At this point, the entire world is aware that Volkswagen programmed about a half-million diesel-powered cars to fully operate their emission controls only while being tested. What remains far than clear is why they did it. We assume they did it to sell more cars and make more money. But did they?

It might seem obvious why they cheated: for the same reason people usually cheat, to pass a test. But the cars do pass the emissions test as long as their pollution controls (of either type) are operating. So why have them operate only when the cars are being tested?


One possibility: fuel economy. This is clearly a factor with the lean NOx trap (LNT) system used in the Jetta, Golf, and Beetle. For the trap to operate properly, it must be kept at a fairly high temperature. With the latest, highly efficient diesels, the engine’s exhaust flow often isn’t hot enough to maintain this temperature. To correct for this, more fuel must be injected than the engine really needs. As I understand it, this harms fuel economy, but not performance.

The thing is, the EPA’s fuel economy tests are the same as the EPA’s emissions tests. So the window sticker mpg would have been calculated with the emissions controls properly operating. The cheat did not improve these figures. It should improve real-world MPG. Was Volkswagen counting on word-of-mouth to spread real-world MPG figures that were considerably better than EPA figures, and thus sell more diesel cars?

The Passat used a different diesel exhaust fluid (DEF) based system that, as far as I can tell, would not require extra fuel to properly operate. Deactivating this system in real-world use might have only reduced the consumption of DEF. DEF costs $5 to $15 per gallon, depending on the brand and source, and a gallon lasts about 2,000 miles in cars with properly functioning systems. The cost of the fluid itself doesn’t seem significant. The hassle of refilling could have been more of a concern, but something worth the risk of cheating the EPA regs?


Maybe there is a performance impact I’m not aware of. But would it have had much impact on sales if the cars were a little less powerful? This doesn’t seem likely. More likely: the engineers felt they had to meet specific performance goals, possibly including real-world fuel economy, and these could not be met with the controls operating. They were ordered to “make it work,” and they did.

Maybe Volkswagen deactivated the controls during real-world use because constant use would cause them to fail, creating both reliability and regulatory problems (since the emissions controls must be effective for 120,000 miles)? Possibly, but I have not been able to uncover any evidence for this.

It’s possible that the cheat was developed for a reason that no longer exists, but that the cheat was left in place anyway. We tend to view past decisions based on knowledge and circumstances that arrived later. In reality, the paths that generate actions, especially corporate actions, can be quite curvy and indirect.

But until new information comes to light, I just can’t find a good motivation for what VW did.

Perhaps I’m thinking too rationally. An excellent article in the New York Times blames Volkswagen’s arrogant, insular culture for a willingness to cheat—while still assuming there was some rational motivation to cheat. But could this culture itself be largely to blame? The article notes that VW is dominated by engineers, and that many of these engineers felt that the EPA’s regulations were unfair to their beloved fuel-saving technology, diesel. These engineers, with their love for efficiency, might have also felt it did not make sense to burn more fuel, and produce more CO2 in the process, just to reduce NOx. Could the code that defeats the pollution controls in normal use have been their way of protesting what they saw an unfair, irrational EPA requirements, while appearing to meet these requirements? Could the cheat have been just a secret (until now) middle finger from some VW engineers to the EPA?

Michael Karesh operates TrueDelta, a provider of car comparisons, including reliability stats, pricing, and specs.