Buy Rate: The Half of the Indirect Auto Price You Actually Set
- Industry Use Cases
dotData
dotData
Key Takeaways
- The buy rate imposed is the one set by the lender, since the dealer’s markup is added to it; each scoring error thus has an impact on the buy rate first.
- Most auto financing runs through dealers: 83% of auto loans are obtained indirectly, according to a paper cited by the Congressional Research Service.
- Overpricing shows up as lost volume, not losses: good borrowers priced too high get routed to other lenders, and look-to-book falls.
- Fix the population’s price, not the dealer program: in an illustrative example, a targeted buy rate cut costs $375,000 a year, against $750,000 for a program-wide increase in dealer participation to save money.
The chief lending officer at a regional lender with a large indirect auto book looks at her near-prime contracts. The pattern is clear: over the past two quarters, contracts from the dealer network for near-prime clients have been falling. Dealer relations has a fast diagnosis and a quick fix: The buy rate is not competitive, so raise dealer participation across the program. The proposal is next up on the agenda, but no one has asked the critical question: Which borrowers have stopped coming?
Every lender that funds through dealers—banks, credit unions, captive auto lenders, independent lending institutions—sets the same number at the same moment in the deal. Call it what the market calls it: the buy rate. Credit score, pricing tier, buy rate. That scoring prices the borrower in the middle of a tier well enough, but borrowers at the edges get the tier’s price instead of their own, before the dealer adds a single basis point.
Instead of changing the dealer program, the buy rate should get a harder look. This post explores what a buy rate is, why it inherits every error in the score, where dealer reserve adds a second layer, and how to find the small but costly populations where the buy rate is wrong. It builds on the larger case for why a rate sheet can be wrong in both directions.
What Is a Buy Rate, and Who Sets It?
The buy rate is the interest rate a lender offers to purchase a dealer-arranged auto contract. Dealers can write the contract at a higher rate, called the contract rate, and keep part of the difference. Lenders set the buy rate; dealers set the markup.
Buy Rate vs. Contract Rate
When a customer starts car buying and decides to pay for their new car by financing their vehicle purchase through a dealer, the finance office submits the application to one or more lenders at once. Each lender, in turn, returns a “buy rate”—the rate at which each lender will purchase the retail installment contract from the dealer. The dealer then writes the contract at the rate of its choosing, within whatever terms and limits the lender allows. The borrower only sees the contract rate.
Most borrowers never learn the buy rate existed. Unless someone builds the report, the lender does not see how far each contract rate drifted from it either.
Why the Buy Rate Carries the Score
Every buy rate is the point where the lender’s credit view becomes a price. Buy rates come off the rate sheet, which means they come off the score and the tier cutoffs behind it—the same two-layer pricing structure that runs through every indirect deal. Whatever the score gets right, the buy rate gets right. Whatever the score gets wrong, the buy rate gets wrong first, and the dealer’s markup is added on top of that error rather than replacing it.
Start by writing down which score version and which tier cutoffs drive each buy rate, split by channel. Producing the rate sheet takes minutes. Finding when each cutoff was last tested against what indirect borrowers actually did usually takes longer, and that test history matters here because the buy rate determines the indirect book’s margin.
Why the Buy Rate Inherits Every Scoring Error
A buy rate prices the tier, not the borrower. When the score underestimates a group’s risk, the buy rate misses the loss. When it overestimates risk, the buy rate is too high, and dealers route those borrowers to a lender that prices them better. The challenge scales with the channel. According to a paper cited by the Congressional Research Service, 83% of car loans are obtained indirectly—so for lenders with a large indirect book, an error in the buy rate runs through most of the portfolio.
Underpriced: The Loss Shows Up Late
Underpricing is the error lenders watch for. If a population the score rates as tier-average defaults faster than the tier, its buy rate never covered the difference. By the time the gap reaches net charge-offs, every contract in that population has already been bought at the old rate. Current data shows why the timing matters: in its Q2 2026 report, the New York Fed put the annualized flow of auto balances into serious delinquency at 3.00%, up from 2.93% a year earlier.
