Guide Updated 2026-10-30
Why cars fail a smog check in California
About 8.5% of California cars failed their first smog check in 2025, and older cars fail far more often than newer ones.
How many California cars fail a smog check
About 1 in 12. In 2025 California ran 11,275,981 initial smog checks and 959,667 failed. That is 8.51%. The word "initial" matters. It means the first test of a car in a cycle, not a retest.
Share of California initial smog checks that failed, 2019 to 2025
The failure rate has fallen every year since 2019, from 10.1% to 8.5%. BAR's tables do not give a reason. The 2026 numbers so far, January to September, show 8,633,014 initial tests and a 8.31% failure rate.
Two things shape that number. One is how old your car is. The other is what kind of test it gets. Let's take them in turn.
The biggest factor is model year
Older cars fail more. It is not close. Cars that go through the OBD style test fail at 7.8%. Cars that get the tailpipe test fail at 17.9%. That is more than twice the rate.
Initial fail rate by model year, 2000 to 2025 cars
Look at the slope. A 2000 model year car failed 16.9% of the time in 2025. A 2005 failed 11.2%. A 2009 failed 7.8%. A 2013 failed 6.4%. A 2017 failed just 4.2%. After that the rate sits between about 4 and 6%. New cars rarely fail. When they do, it is often an OBD or readiness issue.
Initial fail rate by model year, 1976 to 1999 cars
The 1976 to 1999 cars fail about 16 to 21% of the time, and this group is the toughest. In 2025 the 1976 model year failed 21.1%. The 1999 model year failed 15.8%. Most of these cars fail on the tailpipe test itself.
Where the failures come from
Rate is one thing. Volume is another. Most California cars on the OBD style track are mid age, so they supply a lot of the failures even at medium rates. Here are the five groups.
| Model years | Initial tests | Failed | Fail rate | Share of OBD style failures |
|---|---|---|---|---|
| 2000 to 2005 | 1,871,410 | 250,975 | 13.4% | 31% |
| 2006 to 2010 | 2,022,089 | 187,560 | 9.3% | 23% |
| 2011 to 2015 | 3,510,514 | 230,140 | 6.6% | 28% |
| 2016 to 2020 | 2,431,603 | 115,987 | 4.8% | 14% |
| 2021 to 2025 | 615,778 | 28,443 | 4.6% | 3% |
The oldest group, model years 2000 to 2005, fails the most often at 13.4%. It also supplies the biggest share of failures, 31%, from only about 18% of all OBD style tests. The 2011 to 2015 group fails far less often, yet it supplies 28% of failures because so many of those cars are on the road. The newest five model years supply only 3%.
If you drive a car that is 10 to 25 years old, you are in the groups that fill most of the failure column. Set aside time and money for a repair before your due date, and test early in your 90 day window.
What exactly fails
The test has four main parts, so a failure can come from any of them. A car may also fail more than one part, so the rows add up to more than the total.
| Failure type | OBD style tests (2000 and newer) | Tailpipe tests (1999 and older) |
|---|---|---|
| Emissions (tailpipe) | Not measured | 10.2% |
| Functional | 6.9% | 8.9% |
| OBD | 6.8% | 9.4% (1996 to 1999 cars) |
| Readiness monitors | 4.1% | Not listed |
| Visual | 1.3% | 3.1% |
| All failures | 7.8% | 17.9% |
On 2000 and newer cars, the biggest buckets are the OBD scan and the functional checks. The readiness monitors add 4.1%. Visual checks catch very few. On older cars, the tailpipe is the leading reason, at 10.2%. See check engine light and smog and readiness monitors for those two big causes. The test walk through explains each step.
Reasons BAR and the Referee name
- The check engine light is on. This alone fails the car.
- Incomplete monitors after a battery change. Drive one to two weeks first.
- Modified software. Not allowed unless it has a CARB Executive Order.
- Aftermarket or wrong catalytic converters. A 2000 to 2007 Honda Accord SULEV, for one, needs the factory part.
- Gas cap failures on older cars. 1976 to 1995 cars get a fuel cap and a fuel vapor pressure test.
What do roadside checks add? When BAR and CARB look at failing cars on the roadside, the top trouble code group is emission control (P04xx) at 39.9%. Air and fuel metering is next at 25.3%, then ignition and misfire at 15.3%.
Do STAR stations fail more cars
Yes, but that is mostly about which cars go there. BAR sends cars more likely to fail, and a random 2% sample, to STAR stations. So STAR stations see a tougher mix. Non-STAR stations test the easier cars. Referee stations are different again. They take problem cases and disputes, so they fail most of what they see.
| Station type | OBD style tests | Fail rate | Tailpipe tests | Fail rate |
|---|---|---|---|---|
| STAR Test Only | 5,916,245 | 8.1% | 492,244 | 17.6% |
| STAR Test and Repair | 2,766,051 | 8.2% | 236,209 | 20.1% |
| Non STAR Test Only | 375,508 | 6.1% | 18,807 | 10.9% |
| Non STAR Test and Repair | 1,351,143 | 5.7% | 57,373 | 13.1% |
| Government Fleet | 22,007 | 5.2% | 7,604 | 17.7% |
| Referee | 4,419 | 71.8% | 1,287 | 61.8% |
Do not read this as one type of station being easier. Read it as a sorting effect. A STAR notice on your DMV renewal means BAR flagged your car. Read more in our STAR station guide.
