How many insured cars carry collision and comprehensive, by state: NAIC car-year shares, 2020 to 2024
Across the NAIC’s 2024 data, collision was written on 76.3 car-years for every 100 car-years of liability insurance, and comprehensive on 79.2. By jurisdiction the collision figure runs from 66.1 in Arkansas to 94.4 in New Hampshire, and the comprehensive figure from 65.8 in Texas, where the NAIC estimates comprehensive exposures, to 95.9 in New Hampshire. The unit is the insured vehicle-year, not the driver or the policy.
That leaves 56,457,127 liability car-years with no collision and 49,433,439 with no comprehensive. The spread between states is stable: the ranking of the 51 jurisdictions in 2020 and in 2024 has a rank correlation of 0.96 for both coverages. A popular explanation, that lenders force the coverage on financed cars, did not show up in a state-level test against auto-loan balances per consumer, which pointed the other way; the loan section explains why that test is weaker than it sounds. The national dollar side of the same tables is in the national average premium guide.
Estimates and published averages, not a quote. Not insurance advice. State rules and insurer rules vary. Confirm requirements with your state insurance regulator or a licensed agent in your state.
What the share measures and what it misses
The NAIC publishes written exposures, in car-years, for each coverage. Its definition is short: “One car-year is the risk associated with insuring one car for one year.” Table 1B counts car-years written with liability, Table 2B with collision and Table 3B with comprehensive. Dividing Table 2B by Table 1B gives collision car-years per liability car-year, and Table 3B by Table 1B gives the comprehensive figure. The NAIC does not print those ratios. Every share in this guide is the writer’s division of two published tables, rounded to one decimal place.
Because the unit is a car-year, the share describes vehicles, not people. The supplement’s own example is a two-car policy running six months, which it counts as one car-year in the calendar year the policy was issued. A household with three cars, one of them an old pickup kept on liability only, therefore contributes two collision cars out of three. Questions phrased as “what percentage of drivers” are answered here per insured vehicle-year, because that is the only unit the tables carry. A car insured for only part of a year contributes only that fraction of a car-year.
The NAIC states an assumption that shapes how the ratio should be read: “This assumes that all insured vehicles carry liability coverage but do not necessarily carry the physical damage coverages (i.e., collision and/or comprehensive).” That sentence belongs to the NAIC’s average-expenditure method, and it is an assumption rather than a count. If some car-years carry collision or comprehensive without liability (comprehensive-only coverage on a stored vehicle is one possibility), the ratio overstates the share of liability cars that also have physical damage coverage. The supplement does not report how many such car-years exist, so the size of any overstatement is unknown.
The tables also carry no joint count of cars that have both collision and comprehensive. The two shares cannot be combined into a “full coverage” percentage. What the arithmetic allows is a range: if every collision and comprehensive car-year sits on a liability car-year, the share with both is at least 76.3 plus 79.2 minus 100, which is 55.5 countrywide, and at most the smaller of the two, 76.3. Any single published number for “full coverage” is a point inside a range like that, or it comes from a different source and a different unit, as the comparison section below shows.
Coverage of the data matters too. The NAIC lists the statistical agents that supplied it (AAIS, ISO, NISS, ISS, Massachusetts CAR and MAIF, plus the California and Texas departments of insurance) and adds: “Some data from other small statistical agencies may not be included.” The business is the combined voluntary and residual market. Motor homes, recreational vehicles, campers, travel trailers, all-terrain vehicles, antique autos, snowmobiles, golf carts, motorcycles, scooters and mopeds are listed as not included. The supplement is a September 2026 edition covering 2024, and earlier years may be restated: “Data contained in this supplement may differ from data contained in reports from previous years, as the statistical agents periodically obtain updated information from insurers.”
Countrywide car-years, 2020 to 2024
The countrywide rows give the frame for the state table. Liability car-years grew from 226,878,478 in 2020 to 238,174,969 in 2024, a rise of 5.0%. Collision car-years grew faster, 6.6%, and comprehensive car-years grew 5.3%. That is why the collision share edged up from 75.1 to 76.3, a gain of 1.2 points, while the comprehensive share moved from 79.0 to 79.2.
| Year | Liability car-years | Collision car-years | Comprehensive car-years | Collision share | Comprehensive share |
|---|---|---|---|---|---|
| 2020 | 226,878,478 | 170,398,478 | 179,188,851 | 75.1% | 79.0% |
| 2021 | 232,118,214 | 175,908,230 | 184,514,514 | 75.8% | 79.5% |
| 2022 | 233,051,959 | 178,303,888 | 186,833,884 | 76.5% | 80.2% |
| 2023 | 235,568,929 | 180,269,958 | 188,077,873 | 76.5% | 79.8% |
| 2024 | 238,174,969 | 181,717,842 | 188,741,530 | 76.3% | 79.2% |
Source: NAIC 2024 Auto Insurance Database Average Premium Supplement (September 2026), Tables 1B, 2B and 3B, Countrywide rows, checked 2026-11-04. Shares are collision or comprehensive car-years divided by liability car-years, computed by the writer. The 2023 and 2024 California data in the supplement are preliminary, and Texas comprehensive car-years are estimated (see the state section).
