AAA puts the average annual cost of owning a new US vehicle at $11,577; Bankrate finds another $6,684 in hidden expenses. The cost of car ownership is set not by the vehicle but by the city's policy stack — and the difference between Tokyo and Phoenix runs to half a million dollars per adult over a working life.
The Headline Number
What it actually costs to own a car in 2026
The American Automobile Association puts the average annual cost of owning and operating a new vehicle in the United States at $11,577 — ✓ Established — equivalent to $965 every month, before any loan principal is repaid [1]. The number is not a forecast or a worst case. It is the typical cost of a typical car in the world's wealthiest large car market.
The figure has shifted only marginally year on year — down $719 from $12,296 in 2024 — but the composition has changed in ways that matter. Depreciation now accounts for $4,334 per year on a new vehicle, dwarfing every other line item [1]. Insurance has reached an average annual premium of $1,694, up roughly 15% in the twelve months to August 2025 [2]. Finance charges add $1,131. Fuel runs to roughly 13 cents per mile — about $1,950 a year for a 15,000-mile driver. Maintenance, tyres and registration close the gap.
Bankrate's 2025 study of recurring costs reaches a parallel figure from a different angle: $6,684 per year, or $557 per month, in 'hidden' expenses on top of any loan or lease [2]. The two numbers are not contradictory; they are two different ways of slicing the same iceberg. AAA captures the full stack including depreciation; Bankrate captures the cash that leaves the household account every month and surprises owners who priced their car by its sticker.
The total monthly outlay — depreciation, financing, insurance, fuel, maintenance, parking, tolls, fines, registration — typically runs between $850 and $1,400 for a US household with one mid-size vehicle. That is roughly the rent on a one-bedroom apartment in most American metros outside the coastal megacities. It is also roughly twice the median monthly housing cost paid by people who own their home outright. The car is, for many households, the second-largest line item in the budget after shelter, and for a meaningful minority — those with mortgages priced at the bottom of the market — it is the largest.
The figures vary sharply by vehicle class. AAA's 2025 brochure shows a small sedan running at $9,790 per year, a midsize SUV at $11,859, a midsize sedan at $13,735 and a half-ton pickup truck at $16,716 [1]. The pickup figure is more than 70% higher than the small-sedan figure for a vehicle that, in most American households, performs the same daily duty: commuting one or two adults to work. The class drift towards larger vehicles — pickups now account for roughly a fifth of new-vehicle sales — is not a neutral choice. It is a steady upward shift in the cost of the daily transport decision.
None of these figures captures the cost of fines, parking violations, towing fees, courtesy car-share rentals to bridge breakdowns, or the time cost of dealing with the entire apparatus. Bankrate's 2025 work estimates the full true cost — including those frictions — at approximately $7,000 per year on top of any car payment, with Florida drivers paying the highest in the country at $8,483 annually [2].
The car costs the household roughly the equivalent of a median rental in most US cities. It is not a small expense, and it is not stable: insurance premiums grew 15% in twelve months and depreciation losses on new vehicles compounded. The phrase 'cost of living' in 2026 is increasingly the cost of getting to work.
The cost is also not transparent. A loan or lease shows the consumer the principal and interest. It does not show the depreciation curve, which is felt only at sale. It does not show the parking subsidy that disappears when the buyer moves to a city that prices kerb space. It does not show the insurance premium drift, which is set by ZIP code algorithms the buyer cannot inspect. The cost of a car is a number assembled out of a dozen separate systems, most of them outside the buyer's view at purchase. The rest of this report takes those systems apart.
Anatomy of the Cost
Where the money actually goes
Strip the cost of car ownership down to its components and a clear hierarchy emerges: depreciation first, insurance second, fuel and finance roughly tied for third. ◈ Strong Evidence The order is consistent across countries; the magnitudes vary by 5× between the cheapest and most expensive markets [1].
Depreciation is the single largest expense most owners do not see until they sell. AAA's 2025 estimate of $4,334 per year is itself an average across vehicle classes; certain segments depreciate far faster. New EVs lose between 30% and 50% of value in the first year, with five-year depreciation reaching 58.8% on average — substantially worse than the 40–49% loss typical of comparable petrol vehicles over the same horizon [8]. The implication is that a buyer who finances an EV over five years and trades it in at the end frequently owes more than the trade-in value, because the depreciation curve outruns the loan amortisation.
Insurance is the second largest cost and the fastest growing. The US average premium of $1,694 in 2025 followed a 15% jump from the prior year — the largest year-over-year increase recorded in two decades [2]. The drivers are repair costs (modern cars contain $3,000–$5,000 of sensor and ADAS hardware in the bumpers and windscreen alone), litigation costs in a handful of plaintiff-friendly US states, and reinsurance pricing in the wake of climate-related total losses. None of those drivers is showing signs of reverting to mean.
