How we calculate what Orange County cities drive

The data

Everything starts with two public sources:

• California DMV vehicle registrations. Every January, the DMV publishes a count of registered vehicles by ZIP code, model year, fuel type and make. Our city pages use the file as of January 1, 2026, the most recent available.
• California Energy Commission new-car registrations. The CEC tracks every new vehicle registered in Orange County by make, model, model year and fuel type, with nothing hidden. We use 2020 through 2025.

For national comparisons, we use US new-car sales by brand for 2021 through 2025, from the Automotive News Research & Data Center and automakers' year-end reports. Tesla doesn’t publish US sales, so its figures are industry estimates (from Cox Automotive and Automotive News).

The gap we had to fill

Starting with its 2024 file, the DMV’s public data leaves the make blank for about a third of vehicles. Before that, it named the make for nearly every car. Small brands lose their names far more often than big ones, so counting only the named cars would make Toyota and Honda look bigger than they are and shrink everyone else.

What the DMV does still publish exactly is how many vehicles each ZIP code has, by model year and fuel type, including the unnamed ones. And two other things are known: the full make mix in every ZIP code as of January 2023, the last near-complete file, and the exact countywide mix of new cars from the Energy Commission.

How we fill it in

Our model gives each ZIP code’s unnamed cars a make, using a standard statistical technique called raking (iterative proportional fitting):

1. Start with each ZIP code’s own history. The first guess for a ZIP code’s unnamed cars is that ZIP code’s make mix in January 2023, for cars of similar age and fuel type. 2. Match the county totals. For newer cars, we know exactly how many of each make Orange County registered, from the Energy Commission. For older cars, we use the 2023 file. Subtracting the cars the DMV still names tells us how many unnamed cars of each make are out there countywide. 3. Balance both sides. The model adjusts the first guess until every ZIP code’s total and every make’s county total line up exactly. The cars the DMV does name are never changed. Only the blanks are filled in. 4. Add up by city. ZIP codes roll up into cities using official city boundaries, so Corona del Mar and Newport Coast count as Newport Beach, for example.

How we know it works

We tested the model on January 2023, when the DMV still named nearly every car’s make, so the right answers are known. We hid makes the same way the 2024 and later files do, ran the model with the information it would have had (data from three years earlier, the same gap as today), and compared the results with the real numbers.

• It identified 96% of each city’s top five brands (149 of 155, across 31 cities).
• For the biggest brands, the typical miss was less than half a percentage point of a city’s vehicles. For Mercedes-Benz, BMW, Tesla and Hyundai, it was under two-tenths of a point.
• It was far more accurate than counting only the named cars, typically two to seven times closer to the truth.

How we write about the numbers

• “Newer cars” means model years 2021 and later, registered in the city, whether they were bought new or used.
• National comparisons set each city’s newer cars against the national mix of new-car sales over the same years. Very few cars are scrapped in their first five years, so five years of sales closely matches the mix of newer cars on the road.
• Orange County comparisons set each city against the 34 Orange County cities with at least 3,000 newer cars.
• We only say a city has “more than any other” when it leads the next city by a clear margin. When cities are close, we say “one of the highest.”
• When two brands are within our margin of error, we call them about even.
• Electric-car shares come straight from the DMV’s fuel data, which has no gaps.

What the data can't tell us

• Models. The DMV’s public file has no model field, so we report brands, not models, at the city level. Countywide model rankings come from the Energy Commission’s data.
• New vs. used purchases. A registration doesn’t say whether a car was bought new or used, so we report what’s registered.
• Corporate fleets. Companies sometimes register cars at their headquarters. Irvine is home to the US headquarters of Mazda, Kia and Rivian, and where it matters, our city pages say how much of a brand’s presence looks like company cars.
• A small remainder. Even after filling in the blanks, about 1 in 10 newer cars can’t be confidently assigned a make, about the same share as in the DMV’s complete years. That remainder is mostly smaller brands, such as Chrysler, Buick, Mini, Alfa Romeo and Maserati, which have too few cars in each ZIP code for the DMV to report them by name.

For reporters and researchers

We’re happy to share the city-by-brand figures behind these pages, along with more detail on the method. Email jason@shopbestest.com.