The two series peaked in different years. Comprehensive reached 80.2 in 2022 and slid to 79.2 by 2024; collision was 76.5 in both 2022 and 2023 (76.51 and 76.53 before rounding) and dipped slightly in 2024. Neither move is large enough to read as a trend in what drivers choose: a change of a few tenths of a point on 238 million car-years can come from the mix of insurers reporting, from policies restated in later editions, or from a shift in which cars are insured at all. Because earlier years can be restated in each edition, the 2023 row here may differ from the one printed in the previous edition.
All 51 jurisdictions, ranked by collision share
Rank 1 is the highest collision share in 2024, using unrounded values, so two rows that print the same figure can sit one place apart. The comprehensive-minus-collision column is the difference in percentage points. The last two columns show how far each share moved between 2020 and 2024, in points. The list covers the 50 states and the District of Columbia, every row of the NAIC tables except the countrywide total, which closes the table.
| Rank | Jurisdiction | Liability car-years, 2024 | Collision share | Comprehensive share | Comprehensive minus collision (points) | Collision change since 2020 (points) | Comprehensive change since 2020 (points) |
|---|---|---|---|---|---|---|---|
| 1 | New Hampshire | 1,041,591 | 94.4% | 95.9% | +1.5 | +1.9 | +1.2 |
| 2 | Maine | 1,078,527 | 86.8% | 90.6% | +3.8 | +9.0 | +7.1 |
| 3 | Massachusetts | 4,680,957 | 86.4% | 87.5% | +1.0 | +0.6 | -2.7 |
| 4 | District of Columbia | 253,552 | 84.9% | 86.9% | +1.9 | +3.3 | +2.0 |
| 5 | Washington | 5,128,833 | 82.9% | 88.4% | +5.5 | +3.5 | +1.6 |
| 6 | Vermont | 483,476 | 82.8% | 87.8% | +5.0 | -2.4 | +0.4 |
| 7 | Wisconsin | 4,264,262 | 82.6% | 87.0% | +4.4 | +1.6 | +0.4 |
| 8 | Ohio | 8,234,110 | 82.5% | 87.2% | +4.7 | +2.4 | +2.5 |
| 9 | Illinois | 8,175,097 | 82.3% | 86.5% | +4.2 | +0.3 | -0.7 |
| 10 | Pennsylvania | 9,151,357 | 82.1% | 85.8% | +3.7 | +2.5 | +1.6 |
| 11 | New Jersey | 6,077,195 | 81.7% | 84.8% | +3.0 | +2.9 | +2.1 |
| 12 | Maryland | 4,432,039 | 81.7% | 85.2% | +3.6 | +1.4 | +1.4 |
| 13 | Connecticut | 2,449,176 | 81.2% | 87.0% | +5.8 | +1.3 | -0.9 |
| 14 | Delaware | 776,728 | 80.9% | 84.4% | +3.5 | +1.8 | +1.9 |
| 15 | Michigan | 6,039,338 | 80.0% | 90.3% | +10.3 | +1.5 | -0.4 |
| 16 | Minnesota | 4,325,677 | 79.6% | 84.6% | +5.1 | +1.4 | +0.1 |
| 17 | Oregon | 3,110,123 | 79.4% | 86.0% | +6.6 | +2.8 | +1.0 |
| 18 | Florida | 16,088,497 | 79.3% | 81.3% | +1.9 | -2.9 | -4.3 |
| 19 | New York | 9,717,769 | 79.3% | 85.9% | +6.6 | +1.0 | +0.6 |
| 20 | Indiana | 5,176,284 | 78.8% | 82.5% | +3.7 | +2.2 | +1.8 |
| 21 | Hawaii | 942,914 | 78.6% | 83.9% | +5.3 | +3.0 | +3.0 |
| 22 | Utah | 2,383,211 | 78.4% | 81.8% | +3.5 | +1.3 | +0.4 |
| 23 | Virginia | 6,886,425 | 77.7% | 81.1% | +3.4 | +0.5 | -1.0 |
| 24 | Rhode Island | 731,419 | 77.1% | 80.0% | +2.9 | +2.8 | +2.2 |
| 25 | Colorado | 4,545,866 | 75.9% | 77.4% | +1.5 | +2.4 | +0.8 |
| 26 | Iowa | 2,572,374 | 75.8% | 78.8% | +2.9 | +1.3 | +0.9 |
| 27 | Arizona | 5,401,218 | 75.8% | 79.1% | +3.3 | +1.4 | +1.0 |
| 28 | Missouri | 4,411,553 | 75.7% | 77.1% | +1.4 | +1.6 | +0.9 |
| 29 | Idaho | 1,476,491 | 75.7% | 79.6% | +3.9 | +3.5 | +3.0 |
| 30 | Nevada | 2,291,938 | 75.3% | 77.3% | +2.1 | +1.7 | +0.4 |
| 31 | South Carolina | 4,449,593 | 75.1% | 77.9% | +2.8 | +1.6 | -0.4 |
| 32 | North Dakota | 726,658 | 74.6% | 77.2% | +2.7 | +4.2 | +3.8 |
| 33 | West Virginia | 1,321,682 | 74.3% | 79.1% | +4.7 | +2.9 | +2.0 |
| 34 | California | 28,413,710 | 74.0% | 76.7% | +2.7 | +2.8 | +2.1 |