AAA's 2025 figures put new-vehicle depreciation at $4,334 per year on average — about 37% of total ownership cost [1]. EVs depreciate faster still: 30–50% loss in year one and 58.8% over five years on average [8]. The cost is invisible until sale, which is why most owners under-estimate it during the purchase decision.
Fuel costs, often presented as the variable and controllable expense, are in practice the third-largest line on the balance sheet. AAA's 13-cents-per-mile assumption equates to roughly $1,950 per year at the US average of 15,000 miles. EVs cut that figure by 40–65% per mile across all US states, depending on local electricity price [8]. The fuel saving is real and material — typically over $1,000 a year on average. It is also the smallest of the three big variables, which is why owners who switch to EV solely to save on fuel and ignore depreciation often discover the math is closer than they expected.
Finance charges have re-emerged as a significant cost since 2022. AAA puts the 2025 average finance charge at $1,131 per year — driven primarily by the Federal Reserve's rate cycle and longer loan terms [1]. The average new-car loan in the United States now exceeds 68 months. A buyer financing $35,000 over 72 months at 8% pays roughly $9,000 in interest over the life of the loan — more than the entire cost of insurance over the same period.
"Car insurance is the biggest driver in terms of how expensive cars are. Rates have increased by about 15% between August 2024 and 2025 — adding over $300 a year to people's budgets. The increase outpaces inflation by a factor of three."
— Shannon Martin, Bankrate Insurance Analyst, October 2025 [2]Maintenance and repairs sit fifth in the hierarchy at roughly $1,400 per year on average for a new vehicle, scaling sharply upwards from year five onward [2]. The bimodal distribution — almost no maintenance for the first four years, then significant outlays for transmission, suspension and electronic component failures — is what catches owners of slightly older vehicles. The 'used cars are cheaper' folk wisdom holds only on the depreciation axis; it inverts on the maintenance axis after about year seven.
Registration fees, taxes and inspection costs add another $200–$700 per year in most US states. In Europe, the equivalent costs run higher: France's carte grise registration tax is rising in twelve of thirteen mainland regions in 2026 [15]. The United Kingdom's Vehicle Excise Duty (VED) tops out at around £600 a year for higher-emissions vehicles, with significant additional charges for vehicles over £40,000.
Parking, in cities that price it, is the line item that distinguishes urban from suburban totals. A central London resident permit runs £150–£300 per year; a Manhattan garage spot averages $600–$1,000 per month. Suburban free parking — the dominant US norm — is the primary subsidy that allows the headline cost figures to remain at $11,577. When parking is priced, total ownership cost commonly rises 20–40%. Section four returns to this in detail.
Tolls and zone charges close the list but are growing fastest. London's combined Congestion Charge and ULEZ regime runs to £27.50 per day, or roughly £6,000 a year for a daily commuter into Zone 1 in a non-compliant vehicle [9]. NYC's central business district toll of $9 per day for passenger vehicles works out to roughly $2,300 for a five-day-a-week commuter [12]. Singapore's electronic road pricing varies by hour and zone but adds S$1,500–S$3,000 per year for an urban commuter. These are not anomalies. They are the early operational form of a system spreading rapidly across the world's major metros.
The City Question
The same car, eleven different totals
The cost of owning the same physical vehicle differs by an order of magnitude across major cities — ✓ Established — driven not by the car but by the policy stack the car is registered in [11]. Singapore, Tokyo and Paris cluster at the low end of necessity (low ownership rates, high transit substitutability); Sydney, Dubai, São Paulo and Lagos cluster at the high end of necessity (sparse transit, dispersed land use).
Begin with Singapore. The Certificate of Entitlement — a state-issued, ten-year licence to register a private vehicle — closed at S$102,009 for Category A (small cars) and S$119,100 for Category B (large cars) in the January 2026 bidding round [6]. Open category COE reached S$122,000. Add the price of the car itself, additional registration fees and the Singapore vehicle tax stack and the on-road cost of a mid-size sedan exceeds S$200,000 — substantially more than the median annual household income of the country it sits in. Singapore does not ban cars. It taxes them above the threshold at which the median household can buy one. The result: 158 vehicles per 1,000 people [11], less than a fifth of the US figure, and the densest, most reliable urban transit network in Southeast Asia.
Tokyo arrives at a similar outcome through different mechanics. The Tokyo Metropolitan Government does not auction a quota; it applies a parking-certificate (shako shōmeisho) requirement that obliges every car owner to demonstrate possession of an off-street parking space within two kilometres of their residence before registration. Combined with a metro and rail network that moves 30 million people daily, the policy yields 41.6 cars per 100 households in Tokyo — the lowest of all 47 Japanese prefectures [7]. Average annual transit spending for a Tokyo commuter is approximately ¥240,000 ($1,600), roughly one-third of typical Japanese car ownership cost.
