I used a method to get auto maker reliability indicators from ordinary classified listings. This is more like a population study, lacking scientific controls.
Pull a list of old cars. Break down by manufacturer.
Pull a list of middle-aged cars, after breaking down by manufacturer, use this to normalize the list of old cars.
Overall, I have to say, it would be more interesting to write some software to process more data than I did, and use a better control than middle-aged cars. Import or sales data would be ideal.
Utah (ksl.com) favors: Honda (by a landslide), Toyota, and Subaru. Last place went to Dodge, Hyundai, Kia, and Audi. Hyundai and Kia may have appeared to perform poorly just because they are young brands. (No study controls.)
I didn’t pull enough data to represent a lot of makers, however. 1514 middle-aged cars, and 199 old cars. The stats were sufficient only for popular brands. I was hoping to see some data on Suzuki, but Suzuki is just not a popular enough brand in this Utah in order for me to comment on. I deliberately excluded trucks and SUV’s. I’ve heard some people think that Ford is a reliable brand, for example. Since I’m ignoring trucks such as the F-series, I see that the Ford cars were not high performers. They landed in the middle of the range, at best.
The Pontiacs in the old car dataset were all Grand Am’s, and probably the work of public attitude toward the late 90’s Grand Am, rather than the quality of the Grand Am.
My sample sizes in Israel (yad2.co.il) were possibly too small, but the best makers seemed to be Subaru, Toyota, and maybe Suzuki and Mercedes. I had only 51 old cars to work with. Colleagues have told me that Israelis don’t take car of their cars, and it’s true. Utah has 13% as many old cars as middle-aged cars, Israel has less than 2%, according to the classified listings. It seems bizarre, since Israel is taxed at a rate of 99% on new car imports. I didn’t make any effort to exclude trucks and SUV’s from the Israeli data, but there probably were none, or very few. Usually if I just search for cars, I get compacts and city cars (they refer to these classes as “family size” and “minis”.)
In Israel, Ford cars fared the worst by far. Or this could be explained if there were a change in import rates from the US after 2000. I don’t know.
There were three major brands that were quite different between the datasets. Mercedes could be a statistical anomaly, but then so could Suzuki. Mercedes was just an average make in Utah, landing right at the 13% mark, and it was more statistically significant in Utah.
The biggest surprise was how poorly Honda seems to fare in Israel, especially considering how incredibly well they do in Utah. I can think of ways to explain this (especially import change, the lack of controls again), but I’m really not certain.
I was also surprised to see Mazda close to the bottom in both regions. But then, I wonder if some Japanese carmakers have a halo of reliability, just because of Honda and Toyota. I haven’t found any indicator that Mazda is an exceptionally reliable brand in real life.