| 35 | Tennessee | 5,589,820 | 73.8% | 77.4% | +3.6 | +0.8 | +0.1 |
| 36 | Louisiana | 2,975,786 | 73.4% | 75.2% | +1.7 | +1.1 | +0.8 |
| 37 | Alaska | 506,995 | 73.4% | 84.9% | +11.5 | +3.3 | +2.9 |
| 38 | Mississippi | 2,023,372 | 73.1% | 74.2% | +1.1 | +1.7 | +1.0 |
| 39 | North Carolina | 8,857,647 | 72.2% | 75.6% | +3.4 | +0.5 | -0.3 |
| 40 | Georgia | 8,229,337 | 71.8% | 74.4% | +2.5 | +1.1 | +0.1 |
| 41 | Alabama | 4,160,006 | 71.6% | 73.4% | +1.7 | +0.9 | 0.0 |
| 42 | Kentucky | 3,484,114 | 70.1% | 74.3% | +4.2 | +2.8 | +1.6 |
| 43 | Nebraska | 1,714,000 | 69.7% | 71.0% | +1.3 | +0.6 | 0.0 |
| 44 | South Dakota | 751,476 | 69.6% | 71.9% | +2.3 | +1.9 | +2.0 |
| 45 | Kansas | 2,490,413 | 69.3% | 70.3% | +1.0 | +1.2 | +0.9 |
| 46 | Wyoming | 491,312 | 68.7% | 71.4% | +2.6 | +2.3 | +3.6 |
| 47 | Montana | 731,794 | 68.5% | 70.6% | +2.1 | +3.0 | +1.9 |
| 48 | Texas | 21,786,401 | 67.8% | 65.8% | -2.0 | -1.3 | -1.9 |
| 49 | Oklahoma | 3,008,373 | 67.4% | 67.9% | +0.5 | +1.3 | +0.4 |
| 50 | New Mexico | 1,630,946 | 67.0% | 69.3% | +2.2 | +2.9 | +2.8 |
| 51 | Arkansas | 2,503,537 | 66.1% | 67.5% | +1.4 | -0.1 | -0.7 |
| United States (countrywide) | 238,174,969 | 76.3% | 79.2% | +2.9 | +1.2 | +0.3 |
Source: NAIC 2024 Auto Insurance Database Average Premium Supplement (September 2026), Tables 1B, 2B and 3B, 2020 and 2024 columns, checked 2026-11-04; shares and differences computed by the writer from unrounded values. Texas: liability and collision exposures are averages of policies-in-force vehicles and comprehensive exposures are estimated from a ratio, so the Texas comprehensive share and its gap are estimates. California: 2023 and 2024 data are preliminary. Direct comparisons between states call for caution because of the factors the NAIC lists.
The NAIC attaches a warning to the state comparisons in this supplement: “CAUTION: Because of these differences, direct comparisons between state results should be treated with a high degree of caution.” Its list of the differences includes the types of coverages purchased, the relative amounts of coverage purchased, use of telematics, weather, driving locations, vehicle theft rates and auto repair costs. The caution was written for the premium tables, and it applies to the share just as well, since, in the writer’s reading, the share depends on which cars a state’s insurers write and which cars its statistical agents report.
The median of the 51 jurisdictions is 75.8 for collision and 79.6 for comprehensive. The countrywide figure is 76.3 because it weights by cars. The five largest jurisdictions by liability car-years, California, Texas, Florida, New York and Pennsylvania, hold 35.8% of all liability car-years, and two of them, California at 74.0 and Texas at 67.8, sit below the countrywide collision share. Of the 51 jurisdictions, 24 have a collision share above 76.3 and 26 have a comprehensive share above 79.2.
The distribution has three bands. Fifteen jurisdictions have a collision share of 80 or more, nine are below 70, and twenty-seven fall between 70 and 80. The next two sections take the two ends.
The fifteen jurisdictions at 80 and above
In rank order the top fifteen are New Hampshire (94.4), Maine (86.8), Massachusetts (86.4), the District of Columbia (84.9), Washington (82.9), Vermont (82.8), Wisconsin (82.6), Ohio (82.5), Illinois (82.3), Pennsylvania (82.1), New Jersey (81.7), Maryland (81.7), Connecticut (81.2), Delaware (80.9) and Michigan (80.0). Ten of them sit in the Northeast and mid-Atlantic corridor, four around the Great Lakes, and one, Washington, in the West. The fifteen contain no state from the Plains, the Rocky Mountains or the Deep South.