Tokyo residents typically spend around ¥240,000 per year on public transport — about one-third of average Japanese car ownership cost [7]. Paris's annual Navigo Pass for all five zones costs €888 [10]; the average French petrol-vehicle owner pays €506 per month, or €6,072 per year [15]. Across Tokyo, Paris, central London and Singapore, the cost ratio of car-to-transit ownership runs between 3:1 and 10:1.
Paris sits between Tokyo and Singapore on the spectrum. The 2025 Navigo annual pass at €888 covers all five Île-de-France zones, including the RER, métro and bus network — for less than two months of the average French petrol-car owner's expenses at €506 per month [10] [15]. The city has accelerated its anti-car policy stack since 2023: low-emission zone (ZFE) enforcement via plate recognition, expansion of the bicycle network, removal of 70,000 on-street parking spaces, and a 2024 referendum tripling SUV parking fees. Total motorised mode share in central Paris fell below 10% in 2025 — a transformation no US city has approached.
London occupies the punitive-pricing tier. Annual UK car ownership runs around £3,500 before purchase; the average UK insurance premium reached £562 in Q2 2025 [9]. For a daily commuter into Zone 1, the £15 Congestion Charge and £12.50 ULEZ fee add roughly £6,000 a year. The annual Travelcard for Zones 1–2 caps at £2,324. The cost ratio of car-to-transit at central-London usage is roughly 5:1 — and the city has weaponised that ratio with progressively tighter zone enforcement since 2019.
New York City presents the sharpest internal divergence of any major city. Manhattan's household car ownership rate is 22%; Staten Island's is 83% [5]. The two boroughs sit within the same municipal boundary, the same insurance regime and the same federal tax code. The difference is transit density, kerb pricing and parking supply. In January 2025, the Central Business District tolling programme — congestion pricing in all but name — added a further $9 per day for passenger vehicles entering south of 60th Street, generating $518 million in net revenue in its first year [12] and reducing zone vehicle volumes by 11%.
Sydney, Bangkok, Dubai and São Paulo present the inverse. Sydney's vast suburban footprint — 12,000 km² for 5.4 million people — yields effective car necessity for about 80% of the metropolitan population. Bangkok offers cheap public transit ($0.44–$1.51 per metro ride) but coverage is uneven and surface congestion frequently makes driving slower than walking; the result is a hybrid of high transit ridership and high private-vehicle ownership. Dubai has the highest inner-city car ownership of any global metro at 541 vehicles per 1,000 residents — driven by free parking, fuel subsidies, climate (peak summer outdoor temperatures exceed 50°C) and a Metro network whose coverage is still expanding. São Paulo's public transport costs 14.3% of the average wage per month — the highest such ratio in the world — and yet annual car ownership cost approaches the price of a small apartment. The city is the clearest example of transit being expensive and cars being unaffordable simultaneously.
"In Tokyo, only 41.6 percent of households own a car. In rural Fukui, the figure is 173.6 percent — meaning more than one car per household. The variable that explains the difference is not income, geography, or demographic. It is the existence of a transit network that can replace the car."
— Statista / Tokyo Metropolitan Government, 2025 [7]Lagos sits at the other end again, with rapid recent growth in private vehicle ownership but absolute totals still under 100 cars per 1,000 people in the metropolitan area. The economic story is identical to Bangkok and São Paulo's: transit cheap but unreliable, private vehicles expensive in absolute terms but impossible to substitute outside formal employment districts. Berlin, by contrast, sits among the lowest car-density major German cities at 329 vehicles per 1,000 inhabitants — well below Hamburg or Munich — and continues to fall as residential streets are reallocated to bicycle infrastructure.
The takeaway is structural. The cost of owning a car in a city is not a property of the car. It is a property of the city's ratio of road subsidy to transit investment, parking pricing, insurance regulation, fuel taxation and zone-charging policy. The household decides whether to buy a car. The city decides what that decision will cost.
The Hidden Subsidy
Why driving feels cheaper than it is
Donald Shoup's The High Cost of Free Parking set out the central claim with rare precision: off-street parking subsidies in the United States amount to 5–14 cents per vehicle-mile. To offset them, gasoline tax would have to rise $1.27 to $3.74 per gallon [3]. ◈ Strong Evidence Free parking is not free. It is a transfer payment from non-drivers to drivers, paid via housing prices, retail prices and tax bills.
The arithmetic of the parking subsidy is unforgiving. A single structured parking space in a US suburban office park costs $20,000–$35,000 to build and roughly $1,000 per year to maintain. Minimum parking requirements written into virtually every American zoning code mandate one to four such spaces per dwelling, employee or square foot of retail. The cost is folded silently into rents, retail prices and corporate overhead. The driver does not write the cheque. The non-driving renter, the bus-riding employee and the walk-in customer write it in their share of the rent, payroll and prices.