Overpriced: The Volume Disappears Quietly
Overpricing leaves no charge-off behind. A good borrower quoted too high a buy rate doesn’t default; the dealer sends the contract to a lender with a better number, and the application never books. This causes the look-to-book ratio to drop among exactly those borrowers that the lender is aiming to reach. Loss-based dashboards are unable to detect this scenario, and likewise the dealer relations teams, who attribute the decline to dealer compensation and therefore request higher participation budgets. Errors of this kind can continue because the static scorecards adjust their tier cutoffs based on the development samples and release schedules rather than in response to real-time changes in the indirect portfolio. As a result, the buy rate will continue to display outdated pricing.
When examining losses, the look-to-book ratios should be checked according to tier and dealer. A group that has low losses and is showing a falling look-to-book ratio is probably overpriced. Every contract it fails to book is net interest income a competitor now earns, and none of it appears in a loss report. The margin compression that is seen in the total figures can start at this stage.
Dealer Reserve and the Layer the Lender Doesn’t Price
The dealer markup (often referred to as dealer reserve) is the rate added above the buy rate by the dealer and kept as compensation for arranging the financing. The lender does not set the markup directly, but instead establishes a cap through its dealer compensation policy, making the markup an additional pricing layer above the buy rate.
How Participation Caps Work
Caps live in each lender’s dealer agreements, and they vary. Under its consent orders, one large captive lender introduced a new dealer compensation policy on August 1, 2016, capping dealer participation at 125 basis points for contracts of 60 months or less and 100 basis points for longer terms. Within that cap, the dealer decides how much of the spread to take on each contract.
There are two pricing layers, which means there are also two possible sources of error. If the buy rate set for a given population is wrong, the dealer’s discretionary markup can cause the contract rate to move further away from or closer to the population’s real risk. When the data is averaged over the whole network, these effects might be seen as noise, but examining the data by dealer and tier can show patterns. The dealer performance scorecard is a natural place to see this type of analysis.
The Regulatory History
Markup also has a regulatory history. In 2013, CFPB guidance treated establishing a buy rate and allowing discretionary markup as participation in the credit decision under the Equal Credit Opportunity Act (ECOA). Between 2013 and 2016, the CFPB, in conjunction with the Department of Justice, settled four enforcement actions relating to dealer markup policies. In 2018, Congress revoked the guidance in accordance with the Congressional Review Act. ECOA and Regulation B remain in force. For a pricing committee, markup policy is still a matter relating to fair lending—just without the 2013 guidance.
Look at the markup and the level of participation separately for each dealer and tier; a dealer whose contract rates sit further above the buy rate in one tier than in others is showing you a pricing layer the rate sheet never set. Markup dispersion across the dealers you fund is pricing risk on your balance sheet, and it belongs in the same review as the buy rate.
How to Find Where Your Buy Rate Is Wrong
Compare the loss your score predicted with the loss each population actually delivered, and pair it with look-to-book. If a population exhibits low losses alongside a declining volume, it is likely overpriced. Before making any changes to the dealer program, adjust the buy rate for that group. Finding those populations is tedious and slow. dotData’s Signal Intelligence Platform takes in the lender’s application data, the bureau data, and the loan performance data. dotData Core evaluates relational combinations across these tables to rank the Driver Signals that underlie a Registered Metric. Two apps within the platform map to the two directions of buy rate error: Portfolio Health, anchored to 90 days past due, covers the underpriced side, while Origination Funnel, anchored to look-to-book, surfaces yield signals, not just risk signals, which covers the overpriced side.
Each Driver Signal is given in the form of a Glass Box rule, consisting of a condition, the proportion of records to which it applies, and its lift in relation to a specified baseline. Two examples from dotData’s documentation show what this format looks like:
One signal, a specific vehicle model, matches 13% of historical records and raises default risk by 11.5 percentage points above average. Another, customers who shop on weekends, matches 11% of historical records and reduces default risk by 9.5 percentage points below average—the profile an overpriced population takes on when it is priced at the tier average.
Every signal is also given as production SQL, so a validation team can recompute it from source data without dotData software running. The SQL lands in existing post-model adjustments, knock-out rules, and scorecard updates. There is no requirement to rebuild the loan origination system. dotData does not set the buy rate or replace the decisioning engine; the pricing committee still owns the price.
One Lender, One Week
An illustrative example: Go back to the lender from the opening. Its indirect near-prime book holds $300 million of balances, and dealer relations wants a program-wide fix.
