How to determine diminished value

Step 1: The pre-accident value

Everything scales off this number, so get it right. It is the market value of your car on the day of the crash, in the condition it was in that morning, with its actual mileage.

A guide value from Kelley Blue Book or a similar source is a reasonable starting point, not the answer. It is a national estimate for a car described in broad terms. Local listings for your exact year, trim and mileage band are closer to what a buyer would have paid. More on why a guide value is a starting point.

Step 2: The repair-to-value ratio

Divide the final repair total by the pre-accident value.

A $7,200 repair on a $29,582 car is a ratio of about 24 percent. That single ratio carries most of the signal: a scratch on an expensive car and a rebuilt quarter panel on a cheap one are not the same event, and the ratio separates them.

Our method turns that ratio into a starting loss rate on a curve that rises and then flattens out:

| Repair / value | Starting loss rate | |---|---| | 0% | 2% | | 5% | 4% | | 10% | 6% | | 20% | 9% | | 30% | 12% | | 45% | 16% | | 60% | 20% | | 100% | 25% |

A 24 percent ratio sits between the 20 and 30 percent rows, so the starting rate is about 10 percent of the car's value. The curve is capped at 35 percent and floored at 2 percent, because neither end of the scale is really open.

Step 3: Severity, not just cost

Two repairs can cost the same and mean very different things to a buyer. So the starting rate is multiplied by what the repair actually involved:

• Cosmetic work pulls it down (0.80). Moderate is neutral (1.00). Major is 1.25, severe 1.45.
• Structural repair adds 15 percent. Frame or body-off work adds 8 percent. A deployed airbag adds 8 percent. A welded replacement panel adds 4 percent. Driver-assistance sensor recalibration adds 3 percent.
• All of that together is capped at 1.60, because the multipliers are signals, not a licence to stack.

Step 4: The car itself

Then four adjustments for the vehicle, not the crash.

Age. A one-year-old car loses more, proportionally, than an eight-year-old one: 1.10 at one model year, 1.00 at three, 0.92 at five, 0.82 at eight, 0.70 at ten, 0.50 beyond that.

Mileage is measured against what the car should have, at 12,000 miles a year. Well under: 1.06. About right: 1.00. Well over: down to 0.70.

Segment. Luxury 1.10, electric 1.08, trucks and SUVs that hold value 1.05, mainstream 1.00, economy 0.95.

History. A car with one prior accident on record already carries a discount, so a second one adds less: 0.75 for one prior, 0.55 for two or more. And if the accident is not yet showing on history reports, the number drops to 0.85, because the discount you are claiming has not landed yet.

There is one more, and it cuts the other way: if the repair left visible issues or used aftermarket or used parts, the loss goes up, by 6 to 15 percent.

Step 5: Check it against the market

Now the part that separates an estimate from arithmetic. Find listings near you for the same model and year band, some with an accident on record and some clean, and measure the gap between them.

Rules that keep this honest:

• At least three clean and three accident-history listings from the same source. Fewer than that is an anecdote.
• Listings more than 60,000 miles or more than three model years from your car are thrown out.
• The observed gap is weighted by how many accident listings there are: 15 percent weight on a thin sample, 30 percent on a better one, up to 45 percent. Small samples barely move the answer, and that is deliberate.

Retail listings rarely say how bad the accident was, so the gap they show is treated as a moderate-severity baseline and scaled to your repair before it is blended in.

A worked example

From one of our sample reports — a real market listing used as a test case, with everything else generated by the same engine a customer gets.

| | | |---|---| | Vehicle | 2022 Toyota RAV4 XLE Premium, Corpus Christi TX | | Odometer at loss | 42,468 miles | | Damage | Rear bumper cover and lift gate replaced and refinished, OEM parts, park sensors replaced, pre- and post-repair scans. No structural repair, no airbag | | Final repair total | $7,200 | | Pre-accident market value | $29,582 | | Repair-to-value ratio | about 24% | | Market evidence | accident-history listings priced about 1.9% below clean ones in the qualifying source, weighted 15% | | Estimate | $2,750, in a range of $2,200 to $3,300 | | The insurer's usual formula on the same facts | $887 |

That $887 comes from a 10 percent cap on the car's value ($2,958), then a 0.50 damage modifier, then a 0.60 mileage modifier. Same car, same repair, same day: a difference of $1,863.

Example from a sample report, not a promise.

The range matters as much as the point. Ours runs 20 percent either side of the estimate, because a single number implies a precision that does not exist.

What this method cannot know

Any method that does not say this is selling you something.

• It cannot know what a specific buyer will do. Diminished value is a market average, and your car sells once.
• It cannot see your repair. The severity signals come from the repair estimate, not from an inspection. If the repair left problems a document does not show, the estimate is low.
• The listing evidence is thin by nature. Most accident-history listings do not disclose how bad the damage was, which is why the market gap is weighted down rather than trusted outright.
• The coefficients are our starting assumptions, published so you can argue with them, and due to be recalibrated as real outcomes come in.
• It cannot tell you what the insurer will pay. It tells you what the loss looks like on the evidence.