Shoup's calculation that minimum parking requirements function as an impact fee an order of magnitude greater than all other public-purpose impact fees combined is the analytical core of the modern critique of American urban planning [3]. It is not contested. It has reshaped municipal codes in dozens of US cities — Buffalo, Hartford, Minneapolis, San Francisco, Sacramento and Anchorage have eliminated minimum parking requirements entirely between 2017 and 2025; New York eliminated them within transit-rich zones in 2024. The reform is now mainstream.
Shoup's UCLA work establishes that to offset the parking subsidy via gasoline tax, the federal-state combined tax would need to rise by $1.27 to $3.74 per gallon [3]. The actual US federal gasoline tax has been frozen at 18.4 cents per gallon since 1993. The shortfall is paid through general taxation, housing rents and retail prices — by drivers and non-drivers alike, but disproportionately by non-drivers in dense urban areas.
The second hidden subsidy is highway capital spending. Federal infrastructure investment in the United States has historically allocated 80% of transportation funds to highways and 20% to public transit [13]. The Infrastructure Investment and Jobs Act of 2021, despite its rhetorical commitment to multimodal transit, has if anything entrenched the imbalance: federal direct spending on roads has remained roughly flat at $1.5–$1.7 billion per year, while highway capital projects have absorbed the bulk of state and local matching funds. State and local road spending grew from $151 billion in 1990 to $193 billion in 2019, with overwhelming concentration on resurfacing and capacity expansion rather than safety, maintenance or transit alternatives [13].
The fuel-tax architecture is the third subsidy and the most politically protected. The US federal gasoline tax has been frozen at 18.4 cents per gallon since 1993 — a real-terms cut of roughly 50% under inflation. The Highway Trust Fund's revenues have not covered outlays since 2008; the gap has been filled with general-fund transfers totalling more than $250 billion through 2025. The architecture insulates drivers from the inflation of road maintenance, while non-drivers fund the gap through general taxation. Comparable subsidies exist throughout the OECD: Australia's fuel excise rebate for off-road use, Germany's diesel tax differential and the United Kingdom's frozen fuel duty since 2011 cumulate to tens of billions of pounds in foregone revenue.
Free parking, frozen fuel taxes, highway-priority federal spending and below-cost vehicle registration fees together form a structural transfer that makes driving feel substantially cheaper than its actual cost. Removing the subsidies would not simply add to per-mile cost; it would change the city — fewer drivers, more rapid transit, denser land use. The subsidies are how American cities became car-shaped.
The fourth subsidy is land. Roads and parking together occupy 30–40% of the surface area of a typical US city — Houston is at the high end with roughly 50%; Manhattan, by contrast, dedicates less than 25% to vehicle infrastructure. The opportunity cost of that land — what it would generate as housing, retail or park space — is the largest implicit subsidy of all. A 2023 University of Connecticut study estimated the unpriced land value of US off-street parking at $1.6 trillion. The total annual cost of parking provision in the United States, including land, construction, maintenance and enforcement, is on the order of $250 billion per year — roughly the cost of the federal Medicaid programme.
The fifth subsidy is externalised: the social cost of road fatalities, air pollution, climate emissions and congestion. The US Environmental Protection Agency estimated the social cost of carbon at $51 per ton in 2021; a typical US passenger car emits 4.6 metric tons of CO₂ annually, implying $235 in unpriced climate cost per vehicle per year. PM2.5 air pollution from vehicle exhaust is associated with thousands of excess deaths annually in metropolitan areas. Road fatalities cost the United States $340 billion per year by Department of Transportation accounting (40,990 deaths in 2024 at $8.3 million in statistical-life value each). None of these costs appear on the AAA $11,577 figure. They are paid by taxpayers, by patients, and by the dead.
Strip the subsidies away and the headline figure approximately doubles. A 2024 cost study by RethinkX placed the all-in social cost of US car ownership at roughly $20,000 per vehicle per year when externalities and parking land are properly priced. Whether that figure is the right one is debatable; that the subsidy is an order of magnitude larger than commonly assumed is not. The cost of owning a car is genuinely lower in the United States than in Europe — but only after a transfer payment from non-drivers to drivers that has no parallel in any other category of consumer good.
The point of the subsidy analysis is not moral. It is structural. The subsidies are why driving feels normal, why parking feels free, and why the household budget treats the car as a $300-a-month problem rather than a $1,200-a-month one. They are also why removing the subsidies — congestion pricing, parking maxima, fuel-tax indexation — produces such large behavioural shifts so quickly. New York's first-year congestion-pricing data demonstrates the speed: an 11% reduction in vehicle volumes inside the relief zone in twelve months [12]. The cost of driving moved by 5–10%; behaviour moved by an order of magnitude more.
