The auto finance market has seen significant change over the past two years. Car prices have risen substantially, leading to larger loan amounts and higher monthly payments. These more expensive loans are beginning to have an impact on consumer and household financial stability. Recent data show an increase in auto loan delinquencies, particularly for Some consumers may even be getting priced out of the current market. As part of our monitoring the auto loan market for consumer risks, the CFPB is piloting a new collection of auto lending data.
Inquiry into auto lending portfolios
Today, we issued orders to nine large auto lenders to provide information about their auto lending portfolios. Congress has tasked us with ensuring that markets for consumer financial products and services are fair, transparent, and competitive. We routinely ask lenders in different sectors of the market to provide information and data that helps us monitor risks to consumers and to publish aggregated findings that are in the public interest.
These nine lenders represent a cross-section of the auto finance market. The data collected from their responses to these orders will help us build a quality data set that provides insights into lending channels, loan performance, and inform potential future data collection efforts.
We take protecting consumer privacy very seriously and are required by statute to take steps to protect the personal information of consumers. We have taken care to ensure that no directly identifiable information— like name, address, or social security number—is collected as part of this pilot. View a sample order .
Since announcing our intention to build a new auto lending data set last November, we have held multiple discussions with stakeholders and into the auto lending areas most in need of greater transparency. These comments and guidance have helped inform how we structured our pilot data collection.
In December 2022, we convened an auto finance data stakeholder event with market participants, fellow regulatory agencies, other Federal Reserve officials, market analysts, and consumer researchers and advocates to gather input on the current data landscape. The discussions identified three areas where most participants stated that additional data visibility would be important: lending channel differences; data granularity, consistency, and quality; and loan performance trends. These areas are helping guide our pilot efforts.
Lending Channel Differences. Currently, data is generally not broken down by whether the consumer secures financing for the purchase of the vehicle directly with a lender (direct lending) or whether the dealer arranges financing for the purchaser (indirect lending). As a result, it can be difficult—and in some cases nearly impossible—to analyze differences between direct and indirect auto loans.
Data Granularity, Consistency, and Quality. Complete and comprehensive auto lending analyses are nearly impossible because of variations within existing data, the lack of a centralized data source, and the cost and significant burden of combining data sets. For example:
- The variety of lender types in the auto finance market can lead to data gaps. For example, depository institutions are required to submit regular call reports about their activities, while non-depository institutions do not have that same requirement.
- The use of different definitions and terms in various data sources leads to data quality issues. For example, data providers may use different credit score cutoff points when defining credit score tiers (superprime, prime, subprime, etc.). When data sets use different thresholds and data buckets, analysis across data sets is difficult or impossible.
- While some stakeholders may have access to sufficient data, those data sources are often prohibitively expensive, proprietary, and/or only available to certain market participants. Some data sets, even when publicly available, are only useful to individual market participants or small segments of the industry.
Loan Performance Trends. There’s a lack of reliable information on repossessions, including how many days past due a loan typically is before a vehicle is repossessed, how long the consumer has paid on the loan before a repossession, and post-repossession impacts for the borrower and lender. Stakeholders have pointed to a need for more consistent and granular data on delinquency and default trends, specifically the correlation between delinquency and geography, credit score, and income.
Stakeholders have also expressed a desire to gain insight into the kinds of technology used during repossession, like GPS tracking and starter-interrupt devices. Little is known about how often these technologies are used (which raise other privacy, security, and liability concerns), their effectiveness, and their impact on repossession timing.
A common thread from the stakeholder event, the docket comments we received, and other stakeholder discussions is that a fuller understanding of the underlying trends of the market is important for consumers, market participants, and regulators alike. The current data landscape is fragmented and requires significant time and effort to piece together. Most of the participants in our stakeholder events stated that access to high-quality, consistent, and regularly published auto lending data would provide greater transparency and foster a better functioning market.
As we collect data for and conduct this pilot, we will continue to inform the public of our assessment of gaps in auto data, where this data provides new insight into the market, and the next steps in scoping and building an auto finance data set that will help us better understand market trends.