New Hampshire stands apart. Its collision share is 94.4 and its comprehensive share 95.9, and the next-highest collision share, Maine’s, is 7.6 points lower. The supplement gives no reason and the tables cannot supply one. One possibility, which this data cannot test and which is the writer’s reading rather than the NAIC’s, is that the denominator matters: how New Hampshire treats compulsory coverage is covered in the uninsured drivers by state guide, and a smaller or more selective liability-insured pool would push the ratio up. Whatever the cause, 94.4 does not mean that nine in ten New Hampshire drivers choose physical damage coverage. It means that for every 100 car-years with liability there were 94.4 with collision.
Maine is the other unusual row, for a different reason. Its collision share was 77.8 in 2020 and is 86.8 in 2024, a gain of 9.0 points, the largest of any jurisdiction by more than double, and its comprehensive share rose 7.1 points over the same years. Maine moved from 18th in the 2020 ranking to 2nd. Ohio, Pennsylvania, Illinois and Wisconsin, each with between 4.3 million and 9.2 million liability car-years, hold their places with much larger pools than New Hampshire (1,041,591 liability car-years) or Maine (1,078,527), so a change in one insurer’s reporting would move their shares less.
The nine jurisdictions under 70
At the other end, in rank order from the bottom, the nine jurisdictions below 70 are Arkansas (66.1), New Mexico (67.0), Oklahoma (67.4), Texas (67.8), Montana (68.5), Wyoming (68.7), Kansas (69.3), South Dakota (69.6) and Nebraska (69.7). Only four jurisdictions have a comprehensive share below 70: Texas (65.8), Arkansas (67.5), Oklahoma (67.9) and New Mexico (69.3).
Two things keep this band from being simply the poorest states. Mississippi, Louisiana and West Virginia have three of the five lowest median household incomes in the Census series used later in this guide, and none of them is in the bottom nine: their collision shares are 73.1, 73.4 and 74.3. Meanwhile Montana, Wyoming, Kansas, South Dakota and Nebraska are not in the lowest income fourth at all. A single explanation does not cover the band. The supplement carries no split by vehicle age, rural or urban location, or loan status, so any account of why these nine sit low is a reading of the pattern, not a finding.
Texas needs its own note, because its collision and comprehensive series are built differently from the other states’. The supplement explains: “The Texas Department of Insurance collects vehicle and policies-in-force information at the end of each calendar quarter for voluntary bodily injury liability, involuntary (residual) bodily injury liability and collision coverages. The average number of vehicles reported for policies in force is used as an approximation for written exposures. Comprehensive exposures for all years are estimated based on the ratio of comprehensive to collision car-years. The data to estimate this ratio only includes about 70 percent of the private passenger automobile market. This estimation affects combined average premium but not average expenditure.” In other words, the Texas liability and collision exposures behind the collision share are both policies-in-force approximations, and the Texas comprehensive share of 65.8 is an estimate scaled from collision. It is the only jurisdiction among the 51 where the comprehensive share is below the collision share. That would follow arithmetically if the ratio the NAIC used is below one, which is the writer’s reading of the method because the supplement does not print the ratio; a ratio below one would itself be unusual, since the other 50 jurisdictions all report more comprehensive than collision car-years. The Texas comprehensive figure should be treated as indicative only.
California carries a different flag: “The 2023 and 2024 California auto insurance data in these tables is preliminary.” The NAIC adds that the California Department of Insurance tests the data each year after publication and that any adjustments will appear in the next edition. California is 11.9% of all liability car-years, so a revision there would move the countrywide figure as well as its own row; its 2024 collision share of 74.0 and comprehensive share of 76.7 should be read with that in mind.
Comprehensive bought without collision
Comprehensive and collision cover different events. The Triple-I statistics page puts the split briefly: “Collision and comprehensive insurance cover property damage and theft to the policyholder’s car.” In practice a driver can drop collision, which pays for damage from a crash the driver causes, and keep comprehensive, which is the coverage the site’s guides tie to theft, hail and hitting a deer. The comprehensive against collision guide sets the two coverages side by side. The gap column in the state table measures how often that choice is made in aggregate.
Among the 51 jurisdictions, comprehensive car-years exceed collision car-years in 50, and the median difference is 3.0 points. Alaska has the widest gap, 11.5 points (73.4 collision, 84.9 comprehensive), followed by Michigan at 10.3 (80.0 against 90.3), then Oregon and New York at 6.6 each, Connecticut at 5.8 and Washington at 5.5. Eight jurisdictions have a gap of 5 points or more, and fourteen have a gap under 2 points; those fourteen include Texas, whose gap is negative at -2.0 because of the estimation method described above. Massachusetts and Kansas are at 1.0 and Oklahoma at 0.5.
The supplement does not say why Alaska and Michigan stand out. One reading, labelled as the writer’s and not tested, is that comprehensive-only coverage on vehicles that are stored or seldom driven would add comprehensive car-years without adding liability car-years, which would raise the comprehensive figure relative to the denominator. The tables cannot confirm it. The two states also differ in many other ways, including how collision is priced and sold in each, so a single cause is unlikely to explain both.