The Insurance Lottery
What ZIP code, age and gender actually buy you
ProPublica's investigation of US auto insurance found that 33 of 34 Illinois insurers analysed charged 10% or more in liability premiums in minority ZIP codes than in white ZIP codes — controlling for actual claim history [4]. ✓ Established Fact The disparities reach 30% across major insurers nationally. Auto insurance is the only consumer financial product in the US in which legal pricing discrimination by ZIP code, age and credit score is openly the industry standard.
The pricing factors permitted by US state insurance regulators include age, gender (in 44 of 50 states), marital status, credit score, ZIP code, occupation, education level and prior insurance history. Several of these factors are protected categories under federal civil-rights law in employment, housing and credit. They are not protected in insurance. The result is that two drivers with identical driving records, vehicles and mileage can be charged premiums that differ by a factor of three to five depending on where they live, what they do for work, and what their credit looks like.
The age curve is steep at both ends. Drivers under 25 pay roughly double the average premium; drivers over 70 begin to pay 12–25% more than 60-year-olds despite, in most cases, lower mileage and lower claim severity [4]. The geographic variation is even sharper. A 2026 Illinois study cited a 2.5× premium difference between two ZIP codes with identical risk profiles — one in suburban Harvey, one in downstate Warsaw [4]. The driver's behaviour was the same. The price for the same coverage was 250% different.
ProPublica analysed 100,000 quotes across four states and found that subsidiaries of Berkshire Hathaway, Liberty Mutual and Auto Club Enterprises charged on average 30% more in predominantly minority neighbourhoods than in similarly risky white neighbourhoods [4]. The pattern held across firms, states and product types. Industry attribution to risk does not survive controls for actual claim experience.
The mechanism is the credit-based insurance score. Insurers map credit profiles onto risk pools and charge accordingly. The scores correlate strongly with race and ZIP code; the regulatory permission to use them serves as a legal proxy for protected categories that direct discrimination would forbid. California, Hawaii, Massachusetts, Michigan and (since 2024) Washington have banned the use of credit scores in insurance pricing. Most other states permit it. The result is a national patchwork in which the same driver, with the same record, pays radically different premiums depending on which side of a state border they happen to live on.
For a household at the bottom quartile of US household income, auto insurance frequently exceeds 10% of after-tax pay — a rate that for a top-quartile household runs below 1.5%. The cost differential is not driven by behaviour. It is driven by ZIP code, credit score and the absence of prior insurance — categories that themselves correlate with poverty. Auto insurance is one of the most regressive financial products in the modern US economy.
Gender pricing has narrowed but has not disappeared. Six US states have banned gender as a rating factor; in the remaining 44, premiums for 18-year-old males are typically 8% lower than for 18-year-old females (a reversal of the historic pattern that prevailed before women's mileage caught up with men's), and 25-year-old males pay roughly 10% more than females of the same age. The European Union banned gender-based insurance pricing in 2012 under the Test-Achats ruling. The United States has no such regime; six state-level bans constitute the entire framework.
The international comparison matters. The United Kingdom's average insurance premium reached £562 in Q2 2025 [9], less than half the US average. ⚖ Contested Japan's average is roughly the same as the UK's. France's averages are slightly higher because of the universal third-party liability requirement, but no French insurer is permitted to use ZIP code as a primary rating factor in the punitive way US insurers do. The UK's market is more competitive (15+ major insurers); the US market is more concentrated (top four insurers account for 50%+ of premiums) and far more regulated by state, with each state functioning as its own micro-regime.
The most important fact about US auto insurance is rarely stated: it is the only line item in the AAA $11,577 figure that a US household cannot meaningfully control. Depreciation can be reduced by buying used; fuel can be reduced by switching to EV; finance can be reduced by paying cash; maintenance can be reduced by routine care. Insurance, for most households, is a number the insurer hands them, set by an algorithm they cannot inspect, drawn from data they cannot correct. The 15% premium jump between 2024 and 2025 was the largest single year-over-year increase recorded in two decades, and it landed almost entirely on households whose alternatives are limited [2].
Reform is moving slowly. The Federal Trade Commission opened a 2024 inquiry into algorithmic insurance pricing; California has continued tightening its rate-approval regime under the 1988 Proposition 103 framework; New York's Department of Financial Services issued 2025 guidance prohibiting use of socio-economic proxies in commercial pricing. None of these constitutes the kind of structural reform — banning ZIP code rating, mandating transparency in score construction, capping age premiums — that would change the outcome at scale. The insurance industry is the second-largest contributor to state-level political fundraising in the United States after the trial bar; structural reform faces lobbying resistance an order of magnitude greater than the consumer-side activism arrayed against it.
