Residual value sits at the intersection of underwriting and asset finance, and it is frequently mishandled by both. In motor insurance, the residual value of a vehicle determines the maximum indemnity exposure on a comprehensive policy. Underestimate it and you underprice. Overestimate it and the vehicle owner is overinsured against a loss that will settle at a lower market value. Either way, the pricing error originates in the same place: an assumption about the vehicle's current value that was not grounded in its actual, verified condition and history.
Most insurers solve the residual value problem with lookup tables. Make, model, year, trim level, and declared mileage map to a band from a third-party valuation guide. The approach is defensible for new or near-new vehicles where the variance between the best and worst example of a given model is small. For a used car that is four to eight years old, with multiple previous owners and a history that may span more than one country, the variance between the best and worst example of the same make, model, year, and trim level can be substantial.
The lookup table approach prices the average. The actual vehicle may be significantly above or below that average, depending on its specific history.
What Generic Depreciation Curves Miss
Depreciation curves are statistical constructs built on market transaction data. They describe what happens to the average value of a vehicle cohort over time. They are useful for portfolio-level estimation and for setting initial sum insured bands. They are not useful for assessing the specific residual value of a specific vehicle.
The divergence between the cohort average and the specific vehicle grows with age, because the history-dependent factors that create value variation accumulate over time. A four-year-old vehicle with verified single-owner history, documented service intervals, and no incident record occupies a different position in the market from a four-year-old vehicle with four previous owners, inconsistent mileage records, and a structural repair entry. Both are the same make, model, and year. Both index to the same point on the depreciation curve. Their actual market values may differ by 20% or more.
Depreciation curves do not capture this. They cannot, because they describe the cohort distribution rather than the individual vehicle's position within it. A vehicle-specific residual value assessment requires vehicle-specific data inputs.
The Inputs That Drive Vehicle-Specific Residual Value
The factors that most reliably determine where a specific used vehicle sits relative to the cohort average are knowable from registry and history data, provided that data is available at sufficient quality and completeness:
Verified mileage against cohort average. Mileage is the primary driver of depreciation within a cohort. A vehicle with verified mileage 30% below the cohort average for its age carries a meaningful premium over the cohort price. A vehicle with verified mileage 40% above the average sits below it. The qualifier "verified" matters because declared mileage and actual mileage can differ significantly for the reasons we have covered in previous pieces on odometer manipulation.
Ownership count and transfer frequency. Multiple ownership transfers, especially when concentrated in a short period, are correlated with value reduction. A vehicle with four owners in three years is likely to sit below cohort average, controlling for mileage. The mechanism is partly reputational (a vehicle that many owners have chosen to sell is harder to value on other dimensions) and partly practical (each ownership transfer may introduce inconsistencies in maintenance history).
Incident and structural repair history. A vehicle that has had structural damage repaired commands a discount relative to an undamaged equivalent, even where the repair was executed professionally. The discount exists in the market regardless of the repair quality, because potential buyers apply their own uncertainty about the repair to the price they are willing to pay. For an insurer, this affects the realistic settlement value for a total-loss or major repair claim.
Total loss flag. A vehicle that was previously declared a total loss and subsequently repaired and returned to market (a rebuilt or salvage-titled vehicle in US terminology; a category-specific designation in EU markets) has a residual value profile that is categorically different from the cohort average. These vehicles are frequently not identified at the point of insurance application because the declaration in the source country's insurance database is not visible to a registry-only check in the destination country.
How Residual Value Errors Propagate into the Premium
The mechanism by which residual value error affects the premium depends on the policy type. For a comprehensive policy, sum insured is the primary exposure variable: the maximum payout on a total loss is the insured sum, which should reflect the vehicle's actual market value. If the sum insured is set by generic depreciation curve against a market value that significantly underestimates the vehicle's actual value, the policy is underpriced for total-loss exposure.
For motor lenders providing financing against used vehicles, the residual value question is more acute. The loan-to-value ratio at origination depends on the vehicle valuation. If the underlying valuation is drawn from a generic curve that overestimates the vehicle's actual condition-adjusted value, the lender is extending credit against collateral that is worth less than assumed. Downside scenarios, including default with repossession, may yield recoveries below the outstanding loan balance.
Neither of these errors is dramatic at the individual policy level. The issue is that they are correlated: the vehicles most likely to be overvalued on a generic curve are often the vehicles with the least visible history, which are also the vehicles most likely to be involved in adverse claim events. The correlation means the error does not average out across the portfolio.
Residual Value Scoring at the Signal Level
The residual value band we return as one of the eight verification signals in the GoodToKnow API is not a standalone valuation. We are explicit about this distinction because it matters for how the output is used. A residual value band derived from vehicle history signals is an adjustment factor applied to the cohort baseline valuation, not a replacement for it.
The output structure is: cohort band (from the insurer's or lender's standard valuation source), a history-derived adjustment direction and confidence score (above average, within average, below average, with a confidence band), and the specific signal inputs that drove the adjustment. The insurer or lender applies the adjustment to their existing valuation and decides how much weight to give the history-derived signal based on the confidence score.
This approach has two advantages over an integrated valuation approach. First, it does not require us to have market pricing data for every vehicle in every EU jurisdiction, which we do not. Second, it preserves the insurer's or lender's existing valuation workflow and simply adds a history-informed adjustment layer. The signal is additive, not substitutional.
Where the Data Actually Comes From
The honest answer to "where does the residual value signal come from" is: from the same registry and lifecycle data that drives the other seven signals in the API. We are not building a separate residual value model. We are weighting the history signals that are already present in the verified record into a value-adjustment factor.
This is a deliberate architectural choice. A standalone residual value model requires market transaction data to calibrate against. We do not have that data and have not sought to fabricate it. What we have is registry coverage and lifecycle event data. The residual value band is therefore a history-quality signal, not a price signal. It tells the underwriter where on the condition curve this vehicle likely sits, relative to the cohort average. The price implication of that position is for the underwriter to calculate using their own market reference.
We are not saying this is sufficient to replace a full vehicle appraisal for high-value vehicles or specialty cases. We are saying it is a significant improvement over pricing purely on a generic depreciation curve for the large proportion of used-vehicle policies where no physical inspection is performed before policy issuance. That is the realistic comparison point for the EU private motor market, not a full appraiser inspection workflow.
The Practical Case for Including Residual Value in API Output
Motor insurers in the EU have been slow to adopt vehicle-specific residual value signals for a straightforward reason: the data has not been easily available at point-of-quote latency. Residual value feeds from guidebook providers are available, but they are cohort-based, not history-adjusted. Registry data is available, but aggregating and reconciling it across jurisdictions has required custom integration work that most insurers do not want to build and maintain.
An API that returns a history-adjusted residual value band as one output alongside the other verification signals changes that economics. The integration effort is the same as integrating any other VIN verification call. The incremental value is a vehicle-specific input to a variable that currently relies on cohort averages.
For lenders, the case is even more direct: loan-to-value ratios should reflect the vehicle's actual condition-adjusted value, and the history signals we surface are directly material to that calculation. The residual value band makes the lender's collateral assessment less dependent on declared information and more dependent on verified history.