What the gap column does show is where the “full coverage” shorthand is least precise. In a jurisdiction with a gap near zero, the two shares are close, which is consistent with most vehicles holding both coverages or neither. Where the gap is large, the pattern is consistent with a block of vehicles that carry only comprehensive, though the tables do not count them. The sentence “X% of drivers have full coverage” hides that block, and it is one reason the single percentages published elsewhere differ.
Movement between 2020 and 2024
The ranking is persistent. The rank correlation between the 2020 and 2024 collision shares across the 51 jurisdictions is 0.96, and the same figure holds for comprehensive. Within that stable order, the collision share rose in 47 of the 51 jurisdictions between 2020 and 2024 and fell in four: Florida (down 2.9 points), Vermont (2.4), Texas (1.3) and Arkansas (0.1). Eight rose by 3.0 points or more, and 43 moved by less than 3 points in either direction.
The largest gains were Maine (+9.0, from 77.8 to 86.8), North Dakota (+4.2, from 70.4 to 74.6), Washington (+3.5, from 79.4 to 82.9), Idaho (+3.5, from 72.2 to 75.7), the District of Columbia (+3.3, from 81.6 to 84.9) and Alaska (+3.3, from 70.2 to 73.4). Comprehensive fell in eleven jurisdictions: Florida (down 4.3), Massachusetts (2.7), Texas (1.9), Virginia (1.0), Connecticut (0.9), Illinois (0.7), Arkansas (0.7), Michigan (0.4), South Carolina (0.4), North Carolina (0.3) and Alabama (0.03, which the table prints as 0.0). Comprehensive rose in the other 40.
Florida is the clearest reversal. In 2020 its collision share put it fourth highest of the 51; in 2024 it is 18th, with 79.3 for collision and 81.3 for comprehensive, down 2.9 and 4.3 points. The supplement shows the pattern and gives no cause, and the data offer no way to separate a drop in the number of cars carrying the coverage from an increase in cars carrying liability alone. In the same supplement, Florida’s combined average premium rose from $1,469.92 in 2020 to $2,174.55 in 2024, a rise of 47.9%. That fits one reading, that some Florida owners dropped optional coverage as prices climbed, but the shares alone do not establish it, and the price trend guide covers the national price series.
Taken together, the four decreasing states and the eleven comprehensive decreases are a small minority. In most jurisdictions the share crept up by one or two points over four years. For a buyer, the practical point is that the benchmark is slow-moving: the share that applied in a state in 2020 is a good guide to the share that applies now, to within three points in 43 of the 51 jurisdictions.
Testing the loan explanation
A common explanation for why most cars carry both coverages is the loan. A buyer who finances a vehicle signs a contract that requires physical damage insurance, and the financed car guide works through what those contracts say. Age of the car points the same way in one set of quote data: Insurance.com’s 2014 analysis of 557,238 drivers who requested quotes in July 2014 found that about 96 percent of drivers with cars six years old shopped for comprehensive and collision, against 60 percent of those with cars 10 years old, “as car loans are paid off and vehicle value drops to a level that makes owners reconsider the risk.” The explanation is the publisher’s, the data are quote requests rather than policies, and they are twelve years old, so they are a pointer and not a measurement.
A state-level version of the test needs a state-level measure of how many cars are financed. The closest available series is the Federal Reserve Bank of New York’s auto debt balance per capita, from its Consumer Credit Panel, a 5% sample of adults with a credit file. It measures dollars of auto debt per consumer, so it moves with both how many people borrow and how much they borrow. It is not the share of cars with a lien. The Q4 2024 values run from $3,580 in the District of Columbia to $7,920 in Texas.
The result runs the opposite way from what the loan explanation would predict if balance per consumer were a good stand-in for financed cars. Across the 51 jurisdictions the rank correlation between collision share and auto debt per consumer is -0.45: where the average balance is higher, collision is less common, not more. Comprehensive gives the same -0.45 (-0.42 with Texas removed). Median household income, taken from the Census Bureau’s Table H-8 for 2024, correlates positively with both shares, at 0.58 for collision and 0.54 for comprehensive. Debt per consumer and income correlate at -0.48, which on the writer’s reading is behind much of the negative figure: high-balance jurisdictions tend to be lower-income ones. Holding income fixed, the writer’s partial rank correlation between collision share and debt is -0.24, still negative but smaller, and the partial correlation with income holding debt fixed is 0.46. All of these are associations among 51 state aggregates, and they say nothing about any one household.