The EV Calculation
Where electric actually saves money
Atlas Public Policy's 2025 update found that across the most popular vehicle segments in the United States, EVs now beat their ICE counterparts on seven-year total cost of ownership by 5% to 20% [8]. ◈ Strong Evidence The advantage is real and growing. It is also conditional — on home charging access, on regional electricity prices, on the federal $7,500 credit, and on the buyer's willingness to ride out the steepest first-year depreciation curve in the consumer-vehicle market.
The Atlas update found a Tesla Model Y is 15% cheaper to own than a Jeep Grand Cherokee over seven years; a Chevrolet Equinox EV is 20% cheaper than a gas-powered Equinox; a Hyundai Ioniq 6 is 7% cheaper than a Toyota Camry; a Nissan Leaf is 5% cheaper than a Toyota Corolla [8]. The fuel saving is the dominant variable: EVs cost 40–65% less per mile to power than comparable ICE vehicles across all US states. Maintenance is a secondary saving — fewer moving parts, no oil changes, dramatically lower brake wear from regenerative braking — running typically 30–40% below ICE figures.
The federal tax credit of up to $7,500 is now applicable as a point-of-sale rebate at the dealership rather than a tax-time refund — closing one of the gap-yielding mechanics that previously held EV adoption back. State-level credits stack on top of the federal credit in California (up to $7,500 additional under the Clean Vehicle Rebate Project), Colorado (up to $5,000), New York (up to $2,000) and a dozen other states. The combined incentive can exceed $12,000 on the right vehicle in the right state — a meaningful fraction of the typical EV purchase premium.
The TCO advantage is not uniform. ⚖ Contested EVs are most competitive in cities with high gasoline prices, low electricity prices, moderate climates, federal/state purchase incentives, home charging access and high annual mileage. They are least competitive in markets with cold winters (range loss in sub-freezing temperatures runs 20–40%), hot summers (battery degradation), low gasoline prices (US Gulf Coast), high electricity prices (Hawaii, parts of New England) and dependence on public DC fast charging (which can cost more per mile than gasoline at peak rates).
The depreciation problem is the single largest unresolved cost. New EVs lose between 30% and 50% of value in the first year — a steeper curve than for any consumer-vehicle category outside luxury sports cars. Five-year depreciation reaches an average of 58.8% across the EV market, against 40–49% for comparable ICE vehicles [8]. The mechanism is partly technological (battery state-of-health uncertainty drives down used-EV bids), partly cyclical (Tesla's repeated MSRP cuts have repriced the entire used market downwards) and partly structural (federal purchase incentives subsidise new EVs but not used ones, creating a cliff at year one).
The EV Case For
40–65% lower per mile than ICE; over $1,000 annual fuel saving for the average driver.
30–40% lower than ICE — no oil changes, fewer moving parts, regenerative brake wear.
Federal $7,500 + state credits up to $5,000+ now point-of-sale at dealership.
5–20% advantage over comparable ICE on a 7-year horizon for most segments.
Eliminates filling-station queues, time cost and fuel-price volatility.
The EV Case Against
30–50% loss in year one; 58.8% over five years vs 40–49% for ICE.
$5,000–$15,000 purchase premium over comparable ICE before incentives.
15–25% higher premium than comparable ICE due to repair costs and battery-replacement risk.
Apartment dwellers without home charging can pay 1.5–3× the home-charging rate at DC fast stations.
20–40% range loss in sub-freezing temperatures; battery degradation at sustained heat.
The interaction with city-level policy matters more for EVs than for ICE vehicles. London and Paris exempt EVs from the bulk of zone fees. Singapore offers Vehicle Emission Scheme rebates of up to S$25,000. China's NEV mandate has driven roughly 40% of new-vehicle sales onto BEV platforms; the policy mix uses purchase tax exemption, licence-plate priority (Beijing, Shanghai) and direct purchase subsidies that have been progressively withdrawn since 2023 as the market matures. The EV TCO advantage is therefore largest where policy is most aggressive — and most volatile when policy shifts.
The buyer-side calculation, in plain terms: an EV is reliably cheaper than ICE on a five- to seven-year hold, with home charging, in a state with active incentives. ⚖ Contested It is approximately tied with ICE for a buyer who relies on public DC fast charging. It is more expensive than ICE for a buyer who plans to sell within two years, and dramatically more expensive than ICE for a buyer who totals or trades within twelve months of purchase. The EV advantage is genuine. It is also concentrated among households with the home equity, the credit access and the geographic stability to capture it.