Fail rates by program area
Where you live matters less than what you drive. The OBD style fail rate sits within a point of 7.6 to 8.1% in every area. The tailpipe fail rate is higher in enhanced areas (about 18%) and in change of ownership areas (about 22%), and lower in basic areas (about 14%).
| Program area | OBD style tests | Fail rate | Tailpipe tests | Fail rate |
|---|---|---|---|---|
| Enhanced | 9,035,335 | 7.8% | 674,992 | 18.2% |
| Partial Enhanced | 523,295 | 7.9% | 40,699 | 19.8% |
| Basic | 839,586 | 7.6% | 92,938 | 14.3% |
| Change of Ownership | 63,923 | 8.1% | 5,213 | 22.3% |
Change of ownership areas test cars mainly when they are sold. Enhanced areas use the loaded dynamometer test, which measures NOx.
What happens after the first test
A failure is not the end. In 2025 BAR counted 12,200,718 tests of all kinds, including retests. That leaves 924,737 follow up tests after the 11,275,981 initial ones. Of those follow up tests, about 142,670 failed, which is 15.4%. Most follow ups are passes. Still, about 1 in 6 cars that come back fail again. Those owners often need a bigger repair.
Follow the step by step plan in failed smog check. Costs can add up, so also see smog check cost and CAP repair help.
Does a smog check catch every dirty car
Not every one. BAR and CARB compare smog check results with roadside inspections in their performance report. Here is what the 2026 report found.
- About 14% of 2000 and newer cars failed a roadside inspection within a year of their smog check in 2024 and 2025. That is down from 19% in 2018, adjusted for fleet makeup.
- The smog check fail rate rises from about 4% at 8 years old to about 15% at 25 years. The roadside rate rises from 5% to 34% over the same ages.
- Of cars that failed, were fixed and were certified, 65% failed roadside within a year. That is a small sample, only 125 cars.
- BAR and CARB estimate that 44 extra tons per day of ROG plus NOx could have been cut in 2025 through program improvements.
The program also pursues bad actors. Smog data led to 1,534 case filings with the Attorney General from 2016 to 2025, and 115 in 2025. Read about illegal smog check stations if you worry about a clean-pipe or fake test.
What these numbers do not tell you
Good data has limits, and you should know them.
- No make or model data. BAR does not publish fail rates by make and model. You will see claims online about which brands fail most. The public record does not back them up. Only model year, station type, program area and failure type are public. A records request would be needed for more.
- Initial tests only. The headline rate counts the first test. A car that fails, gets fixed and passes is still counted as a failure in these tables.
- One car can fail more than one part. That is why the cause rows add up to more than the overall rate.
- Reported repairs are a floor. Repair costs cover only the repairs that stations log in the BAR system.
- Mix matters. A station type with a high fail rate may simply get the harder cars.
One more point on the newest cars. A 2025 model year car failed 5.6% of the time, a bit higher than the 2022 to 2024 model years. The group is small, with about 43,000 tests, so a few cars can move the rate. Brand new cars are exempt from the renewal check for their first eight years, so the ones that do get tested may be tested for another reason, such as a sale or a move into the state.
How to lower your odds of failing
- Heed the light. If the check engine light is on, fix it first.
- Let monitors finish. Do not reset the car right before a test.
- Know your warranty. A young car may have free coverage.
- Keep the factory parts. Especially the catalytic converter and engine software.
- Test early in your window. You can test up to 90 days before it is due, which leaves time to repair.
Questions people ask
What percent of cars fail a smog check in California?
In 2025, 8.5% of initial smog checks failed, which was 959,667 of 11,275,981 tests. The rate was 10.1% in 2019.
Which model years fail smog the most?
The oldest. In 2025, 1976 to 1995 cars failed about 19 to 21% of the time. Among 2000 and newer cars, the 2000 model year failed 16.9% and the 2017 model year just 4.2%.
What is the most common reason for failing a smog check?
On 2000 and newer cars, the OBD scan and the functional checks lead, with readiness monitors a big share. On 1999 and older cars, the tailpipe emissions test is the top cause at 10.2%.
Do STAR stations fail more cars than other stations?
Yes, 8.1% against 6.1% for non-STAR test-only stations on OBD style tests. BAR directs cars more likely to fail to STAR, so the mix of cars is different.
Why do Referee stations fail so many cars?
Referee stations see problem cases and disputes. They failed 71.8% of OBD style initial tests in 2025.
Does failing once mean a big repair bill?
Not always. About 925 thousand follow up tests happened in 2025, and most passed. In BAR data, about 4 in 10 logged repairs cost over $650.
Can BAR tell which makes and models fail most?
BAR does not publish fail rates by make and model. The public numbers cover model year, station type, program area and failure type.
Do cars that pass a smog check still fail on the road?
Some do. BAR and CARB found that about 14% of 2000 and newer cars failed a roadside inspection within one year of their smog check in 2024 and 2025.
Is the fail rate going down?
Yes. It fell from 10.1% in 2019 to 8.5% in 2025, and January to September 2026 ran at 8.31%.