IFRS 9 in motor finance: used-car prices, exits and loss

The personal contract purchase (PCP) is the most common form of car finance in the United Kingdom, accounting for over half of motor finance agreements in the Financial Conduct Authority’s (FCA) consumer credit data.¹ Under a PCP, the customer pays monthly instalments and, at the end of the agreement, decides whether to pay a final balloon and keep the car or to hand it back.

Under IFRS 9, a lender’s estimate of expected credit losses (ECL) has to allow for these choices, including the customer’s right to end the agreement early. Customers tend to return the car when it is worth less than the balloon, and in those conditions the lender also recovers least when it comes to resale. For a PCP lender, used-car prices therefore determine both how many cars are returned and how much is lost on each.

This article focuses on consumer PCP agreements, a choice driven by the path of used-car prices, which rose sharply through 2021 to a peak in early 2022 and have fallen since.² Agreements written in 2022 had their balloons set when car values were near that peak. Those on four- or five-year terms mature in 2026 and 2027, when prices are lower, so more of these cars are likely to be worth less than their balloons.

Exposure sits at the end

A PCP splits the cost of a car between a deposit, monthly instalments and the balloon. The lender guarantees that the customer can settle the balloon by handing the car back, regardless of the car’s value at that point. The monthly instalments repay the rest of the loan, with interest. The balloon is fixed when the agreement is signed, at the value the car is forecast to have when the agreement ends, so a PCP usually defers a substantial portion of the principal to the end of the contract.

This deferral determines how quickly the exposure at default (EAD) in the ECL calculation reduces over the term. A hire purchase agreement pays the whole advance down in instalments across the term¹, so the balance and the risk attached to it reach zero at maturity. By contrast, a PCP leaves the balloon outstanding on the final day as the largest single cash flow of the agreement. A large exposure therefore remains at maturity, just when a customer whose car is worth less than the balloon can return it and leave the lender with the shortfall.

Exit routes from a PCP

Handing the car back at maturity is only one of the ways a PCP can end. On most loans, an agreement that ends early is simply repaid, so the lender loses money only when the borrower defaults. A PCP agreement can end in five ways.

• Early settlement: the customer pays the outstanding balance and keeps the car.
• Termination under section 99: the customer hands the car back part-way through the term and pays any additional amount section 100 requires.
• Hand-back at maturity: the customer declines the balloon payment and returns the vehicle.
• Balloon paid or refinanced: the customer keeps the car and the agreement runs to term.
• Default: the account meets the lender’s definition of default, which IFRS 9 presumes has occurred by 90 days past due unless the lender rebuts it (paragraph B5.5.37)³, and recovery begins.

Because each exit changes how long the lender is exposed and how much it recovers, IFRS 9 expects the loss estimate to reflect all five. The standard does not mention motor exits by name. It does, however, require lenders to allow for prepayment and similar options in a contract (Appendix A and paragraph B5.5.51)³, and both a section 99 termination and a hand-back at maturity count as such options.

Losses by exit route

Every exit leaves the lender with either cash or the car. When the customer settles early or pays or refinances the balloon, the lender is repaid in cash and makes no loss. After a default, a section 99 termination or a hand-back at maturity, the lender usually gets the car back and sells it. Across these three exits, the difference lies in how much of any shortfall the lender can still recover from the customer. On a default, the customer still owes the full balance and the lender can pursue any amount the sale does not cover, so the loss is modelled much like the loss given default (LGD) on a mortgage or any other secured loan. On the two voluntary exits, the law or the contract caps what the customer owes, so the lender bears the rest of the shortfall itself.

Under section 99 of the Consumer Credit Act 1974, a customer can hand the car back at any point before the final payment⁴, even if they are behind on payments.⁵ Section 100 then limits what the customer owes to the amount needed to bring their payments up to half the total price, plus any arrears, provided they have taken reasonable care of the car.⁵ Customers do not always pay this in full, so the lender loses the balance still owed, less what it collects and the car’s sale proceeds, plus recovery costs.

A hand-back at maturity works in a similar way, without the termination payment. Returning the car settles the balloon under the guarantee, so the lender cannot claim the shortfall from the customer, apart from charges for excess mileage or damage. Its loss is the balloon less what the car sells for, plus the cost of selling it. On both voluntary exits, the loss depends on what the car is worth at the time of termination.

Why averages fall short

Used-car prices are therefore a core input to motor IFRS 9 models. They drive both the customer’s choice of exit and the lender’s loss on it.

However, in the motor ECL models we have seen, customers’ exits are often captured with two averages taken from past years. One is an assumed average life, which sets how long agreements are expected to stay on the book and thus how long the lender is exposed. The other is a flat early-settlement rate, which sets how many agreements are expected to end early. Neither average responds to used-car prices, even though prices drive how many cars are returned and what each one fetches.

When prices fall, more cars end up worth less than their balloons, and more customers are likely to return them than the averages assume. Those cars also sell for less, so the model misses both the extra hand-backs and the lower sale values. When prices rise, the reverse happens. A fixed exit mix therefore understates the loss in a falling market and overstates it in a rising one. Therefore, exit assumptions should follow used-car price forecasts.

Linking exits to a price forecast solves only part of the problem. Because the customer can always settle the balloon by returning the car, they are most likely to do so when it is worth less than the balloon. The lender bears the shortfall in that case and gains nothing when the car is worth more. A model that uses only one central forecast of used-car prices will still understate the loss.

Consider a balloon of £10,000 with two equally likely outcomes, in which the car is worth either £8,000 or £12,000. In the first case the customer hands the car back and the lender loses £2,000. In the second the customer keeps the car and the lender loses nothing. On average, the lender loses £1,000. A model that relies on the central forecast of £10,000 alone is therefore blind to that £1,000 loss.

So the model needs to work out exits and losses under several price scenarios, including a downturn, and weight the results by how likely each one is. Each scenario also needs translating into values for the lender’s own cars, since balloons are usually set with some prudence and a lender’s cars differ from the wider market by age, brand, fuel type and mileage.

What to check in a motor IFRS 9 model

Getting a motor IFRS 9 model right therefore comes down to a handful of questions: what kinds of exits should be accounted for, what drives each one, what each costs the lender, and how all of this moves with macro factors. A lender reviewing its own model can test it against six points.

• Does the model allow for every exit route? A PCP can end in five ways, and each affects the loss differently.
• Does the data include agreements still running? The model needs every agreement’s monthly history, including those still on the book, with the type of each exit recorded.
• Do the exit rates move with used-car prices? Early settlement, section 99 termination and hand-back at maturity should each be modelled separately and linked to forecasts of used-car prices, rather than taken as fixed averages from past years. Default should follow the usual credit drivers, such as unemployment.
• Does the loss on each exit reflect what the customer can still be asked to pay? On a default, the lender can pursue the customer, as on any secured loan. On a section 99 termination, the customer owes at most enough to reach half the total price, plus arrears, and may pay only part, leaving the lender the remaining shortfall and recovery costs. On a hand-back at maturity, only excess mileage or damage charges can be claimed.
• Are exits and losses modelled under several price scenarios? The model should use a range of weighted used-car price scenarios, including a downturn, with appropriate adjustments.
• Is it clear where each loss is recognised? On a loan, a hand-back shortfall at maturity sits outside ECL, as the balloon was optional. On a finance lease, it appears as a write-down of the car’s expected end value. Either way, the model must estimate it.

With the 2022 agreements now maturing, these checks show a finance director the full cost of a falling used-car market, including the part that ECL does not cover.

Frequently asked questions