Methodology.

The analytical framework

Drive Economics analyzes the total cost of owning a vehicle over a defined period — typically 5 years — by summing every dollar out and subtracting any incentive dollars in. The six cost categories we model are: depreciation, fuel or electricity, insurance, maintenance, and (where applicable) state or federal incentives. Financing costs are handled separately in the calculators when the user selects loan or lease.

What makes this different from typical online calculators is a commitment to per-user data rather than national averages. When you enter a ZIP code, we look up the actual utility serving that address, identify the specific rate plan most favorable to EV charging (if you're comparing an EV), and blend the time-of-use structure with realistic charging patterns. We do the same for gasoline pricing (PADD-region and state-adjusted), insurance premiums (density-and-MSRP adjusted for metro), and incentive eligibility (state-specific and program-current).

The result is a five-year total that's honest at the individual level rather than technically defensible in aggregate. Consumer decisions happen at the individual level, so that's where the math needs to work.

Depreciation modeling

Depreciation is often the single largest cost of vehicle ownership and the one most calculators handle worst. We use a geometric depreciation model — each year the vehicle retains a fixed percentage of the prior year's value — because that matches real market behavior far better than a linear straight-line depreciation.

The annual depreciation rate is calibrated per make and category, sourced primarily from iSeeCars 2024 retention data with EV-specific adjustments where new-model market data is thin. Representative rates from our current library:

• Toyota mainstream (RAV4, Camry Hybrid): 11–12% annually (best-in-class retention)
• Ford mainstream gas (Escape, F-150): 9–11% annually
• Hyundai/Chevy mainstream: 10–12% annually
• Tesla stabilized (Model 3, Model Y): 12–13% annually (post-2023 pricing corrections)
• Ford Mach-E: 12% annually (per iSeeCars 2024 retention data of ~51% at 5 years)
• Ford F-150 Lightning: 16% annually (early EV depreciation shocks persist)
• New-model EVs (Chevy Equinox EV, Hyundai Ioniq 5): 13–15% annually (conservative pending 5-year real data)

The formula is simple: residual = MSRP × (1 − annual_rate)^years, and the depreciation loss is MSRP − residual.

Where this could be wrong: new-model EVs launched in 2024 or later have essentially no 5-year retention data. We use conservative estimates until real market data emerges, which means EV totals may be pessimistic for models that end up holding value well. When better data arrives, affected case studies are re-verified.

Fuel and electricity costs

Fuel and electricity are where our approach differs most dramatically from national-average calculators.

Gasoline pricing

Gas prices come from the U.S. Energy Information Administration's weekly retail data, organized by PADD (Petroleum Administration for Defense Districts) region. Each state is mapped to its PADD, and the PADD average is adjusted for state-level gas tax variations. This produces meaningfully different numbers than a national average — for example, our Washington state gasoline assumption ($4.30/gal) is roughly 30% higher than our Texas gulf coast assumption ($3.10/gal).

Electricity rates

Electricity is harder because the U.S. has approximately 3,000 electric utilities, each with multiple rate plans. Our approach:

1. The user's ZIP code maps to the utility serving that address (via our internal ZIP-to-utility lookup, built from published service territory data).
2. For that utility, we identify whether an EV-specific rate plan exists (many major utilities offer time-of-use rates specifically designed for EV charging).
3. If an EV rate plan exists, we blend the tiered rates assuming realistic charging patterns — typically 85% overnight off-peak charging, 10% off-peak, 5% peak.
4. If no EV rate plan exists, we use the utility's residential default.

Sample blended rates from our case study library: Xcel Denver at $0.120/kWh (TOU with EV rider), Eversource Boston at $0.240/kWh (high-cost Northeast market), Reliant Dallas at $0.080/kWh (Truly Free Nights plan), SRP Phoenix at $0.060/kWh (super off-peak overnight), Seattle City Light at $0.115/kWh (municipal flat rate).

The 5-year fuel or electricity cost is then annual_miles × (MPG or kWh/100mi) × price × years, adjusted by vehicle-specific efficiency (kWh/100mi for EVs, MPG for gas). Vehicle efficiency numbers come from EPA fueleconomy.gov combined ratings.

Insurance premium modeling

Insurance is one of the hardest costs to estimate accurately without a specific quote, because premiums depend on driver history, credit, marital status, vehicle safety features, and dozens of other factors we don't ask about. We use a density-and-MSRP model that estimates typical premiums for a given ZIP code and vehicle value.

