us facts to check by filling out this form

Follow The Fact Checker on Twitter and friend us on Facebook