Methodology deep dive

This page is the full specification of the Car Cost of Ownership Calculator: every formula it evaluates, how each default was sourced and validated, when those defaults get refreshed, what the model deliberately cannot tell you, and how our approach differs from the two calculators most people compare it against. The datasets themselves are catalogued on our data sources page.

The governing principle

A cost-of-ownership number is only as good as the inputs behind it, and most of those inputs are personal rather than national. So the model is built on one rule: every figure the calculator produces is either arithmetic you can reproduce from inputs stated on screen, or a value from a named public dataset. Nothing in between. When a figure would require a proprietary dataset we cannot show you, we ask you for the input instead of inventing an average.

The formulas

Loan and financing

Monthly payment uses standard amortization: P × [r(1+r)ⁿ] / [(1+r)ⁿ − 1], where P is the amount financed (purchase price − down payment − trade-in + sales tax + fees, where those are rolled in), r is APR ÷ 12, and n is the term in months. Total interest is (payment × n) − P. We assume a simple-interest loan with no prepayment, which is the norm for U.S. auto lending; precomputed-interest contracts (rare, and worth avoiding) behave differently on early payoff.

Depreciation

Depreciation compounds on the previous year's residual: value(n) = price × (1 − rate)ⁿ. The 15% default produces a ~49.6% residual at year five, which lands in the middle of the range you can observe yourself by comparing new prices with three-year-old asking prices. It is a modelling choice, not a measured national statistic, and the guide on measuring depreciation shows how to replace it with a rate derived from real listings for your exact model: rate = 1 − (residual)^(1/years).

Fuel and electricity

Gas: (annual miles ÷ MPG) × price per gallon. Electric: annual miles × kWh per mile × price per kWh. Use the EPA combined rating for your exact year and trim rather than a manufacturer figure, and your own utility rate rather than a state average if you charge at home on a time-of-use plan.

Insurance, maintenance, taxes and fees

• Insurance is a straight annual figure × years. The default assumes full coverage at 100/300/100 limits with $500 deductibles for a driver with a clean record — a coverage definition, not a prediction of your premium.
• Maintenance escalates: a base year-one figure growing ~12% annually, reflecting wear items (tires, brakes, fluids) that cluster in years three to five. EVs take a ~30% reduction on scheduled maintenance and a tire surcharge for curb weight.
• Sales tax applies to purchase price minus trade-in credit in states that allow it, at the rate you enter. Registration is annual and flat by default; value-based states decline as the car depreciates.

How we validate a data source

Before a default or a cited figure goes on the site it has to clear four tests:

1. Primary, not secondary. The publisher must be the body that collects the data — EIA for fuel prices, EPA for efficiency, the state DMV for fees, the IRS for credits. An article citing a study is not a source; the study is.
2. Publicly retrievable. A reader must be able to open the link and find the same number without a subscription. If they cannot verify it, we do not print it.
3. Dated and versioned. The source has to state when the figure applies. Undated statistics get cut, because we have no way to know when they go stale.
4. Cross-checked where possible. Energy costs get a second pass through the DOE Alternative Fuels Data Center calculator; loan arithmetic is recomputed independently before publication, so the interest totals in our examples are exact rather than rounded estimates.

Figures that fail these tests are not softened with "approximately" — they are removed. That is how the site lost its old APR-by-credit-score tables and "cheapest states to own a car" rankings.

Update frequency and process

• Fuel and electricity defaults — quarterly. Re-pulled from EIA's weekly gasoline series and monthly Electric Power Monthly.
• State tax and registration defaults — twice yearly, aligned with the January and July effective dates most legislatures use, plus an out-of-cycle fix whenever a reader flags a change.
• Tax credit rules — on publication of new IRS or Treasury guidance.
• Guide prose — reviewed on a rolling basis, with the "Updated" date bumped by hand only when the content genuinely changes. We do not auto-stamp today's date on unchanged articles; a freshness signal that is always true is worthless.
• Formulas — versioned in the codebase. Any change to how a cost is computed is described here and noted on the news & updates page.

Limitations and assumptions, stated plainly

Where this model will be wrong, and by roughly how much:

• Constant-rate depreciation understates year one and overstates later years. Real depreciation is front-loaded. Over a full five-year hold the totals converge; if you sell at 18 months, the model is optimistic.
• Insurance is your input, not our forecast. Two drivers on the same street with the same car can differ by a factor of two. We will not pretend otherwise.
• Maintenance is a smooth curve over a lumpy reality. Real repair spend arrives as a $0 year followed by a $1,900 year. The five-year total is the useful figure; any single year is not.
• No inflation adjustment. Costs are in today's dollars. Fuel and parts prices will move; we would rather you see nominal arithmetic than a compounded guess.
• No opportunity cost on the down payment, and no modelling of parking, tolls, tickets, at-fault premium surcharges, aftermarket accessories, or gap insurance.
• EV assumptions favour home charging. A driver relying on public DC fast charging can pay several times the residential rate, which materially changes the EV comparison.

How this compares to KBB and Edmunds

Kelley Blue Book's 5-Year Cost to Own and Edmunds' True Cost to Own are both excellent, and both answer a different question than we do.

• They start from a specific VIN or trim; we start from your inputs. KBB and Edmunds apply proprietary residual models and licensed regional cost data to a car they have in their catalogue. That is powerful for a mainstream new model and unavailable for an unusual trim, an older used car, or a private-party purchase — which is exactly where our input-driven approach still works.
• Their depreciation curves are proprietary; ours is a rate you control. You cannot audit theirs. You can audit ours, and override it with a residual you derived from live listings.
• Their insurance and maintenance figures are modelled regional averages; ours are your quote and your assumption. An average premium is a good starting point and a poor budget.
• They do not show the arithmetic; we do. Every intermediate figure here is visible, which is the difference between a number you accept and a number you understand.

The honest recommendation: run your car through Edmunds TCO and KBB as well. If all three land in the same neighbourhood, your budget is sound. If ours differs sharply, the gap is in an input you can now identify — usually depreciation rate or insurance — and that gap is the most useful thing any of these tools will tell you.

Corrections

If you find a figure you cannot verify, or a fee schedule we have out of date, tell us at contact@ynbautomotive.com and we will correct it and note the change. Every output on this site is an estimate for planning purposes, not financial, tax or insurance advice — see our disclaimer.