Alternative #1 (Program-wide dealer participation increase): The lender considers adding an extra 0.25 percentage point (25 basis points) of dealer participation across $300 million of balances. This proposal costs $750,000 per year, spread across all dealers and all borrowers in the tier, regardless of risk. Cost: $750,000 a year.
Alternative #2 (Targeted buy rate cut): Before voting on the blanket increase, the risk team compares expected and actual losses in the tier alongside look-to-book. One specific group holding $50 million in balances charges off at 1.8% per year, below the 3.0% expected, making its buy rate 1.2 percentage points too high. Its look-to-book has dropped faster than any other group’s, confirming that dealers were right about price being uncompetitive, but only for this specific borrower segment.
Instead of altering the dealer program, the committee cuts the buy rate by 0.75 points exclusively for that $50 million segment. This leaves the rate 0.45 points above what its losses require while making it competitive. The targeted cut costs $375,000 a year—half the cost of Alternative #1—and directs the price adjustment solely to the overpriced population. Cost: $375,000 a year.
Rank each population by its gap multiplied by its balances, then check look-to-book before touching dealer compensation. Program-wide fixes spend money on every borrower; population-level fixes spend it only where the price was wrong.
A Four-Step Buy Rate Review
- Map buy rates to the score: record the score version and tier cutoffs behind each buy rate, split by channel, with the date each cutoff was last tested against indirect performance.
- Measure both directions: compare expected and actual loss by population, in percentage points, alongside look-to-book. Report both on one page so the committee sees them together.
- Separate buy rate from markup: break contract rates out by dealer and tier. That shows which layer moved the price.
- Fix the smallest thing that works: reprice the population before the program, and send each finding to the pricing committee and compliance as a rule they can understand, read, and test.
The Price You Set Is the One You Can Fix
Every indirect contract has two parts, and the buy rate is the lender’s. Because it carries the score, the buy rate inherits every error the score makes—underpricing surfaces as losses, and overpricing surfaces as contracts dealers send elsewhere. Neither the markup nor the dealer program caused those errors.
The answer on the table cost $750,000. Measuring the gap turned it into a $375,000 one: half the cost, spent only on the $50 million of balances the buy rate had priced too high.
Frequently Asked Questions About Buy Rates
What is a buy rate on a car loan?
A buy rate is the rate at which a lender offers to purchase a financing contract from a dealer. As the CFPB explains, the dealer then offers the borrower a contract rate, which may be higher, and the borrower signs at that rate.
What is the difference between a buy rate and a contract rate?
Lenders offer the buy rate to purchase the contract, while borrowers pay the contract rate; the difference between the two is the dealer’s markup on that loan. Dealers keep part of that markup as dealer reserve. Both rates rest on the lender’s loan pricing decisions.
How much can a dealer mark up an auto loan rate?
Each lender’s dealer compensation policy sets the cap. As one example, a large captive lender’s 2016 consent orders limited dealer participation to 125 basis points for contracts of 60 months or less and 100 basis points for longer terms. Caps vary by lender and dealer agreement.
What is dealer reserve?
Dealer reserve (also known as dealer markup) is the compensation a dealer retains from the spread between the buy rate and the contract rate for arranging the financing. Rather than being a separate portion of the markup, dealer reserve is the markup itself, and it increases with each point added above the buy rate. Tracking it by dealer belongs in a dealer performance scorecard.
Who sets the buy rate on an indirect auto loan?
Lenders set it based on their scoring and tier structure, and regulators have treated the establishment of a buy rate as part of the credit decision. As part of the contract rate the lender controls, it is where a pricing review should start.
How do other financial terms compare to auto financing?
Unlike a fixed-rate loan where a borrower receives a single fixed rate directly from a financial institution or a bank, indirect car financing involves intermediary factors. In lease transactions, lenders use a money factor instead of a standard interest rate, but the basic mechanics of markup and spread remain similar. Borrowers looking to qualify for the best interest rate need a strong credit history to get approved and receive fast approval when they apply. When borrowers realize how much they can save in fees or currency exchange scenarios, they are ready to negotiate with the dealer at the end of the month to mark down rates before they hand over any payment.
Find out where your buy rate is wrong. dotData helps lenders identify the populations their scores misprice and delivers each one as a rule the pricing committee can review. Reach out to the dotData lending team to get started.
dotData