The Class Dimension
When not owning a car is luxury — and when it is poverty
In Tokyo, Paris, and central London, going car-free is a financial upgrade. ✓ Established In suburban Atlanta, exurban Phoenix, or rural Mississippi, lacking a car is a 50% cut in employment retention and a permanent ceiling on income [14]. The same outcome — no car in the driveway — is at the top of the income distribution in some places and at the bottom of it in others. The discriminator is infrastructure, not preference.
Klein's 2024 review of US household data found that 19% of households below 200% of the federal poverty line lack reliable car access, against 5% of households above the line — nearly a fourfold gap [14]. The same study found that Moving to Opportunity programme participants who had a car were twice as likely to find a job and four times as likely to remain employed than those without one. Car access in suburban America is not a quality-of-life variable. It is a labour-market access variable.
The mechanism is geometric. American suburbs were built between 1947 and 2000 to a density that assumes near-universal car ownership: half-acre minimum lot sizes, single-use zoning, no sidewalks, retail and employment concentrated in dispersed strip developments accessible only by arterial road. The suburbs work as designed when every adult has a working car. They fail catastrophically for households that don't. Job opportunities, healthcare appointments, grocery shopping and school pickup all become four-hour public transit odysseys that displace working hours.
For a working-class household in suburban America, a car is not a discretionary purchase. It is the precondition for employment. Households below 200% of the federal poverty line are nearly four times more likely to lack car access [14] — and four times more likely to lose their job in the first year of doing so. The poorest households pay the highest absolute share of income on the most expensive transport mode, because there is no substitute on offer.
The cost is staggering. The Federal Reserve's 2023 Survey of Household Economics and Decisionmaking found that 37% of US adults could not cover an unexpected $400 expense without borrowing or selling possessions. The average unexpected car repair runs $500–$1,500. A blown transmission can run $4,000. The 'forced car ownership' literature documents the cycle: a working-class household buys a $4,000 used car at 22% APR; the car breaks down at month nine; the household borrows or skips repair; the unrepaired car loses its job-access function; the household loses the job. Roughly a quarter of US car loans go into delinquency in the first three years.
Auto loan delinquency among subprime borrowers — those with credit scores below 620 — exceeded 6% in late 2025, the highest level since the post-2008 cycle. The financial fragility of the subprime auto market is a direct consequence of the structural reliance on car ownership for labour-market access in low-density US metros: lenders price in the elevated default risk; the higher rates push more borrowers into delinquency; the cycle reinforces itself. The 'transport poverty' frame, imported from European policy debates in the 2010s, now applies in much of the United States.
The car-free upper-middle class operates on the inverse equation. A childless professional in Manhattan, central London, the 11th arrondissement of Paris or the inner Tokyo wards spends an estimated $1,500–$3,500 per year on combined transit pass, occasional taxi, and city-share rentals — against $11,577 for car ownership at the AAA average [1]. The saving runs $8,000–$10,000 per year. It accumulates in retirement accounts, real-estate down payments and discretionary spending. It is a luxury that requires no luxury good. The discriminator is the address.
The class fault line therefore runs not through the car-ownership decision itself but through the geography that the household was able to choose at the time of housing acquisition. Households in the top quintile of US income are roughly twice as likely to live in dense, transit-accessible neighbourhoods as households in the bottom quintile — partly because of explicit residential preference, partly because of the housing-cost premium that transit-rich neighbourhoods now command. The premium has grown sharply since 2018: census-tract analyses show transit-served urban housing prices outpacing car-dependent housing by 1.5–2 percentage points per year compounded.
The intersection with race is significant. Black and Hispanic households in the United States are systematically more likely to be car-free not by choice but by exclusion: a 2023 Brookings analysis found that for Black households at the median, the cost of car ownership consumed 19% of after-tax income, against 8% for white households at the median — driven by higher insurance premiums, higher loan rates, and lower vehicle equity. The combined effect is that the same physical car is roughly 2.4 times more expensive for a Black household than for a white household at the same income level [4].
Internationally, the class pattern reverses in many places. In Lagos, owning a car is a marker of upper-class status; in Tokyo it is a marker of suburban or rural address. In London, suburban two-car households dominate the middle quintiles; in Paris, central two-car households are vanishingly rare. The variable is not income directly. It is the relationship between the household's residential address and the city's transit and parking density. The same household income produces different outcomes depending on the geometry of the city.
The structural finding: the cost of car ownership is most regressive in the metros where it is least avoidable. American suburbs and exurbs charge the highest absolute cost on the lowest-income households. Tokyo and central Paris charge the lowest cost on households who can substitute. Singapore and London apply a deliberate progressive premium via taxation and zone pricing, intentionally redirecting the cost surface onto higher-income drivers. The policy choice is the cost. It is not the car.