| Measure | Group | Size | Range of the measure | Collision share, pooled | Comprehensive share, pooled | Median of the jurisdictions’ collision shares |
|---|---|---|---|---|---|---|
| Median household income, 2024 | Lowest fourth | 13 jurisdictions | $55,620 to $76,220 | 74.3% | 76.9% | 73.1% |
| Median household income, 2024 | Second fourth | 13 jurisdictions | $77,560 to $83,090 | 75.8% | 78.2% | 75.7% |
| Median household income, 2024 | Third fourth | 13 jurisdictions | $84,350 to $91,890 | 77.4% | 82.0% | 77.1% |
| Median household income, 2024 | Highest fourth | 12 jurisdictions | $95,980 to $112,300 | 78.1% | 81.1% | 81.4% |
| Auto debt per consumer, Q4 2024 | Lowest fourth | 13 jurisdictions | $3,580 to $4,940 | 81.0% | 85.6% | 81.2% |
| Auto debt per consumer, Q4 2024 | Second fourth | 13 jurisdictions | $4,950 to $5,530 | 76.3% | 79.9% | 75.7% |
| Auto debt per consumer, Q4 2024 | Third fourth | 13 jurisdictions | $5,610 to $6,140 | 76.6% | 80.1% | 77.7% |
| Auto debt per consumer, Q4 2024 | Highest fourth | 12 jurisdictions | $6,220 to $7,920 | 72.1% | 72.8% | 72.5% |
Sources: NAIC 2024 supplement, Tables 1B, 2B, 3B (2024); U.S. Census Bureau, Table H-8, 2024 column (CPS ASEC), checked 2026-11-04; Federal Reserve Bank of New York, State Level Household Debt Statistics 2003-2025 (February 2026), auto sheet, Q4 2024 column, checked 2026-11-04. Groups are the lowest to highest fourths of the 51 jurisdictions (13, 13, 13 and 12 jurisdictions). Pooled shares add collision or comprehensive car-years over the group and divide by the group’s liability car-years. Texas comprehensive is estimated.
In the table the income pattern is gradual: pooled collision rises from 74.3 in the lowest income fourth to 78.1 in the highest, and the median jurisdiction rises from 73.1 to 81.4. The debt pattern is steeper and reversed: the lowest-debt fourth (the District of Columbia, Hawaii, Oregon, Massachusetts, New York, Connecticut, Wisconsin, Minnesota, Rhode Island, Pennsylvania, Kansas, Washington and Illinois) pools to 81.0 for collision and 85.6 for comprehensive, while the highest-debt fourth (Alabama, Nevada, Oklahoma, West Virginia, Arkansas, Mississippi, Georgia, Florida, North Dakota, New Mexico, Louisiana and Texas) pools to 72.1 and 72.8.
What this does and does not show. It does not show that lenders do not require physical damage coverage on financed cars; that requirement operates one vehicle at a time and the tables cannot see liens. It shows that a state-level proxy for borrowing, balance per consumer, is associated with less collision coverage rather than more, and that income tracks the shares more closely. A reasonable reading, labelled as the writer’s, is that vehicle value and household resources matter at the margin: in a jurisdiction where cars are cheaper and older, more owners can reasonably carry liability only, whether or not they ever had a loan. A direct test would need the share of insured vehicles with a lienholder, by state; neither the NAIC tables nor the Census tables used here carry it.
How the shares move the average price
Two NAIC price measures differ because of these shares. The combined average premium adds the liability, collision and comprehensive averages, each divided by the car-years that carried that coverage, so it prices a vehicle that has everything. The average expenditure divides all three coverages’ written premium by the liability car-years only, so it spreads the physical damage premium over every liability car, including those without it. Written as an identity, the combined premium minus the average expenditure equals one minus the collision share, times the average collision premium, plus one minus the comprehensive share, times the average comprehensive premium. Recomputed from the published averages and shares, the identity matches the printed Table 4 and Table 5 values to within a cent in every jurisdiction and year.
Ohio shows it. Its collision share is 82.47% and its average collision premium is $391.50, so the cars without collision account for $68.62 ((1 - 0.8247) x $391.50 gives $68.63 with the rounded share; the figures here use unrounded shares). Its comprehensive share is 87.19% and its average comprehensive premium is $224.96, which gives $28.83 ($28.82 with the rounded share). The two shortfalls sum to $97.45, and Ohio’s combined average premium of $1,133.98 less its average expenditure of $1,036.52 is $97.46, the one-cent difference being rounding. The Ohio combined and expenditure figures are the same ones used in the share-of-income guide.
The table runs the identity for the five highest and five lowest collision shares and for the countrywide row. The pattern is mechanical: the less common the coverage, the larger the gap between the two measures, and the further the combined premium sits above the NAIC’s estimate of what consumers in the state spend on average.
