Is Car Insurance More Expensive for Men?

Why Insurers Started Pricing by Gender in the First Place

The most authoritative source of American crash data is the Fatality Analysis Reporting System (FARS)— a nationwide census built by the National Highway Traffic Safety Administration (NHTSA) that has tracked every fatal traffic crash in the country since 1975^[2]. State-employed FARS analysts code more than 140 data elements per crash from police reports, driver records, and death certificates, including the sex of every driver, passenger, and pedestrian involved^[2]. That decades-long dataset is part of the aggregate risk picture insurers and actuaries point to, and it draws a consistent line: for nearly every year between 1975 and 2024, male crash deaths ran more than double female crash deaths^[3].

The gap is not just about crash frequency — it is about behavior. Men drive more total miles than women, which alone increases exposure to a crash. On top of that, men wear seat belts less consistently, and male-involved crashes disproportionately involve excessive speed and a vehicle leaving the roadway to strike a fixed object such as a tree or utility pole^[3]. Put together, male drivers have a fatal crash involvement rate roughly 48% higher per mile traveled than female drivers^[3]. Alcohol widens the gap further: at every age interval, a higher share of fatally injured male drivers test at or above the 0.08% legal blood alcohol limit than fatally injured female drivers^[4].

Historically, the physical outcome of a crash of similar severity was also worse for women — differences in the vehicles men and women tended to drive, and in the circumstances of their specific crashes, left women more likely to be killed or injured in a crash of comparable force^[3]. That gap is closing quickly as vehicle safety engineering improves. A NHTSA analysis of the female fatality-risk disparity found it fell from an 18.3% gap in vehicles built between 1960 and 2009, to 6.3% in vehicles built 2010–2020, to just 2.9% for the newest 2015–2020 model years — a decline NHTSA attributes to stronger Federal Motor Vehicle Safety Standards for seat belts and dual air bags^[5]. As that physical-risk gap narrows, one of the original actuarial justifications for gender-based pricing narrows with it.

Translating a national risk statistic into an individual policyholder’s premium is the job of actuarial ratemaking. In a simplified multiplicative rating plan, an insurer sets a base rate and multiplies it by weighted factors for each policyholder trait — territory, age, gender, vehicle usage, and dozens of others^[1]. Charging men more because men, as a statistical group, generate more expensive claims is a textbook example of statistical discrimination— an insurer using a group characteristic as a proxy for individual risk, without any implication of malicious intent^[6]. That practice is mathematically defensible at the aggregate level. Whether it is defensible for any one 22-year-old man who has never had an accident is a separate question, and it is the question that pushed several state legislatures to intervene.

The Paradox: Real Quotes Don’t Match the National Story

If insurers priced purely off the FARS risk data, men would pay more than women in virtually every quote, every city, and every age bracket. That is not what happens. The Consumer Federation of America (CFA) — a nonprofit association of consumer advocacy organizations — tested this directly by pulling online quotes for simulated drivers aged 20, 40, and 60 across ten major U.S. cities, holding every variable identical except sex, and giving every simulated driver a perfectly clean record^[7].

The results ran against the traditional safety narrative. Women aged 40 and 60 with perfect driving records were charged more than identical men nearly twice as often as men were charged the higher rate^[7]. Sixty-year-old women were penalized in 58% of the instances where an insurer used sex to set the price at all. In 38 specific test cases, a woman with a clean record paid at least $100 more per year than an identical man; in six of those cases, the gap exceeded $500 a year^[7]. Even among 20-year-olds, where premiums generally ran lower for women, there were exceptions — GEICO charged young female drivers more than young male drivers in nine of the ten cities tested, by an average of $143^[7].

What makes this look less like careful risk math and more like inconsistent guesswork is how differently competing insurers priced the exact same driver. In more than two-thirds of the CFA’s test cases, at least one major insurer treated a female driver as higher risk than an identical male driver, while a competitor in that same city treated her as lower risk^[7]. Among 40-year-old drivers in Tampa, Florida, Allstate quoted a woman 21% less than an identical man, while Progressive quoted the same woman 32% more — a 53-percentage-point swing between two national carriers pricing the identical risk profile^[7]. A rating factor that produces that much disagreement between companies is not behaving like an objective measurement of physical risk.

State-level regulatory testing tells the same story. A CFA analysis of Delaware quotes found insurers in the state charged women 8% more than men on average, with seven of the ten largest carriers charging women more — GEICO by 21% and Progressive by 20%^[8]. One GEICO quote for a 20-year-old driver in Dover, Delaware, came back at $3,198 a year for a man and $3,951 for a woman with an identical profile, a $753 penalty tied to nothing but the gender field on the application^[8]. Oregon showed a smaller but still measurable version of the same pattern: a 35-year-old woman with a safe record paid $976.05 annually for basic coverage against $876.20 for an identical man^[9].

Public perception has not caught up with any of this. In a national survey conducted for the CFA, 48% of Americans believed insurers charge men more, only 23% believed women are charged more, and 17% believed pricing is equal^[7]— even though the CFA’s own market testing found women penalized more often than men across the specific age groups it studied.