The model uses:

• Metro-level cost-of-insurance patterns (Detroit and Los Angeles substantially higher than Cleveland or Kansas City)
• State regulatory environment (Michigan's no-fault reforms, California's Proposition 103 rate limits, etc.)
• Vehicle MSRP scaling — a $50,000 vehicle costs more to insure than a $30,000 one, all else equal
• EV premium adjustment where warranted (some markets charge slightly more for EVs due to repair cost profiles)

The model is calibrated against published state average rates from Bankrate, ValuePenguin, and NAIC data. It's directionally accurate for our comparative purpose but should not be used as a substitute for actual quotes.

What we don't capture: individual driver profiles, telematics-based discounts, multi-policy discounts, and specific carrier variation. Real-world premium shopping can vary 20–30% below our estimates on either side, and users should treat insurance line items as reasonable defaults rather than binding predictions.

Maintenance cost modeling

Maintenance costs are estimated by vehicle category (mainstream, luxury, EV) and drivetrain type. Typical assumptions from our current library:

• Mainstream gas vehicles: $550–$650 per year (oil changes, fluids, brakes, tire rotations, minor repairs)
• Luxury gas vehicles: $700–$900 per year (more expensive parts, dealer-only service in some cases)
• Hybrid vehicles: Similar to mainstream gas — hybrid systems require some additional service but brake wear is significantly reduced
• Electric vehicles: $400–$500 per year (no oil changes, drastically reduced brake wear, essentially no drivetrain service, but tire replacement is more frequent due to weight and torque)

Case studies apply these baselines with adjustments for known model quirks (some Tesla owners report higher-than-average tire costs; Ford Lightning owners report software-update-related service visits at rates similar to gas F-150s).

Incentive application

EV purchase incentives come from three sources:

• Federal: Historically the Section 30D tax credit ($7,500). This credit expired September 30, 2025 and is currently excluded from all case studies. If reinstated by future legislation, we will re-verify affected case studies immediately.
• State: Programs vary dramatically. Colorado offers $5,000 for new EVs (still active). Massachusetts's MOR-EV rebate is $3,500 (income-limited). Washington offers a sales tax exemption on EVs under $45,000 MSRP. Texas, Arizona, and Michigan have no state EV purchase rebates.
• Utility: Some utilities offer rebates for EV purchases, off-peak charging incentives, or Level 2 charger installation subsidies. These are typically small relative to state programs but worth capturing when applicable.

We apply only currently-active programs, verified against the issuing agency's own documentation. Case studies note the "last verified" date for incentive data, and re-verification happens quarterly.

Limitations and assumptions

Every model has limits. Here are ours, stated plainly:

• New-model EV depreciation: vehicles launched in 2023–2025 have thin 5-year retention data. We use conservative estimates that may prove pessimistic for models that end up holding value well.
• Insurance is estimated, not quoted: our density-and-MSRP model is directionally accurate but individual quotes can vary 20–30% in either direction depending on driver profile and carrier.
• Charging infrastructure costs excluded from base analysis: a Level 2 home charger installation typically costs $500–$1,500 depending on electrical panel and wiring. Case studies note this as a "what could change this" line but exclude it from the base math since some drivers already have infrastructure.
• Resale timing: we assume the vehicle is held for the stated ownership horizon and then valued at the residual. Early sale or extended ownership shifts the depreciation math meaningfully.
• Fuel and electricity prices are current: our numbers reflect data as of the case study's "last verified" date. Rates can shift meaningfully over 5 years, particularly for utility EV programs which are often pilot rates subject to renewal.
• Federal EV credit excluded: the previous $7,500 Section 30D credit expired September 30, 2025. Case studies include this as a sensitivity in "what could change this" but exclude from base math until reinstated.

How we keep data current

Rate schedules, incentive programs, and vehicle specifications become stale fast. Our approach to freshness:

• Quarterly verification cycles across major utility service territories, state incentive programs, and federal policy environment.
• Immediate re-verification when significant policy changes occur (e.g., the federal EV credit expiration in September 2025 triggered a full library update).
• Per-case-study "last verified" dating — every case study shows the quarter its rates and programs were most recently confirmed accurate.
• Corrections prioritized — user-reported data errors get priority routing over feature additions. See the Contact page for how to report a correction.

Case studies where any material assumption has changed are re-dated and updated with a note about what changed. Our commitment is not that data is perfect — it's that we're honest about vintage, corrections happen fast, and users can trust the "last verified" line.