| Jurisdiction | Collision share | Comprehensive share | Combined average premium | Average expenditure | Gap | Part from cars without collision | Part from cars without comprehensive |
|---|---|---|---|---|---|---|---|
| New Hampshire | 94.4% | 95.9% | $1,103.33 | $1,073.18 | $30.15 | $22.75 | $7.40 |
| Maine | 86.8% | 90.6% | $1,000.13 | $934.54 | $65.59 | $47.20 | $18.39 |
| Massachusetts | 86.4% | 87.5% | $1,610.84 | $1,503.52 | $107.32 | $80.72 | $26.60 |
| District of Columbia | 84.9% | 86.9% | $1,986.50 | $1,840.36 | $146.14 | $106.40 | $39.75 |
| Washington | 82.9% | 88.4% | $1,521.63 | $1,416.32 | $105.31 | $80.06 | $25.25 |
| United States (countrywide) | 76.3% | 79.2% | $1,612.81 | $1,431.63 | $181.18 | $122.41 | $58.77 |
| Montana | 68.5% | 70.6% | $1,321.59 | $1,080.94 | $240.65 | $121.39 | $119.26 |
| Texas | 67.8% | 65.8% | $1,917.05 | $1,578.45 | $338.60 | $183.04 | $155.56 |
| Oklahoma | 67.4% | 67.9% | $1,453.74 | $1,182.76 | $270.98 | $149.72 | $121.26 |
| New Mexico | 67.0% | 69.3% | $1,421.95 | $1,180.17 | $241.78 | $151.31 | $90.47 |
| Arkansas | 66.1% | 67.5% | $1,465.02 | $1,172.35 | $292.67 | $172.29 | $120.38 |
Source: NAIC 2024 supplement, Tables 1C, 2C, 3C (average premiums), 4 (average expenditure) and 5 (combined average premium), with shares from Tables 1B, 2B and 3B, checked 2026-11-04. The two parts are computed by the writer from unrounded shares and sum to the gap within one cent. Texas comprehensive exposures are estimated.
New Hampshire’s gap is $30.15 on a combined premium of $1,103.33, because 94.4% of its liability car-years carry collision. Arkansas’s gap is $292.67 on $1,465.02 and Texas’s is $338.60 on $1,917.05, though the NAIC says its Texas comprehensive estimate affects the combined average premium, so the Texas figures are indicative. A buyer comparing a full-coverage quote with a state benchmark should take the combined average premium as the benchmark for that coverage and the average expenditure as the benchmark for the mixed pool of cars; the national average guide walks through the countrywide version of the same split. Neither is a quote, and neither says what a particular driver’s policy should cost.
Premium against expected claim cost
A share tells how often the coverage is carried. It does not tell whether carrying it pays. The Triple-I statistics page gives 2024 claim figures: “In 2024, 4.16 percent of collision insurance policyholders had a claim, while 3.95 percent of people with comprehensive coverage had a claim,” and “In 2024, the average collision claim was $5,489; the average comprehensive claim was $2,306.” Multiplying frequency by size gives a rough expected claim cost per covered policyholder of about $228 for collision (0.0416 x $5,489) and $91 for comprehensive (0.0395 x $2,306). Against the NAIC countrywide average premiums of $516.39 for collision and $283.15 for comprehensive, those are 44% and 32% of the premium.
That comparison is deliberately rough. The claim figures are per policyholder from one source and the premiums are per car-year from another; the page’s notes say the physical damage figures are paid claims per 100 earned car years from ISO, based on coverage with a $500 deductible, that they exclude Massachusetts, Michigan and Puerto Rico, and that the comprehensive figures leave out wind and water losses, which comprehensive coverage does pay for; and premium pays for the insurer’s expenses and for large losses that an average hides. It cannot be used as a loss ratio. It does show why individual decisions differ from the aggregate: on these figures the collision premium is about 2.3 times the expected claim cost and the comprehensive premium about 3.1 times, a multiple that runs high because the comprehensive claim figures leave out wind and water losses. An average hides the distribution: on the collision figure above, about 96 in 100 policyholders had no claim in 2024 and a few had a large one, and that spread, not the average, is what the coverage is bought for.
Publishers use rules of thumb to turn that into advice. MoneyGeek’s page on dropping coverage, last updated October 11, 2026, lists a “10% rule: Your annual premium for collision and comprehensive exceeds 10% of what your car is worth.” It is one of four conditions that the page says should all apply before dropping coverage makes financial sense. At the countrywide average physical damage premium of $799.54 ($516.39 plus $283.15), that rule would flag any car worth less than about $7,995. The average, though, is taken over all vehicles, so a given car’s premium can sit well above or below it; the arithmetic shows the scale, not a verdict. Whether to carry the coverage on a given car depends on its value, the deductible, savings and any lender requirement, which the liability-only against full coverage guide and the deductible guide work through.
Other published percentages and why they differ
The Triple-I statistics page states that “80 percent of insured drivers purchase comprehensive coverage in addition to liability insurance, and 77 percent buy collision coverage, based on a Triple-I analysis of 2023 NAIC data.” Restated in the 2024 supplement, the 2023 countrywide rows give 79.8 for comprehensive and 76.5 for collision, which round to 80 and 77. The figures agree if Triple-I divided collision and comprehensive car-years by liability car-years, which is the writer’s reading because its page does not give the method; the difference between the 76.3 here and the 77 there is then the year (2024 against 2023) plus rounding. Triple-I’s wording is “insured drivers”; the NAIC tables count car-years, so “drivers” is shorthand for vehicles.
Insurance.com’s 2014 release gave collision as “selected by 60 percent of all drivers” and comprehensive by 61 percent. That figure comes from online quote requests from 557,238 drivers in one month, so it describes people shopping for a policy, not the stock of insured vehicles. It is also from 2014. A gap of about 15 points between it and the NAIC’s 2020 collision figure of 75.1 is a gap of method, year and population, and it is not evidence that coverage changed.