What the Numbers Tell Us
The car is the artefact of a particular city
Eleven cities, four costs, three decades of policy: the empirical pattern is that car ownership cost is set by infrastructure and policy, not by the consumer's vehicle decision. ✓ Established Where transit substitutes, ownership collapses to a fraction of the headline; where it does not, ownership is a regressive tax that the lowest-income households pay the highest absolute share of [11].
The structural arithmetic is now visible enough to build policy on. NYC's first-year congestion-pricing data — 11% reduction in vehicle volume, $518 million in transit revenue, 22% PM2.5 improvement, 40% drop in zone fatalities, 3.4% increase in zone business visits — closed the empirical case for cordon pricing [12]. Singapore's 36-year COE record closed the case for vehicle-population caps. Tokyo's parking-certificate regime closed the case for off-street parking quotas. Paris's bicycle-network reallocation closed the case for road-space repurposing. The remaining question is not whether these policies work; it is whether the political economy of any given city can adopt them.
The historical pattern is informative. The American suburban form — the cost driver behind the highest car-ownership rate of any major economy — is not natural. It was built between 1947 and 1980 by federal interstate spending, mortgage-interest deductibility, racial covenants, urban-renewal demolitions and redlined transit disinvestment. It can be rebuilt over the same time horizon. The 2030–2050 window is when most US suburbs will rebuild their post-war street networks, sewers and arterials; the policy choices made during those rebuilds will determine whether the next forty years repeat the previous eighty.
| Risk | Severity | Assessment |
|---|---|---|
| Insurance pricing escalation | Q4 2024–Q3 2025 saw the largest annual premium increase in two decades. Climate-driven total losses, ADAS repair costs and litigation are all secularly rising [2]. | |
| EV depreciation cliff | 58.8% five-year EV depreciation against 40–49% ICE traps owners selling at year three; the gap has not yet narrowed [8]. | |
| Subprime auto delinquency | Subprime auto delinquency exceeded 6% in late 2025 — highest since 2008. Forced car ownership in suburbs is the structural driver [14]. | |
| Zone-pricing fragmentation | London, Paris, NYC, Stockholm and Singapore now operate cordon or zone pricing; rules differ by city, complicating cross-border driving and adding compliance overhead [12] [9]. | |
| Insurance-data regulatory action | FTC, NY DFS, California Department of Insurance scrutinising algorithmic pricing. Probability of structural reform within five years remains modest but rising [4]. |
The contested questions remain open in good faith. ⚖ Contested EV TCO advantage depends on residence, mileage and home charging access — averages mislead individuals. Auto-insurance pricing reflects both genuine risk pooling and ZIP-code proxy discrimination — the boundary is empirically resolvable but politically encrusted. Congestion pricing produces clear aggregate benefits — ⚖ Contested but the distributional impact on specific outer-borough commuters is real and remains to be addressed via transit investment proceeds. The emergence of effective car-sharing and ride-hailing changes the cost calculus for marginal urban users without addressing the core suburban-car structure.
The synthesis: the cost of owning a car is real but never primarily a vehicle decision. It is a function of urban form, fiscal policy, infrastructure investment ratios, insurance regulation and parking economics — all of which are public choices the household made no contribution to. The household is presented with a bill at the end of the month for $850–$1,400. The bill is the visible artefact of a series of public decisions that set the level of the bill ten, twenty or fifty years before the household existed.
The policy implication is straightforward: cities that wish to lower the household cost of mobility cannot do so by lowering the price of cars. The cost stack is structurally fixed by depreciation, insurance, fuel, parking, finance and externalities. The price of cars can fall by 20% without changing the AAA $11,577 figure meaningfully — depreciation eats most of the saving. The cities that have meaningfully reduced household mobility cost have done so by replacing the car: dense transit, paid parking, zone enforcement, off-street-parking quotas, fuel-tax indexation and direct caps. Tokyo, Paris, Singapore, central London and increasingly New York have followed this playbook to substantively different totals — three to ten times lower than the headline US figure for the equivalent household.
The cost of owning a car is the cost of a particular kind of city. The household chooses the car. The city chooses what the car costs. For the past eighty years, most American and Australian metros have chosen high; Tokyo, Paris, Singapore and central London have chosen low. The next forty years will determine whether the post-war suburban experiment is the historical norm or the historical aberration. The household budget of the median family will be decided by that answer.
The car will not disappear from the developed-world household. The middle-class American suburb, the rural French village, the Australian regional town and the suburban Tokyo edge city will continue to require a vehicle for the foreseeable future. But the structural policy question is no longer whether the car is necessary; it is whether the city is designed in a way that makes the car necessary, or in a way that makes it optional. The cost difference between those two designs, measured in the household budget over a working life, runs to the order of half a million dollars per adult. It is not a marginal number. It is the largest line item the typical household has ever paid for a public-policy choice nobody on the household ballot ever explicitly cast.