Where Gender Rating Is Now Illegal

Faced with a rating factor that lacks industry-wide consensus and produces the kind of 50-percentage-point swings documented above, six states currently ban insurers from using gender to set private passenger auto insurance rates at all (Montana banned it from 1985 to 2021 before repealing the ban). The stated goal behind each ban is the same: shift pricing toward factors a driver actually controls — driving record, annual mileage, years of experience — and away from an immutable personal characteristic^[11].

California’s ban is the most heavily documented. Effective January 1, 2019, the state’s Gender Non-Discrimination in Automobile Insurance Rating Regulation amended 10 CCR § 2632.5 to prohibit gender as a rating factor either alone or blended with any other variable, and required every insurer to file a revised, gender-free Class Plan by July 1, 2019 or risk having its entire rating structure rejected^[12]. With gender off the table, California leans harder on the three primary factors established by Proposition 103: driving safety record, annual mileage, and years of driving experience — and it regulates how insurers estimate that mileage in granular detail, down to requiring insurers to ask for a workplace or school location and a current odometer reading^[12].

Massachusetts takes a related but distinct approach under 211 CMR 79.00, banning gender, credit score, income, and occupation alike, and replacing them with the Safe Driver Insurance Plan— a merit-rating system that applies a 15% surcharge per point (7.5% per point for inexperienced operators) for a traffic violation or an at-fault accident claim exceeding $1,000, and rewards five or six years of clean driving with an Excellent Driver Discount^[13]. Pennsylvania’s ban rests on a different foundation entirely — constitutional case law. In Hartford Accident & Indemnity Co. v. Insurance Commissioner(1984), the Pennsylvania Supreme Court upheld the state insurance commissioner’s authority to bar gender-based rates, establishing that if a pricing practice has a discriminatory effect and an adjustment exists that satisfies both the insurer’s need for accurate pricing and the public’s interest in fairness, the insurer must make that change^[15]. North Carolina has barred gender-based rating by statute since September 1, 1975, under N.C. Gen. Stat. § 58-3-25^[14].

Montana’s Ban-and-Repeal: A Cautionary Tale, Not a Clean Answer

Montana offers the clearest real-world experiment on what happens when a ban actually takes effect. In 1985, Montana became the first and only state to prohibit gender-based pricing across every line of insurance — auto, life, and health — and it held that policy for over three decades^[16].

The result was not the evenly blended, unisex rate consumer advocates might have expected. Instead, insurers successfully argued for the ability to charge every driver the historically higher rate previously reserved for high-risk male drivers. Female drivers in Montana ended up paying close to 30% more for auto insurance than they would have in neighboring states like Idaho and the Dakotas, where their lower-risk gender was still factored into the price^[16]. Arguing the ban had artificially inflated women’s premiums, the Montana legislature repealed it in 2021, reinstating gender as a rating factor^[16].

A year after the repeal, the CFA went back and tested Montana’s market again. Among four major insurers — GEICO, Farmers, Liberty Mutual, and Progressive, together holding about 40% of the state’s market — two still charged women more, one charged men more, and one priced both genders equally^[16]. Some women paid up to $230 more per year than men for the same state-required minimum coverage. Neither the 36-year ban nor its repeal produced a consistent, defensible outcome for women in Montana, which is the strongest evidence in this entire research file that gender rating is not a precise measurement of risk so much as a rough proxy that different insurers apply in different directions.

What Replaces Gender Once It’s Banned

Removing gender from a rating plan does not remove the underlying pressure to distinguish risk between drivers — it just forces insurers toward other variables. The most direct replacement is telematics, where a plug-in device or smartphone app measures a driver’s actual hard-braking events, sharp turns, and mileage, replacing a demographic guess with a direct behavioral measurement^[17]. Consumer acceptance is still mixed: roughly half of adults consider hard-braking and cornering monitoring a fair way to set prices, even though it is a far more precise measure of risk than sex, age, or zip code^[17].

The harder regulatory problem is proxy discrimination— when a facially neutral rating variable, like occupation, education, or a specific vehicle model, produces the same disparate outcome that an explicit gender factor would have produced, without ever asking a driver to disclose their sex^[18]. Modern algorithms trained on large datasets are especially prone to this, because they are mathematically optimized to find whatever pattern predicts claims cost most accurately — and if gender-correlated proxies predict cost well, an unsupervised model will find and use them even when gender itself is banned^[18]. Regulators in gender-ban states now have to audit rating plans for these backdoor correlations, not just check whether a “gender” field appears anywhere in the formula.

Put together, the research points toward a gradual shift away from broad demographic categories and toward direct, individualized measurement — a shift already visible in how insurers price other non-driving factors, such as the actuarially documented effect of enrollment status covered in our companion research on how being a student affects car insurance. As vehicle crashworthiness keeps narrowing the physical injury gap between men and women, and as telematics data becomes cheaper to collect than it was even five years ago, the technical case for pricing an individual driver based on their sex keeps getting weaker, even in the states that still allow it.