A third kind of figure divides by all registered vehicles, which would include uninsured cars; the uninsured drivers guide reports uninsured rates by state. When a number is quoted for “the percent with full coverage,” ask three questions: what unit is counted (drivers, policies or cars), what is in the denominator (insured cars, quoted drivers or all vehicles), and what year the data describe.
Using the state shares for a decision
For a buyer the table is a base rate, not a recommendation. In a state above 85, such as New Hampshire, Maine or Massachusetts, a policy without collision is unusual among insured cars, and a quote without it is a deliberate choice. In a state under 70, such as Arkansas or Oklahoma, one in three liability cars is already carried without collision, and a liability-only quote is ordinary. In both cases the choice rests on the car, not the state.
Three checks come before dropping either coverage. The first is the contract: a lender or lessor can set requirements that the state does not, and the financed car guide and the gap insurance guide cover what financing adds. The second is the deductible, which the deductible guide covers. The third is the price of the coverage itself: ask for a quote with and without it so the difference is visible, and place the full-coverage figure beside your state’s NAIC benchmark on the state rate map.
The state shares will update. The September 2026 supplement is the newest NAIC edition found on the check date, with data through 2024; the NAIC says the full 2023/2024 Auto Insurance Database Report will follow, and a later supplement would carry 2025 data and may restate these rows, as this one says earlier-year data can differ. The table here is a benchmark for 2024, built from published tables, and it is not a forecast or a quote.
Questions readers ask
What percentage of insured cars have collision and comprehensive?
In the NAIC’s 2024 data, collision car-years equal 76.3% of liability car-years and comprehensive car-years equal 79.2%. The tables do not count cars that have both, so no single “full coverage” share can be read from them; arithmetically it is at least 55.5% and at most 76.3%, assuming every collision and comprehensive car-year also carries liability.
Which state has the highest share of cars with collision coverage?
New Hampshire, at 94.4% of liability car-years for collision and 95.9% for comprehensive, followed by Maine (86.8% collision) and Massachusetts (86.4%). The NAIC supplement does not explain New Hampshire’s lead. The lowest collision share is Arkansas, at 66.1%.
Why is Texas comprehensive coverage lower than collision?
Texas is the only one of the 51 jurisdictions where the comprehensive share (65.8%) is below the collision share (67.8%), and the NAIC says Texas comprehensive exposures are estimated from a ratio of comprehensive to collision car-years using data that cover about 70% of the market. The estimation method, not a buying pattern, may produce the gap (the writer’s reading, since the NAIC does not print the ratio), so treat the Texas comprehensive share as indicative.
Do these figures count drivers or cars?
Cars. The NAIC measures exposure in car-years, where one car-year is the risk of insuring one car for one year. A household with three cars counts as three car-years. Triple-I describes the same data as a share of “insured drivers”, which is shorthand.
Do states with more car loans have more collision coverage?
Not in this test. Across the 51 jurisdictions the rank correlation between collision share and auto debt per consumer from the New York Fed’s credit panel is -0.45, and it is -0.24 after holding median household income fixed. The debt series is a balance per consumer, not the share of cars with a lien, so the test cannot show whether lender requirements work vehicle by vehicle.
Are the collision shares rising?
Slowly. Between 2020 and 2024 the countrywide collision share rose from 75.1% to 76.3%, and it rose in 47 of the 51 jurisdictions and fell in four: Florida, Vermont, Texas and Arkansas. Maine rose most, by 9.0 points. The ranking of states is stable, with a rank correlation of 0.96 between 2020 and 2024.
Does this tell me whether I should keep collision and comprehensive?
No. The shares are a benchmark of how common the coverages are, not advice. Whether to carry them depends on the car’s value, the deductible, your savings and any lender requirement. The liability-only against full coverage guide and the deductible guide lay out the decision, and a quote with and without the coverage shows its price.
Sources, checked 2026-11-04
- NAIC, 2024 Auto Insurance Database Average Premium Supplement (September 2026): Introduction, Terms and Calculations, Technical Notes, Additional State Information, and Tables 1A to 5 (2020 to 2024, state and countrywide rows), checked 2026-11-04.
- Triple-I, Facts + Statistics: Auto insurance (coverage shares from 2023 NAIC data; 2024 claim frequency and average claim), checked 2026-11-04.
- Insurance.com, Anatomy of a car insurance buyer (press release dated August 13, 2014; quote data for July 2014), checked 2026-11-04.
- Federal Reserve Bank of New York, State Level Household Debt Statistics 2003-2025 (February 2026), auto debt balance per capita, Q4 2024 column; Consumer Credit Panel / Equifax, 5% sample, checked 2026-11-04.
- U.S. Census Bureau, Table H-8 Median Household Income by State: 1984 to 2025 (CPS ASEC), 2024 column, checked 2026-11-04.
- MoneyGeek, When to Drop Collision and Comprehensive (Full Coverage) (updated October 11, 2026; the 10% rule is one of four conditions listed), checked 2026-11-04.