The behavior of Volkswagen has been heinous and the company and probably some of its executives deserve some serious punishments. Yet our reaction to the scandal is as illuminating as the misbehavior itself. We get much more upset when people do wrong out of deliberate fraudulent intent rather than through accidental negligence, or sheer inability to solve problems, even if the latter phenomena are often the greater risks.
The falsification of Volkswagen emissions software has meant more nitrogen oxide in the air, but how costly is this extra pollution in economic terms? One plausible estimate suggests this additional pollution has been killing 5 to 27 Americans each year, with that number worldwide reaching up to 404 as a maximum.
To put that number in context, the World Health Organization estimates that about seven million people die each year worldwide from air pollution. Even within the United States, early deaths from air pollution have been estimated to run about 200,000 a year, in comparison to which the losses from the Volkswagen scandal are a rounding error. For the American deaths, however, the culprits are often cars, trucks and cooking and heating emissions, so there is no single, evil, easily identified wrongdoer at fault. As Pogo recognized, often the real enemy is us.
Although the practice is ethically controversial, some economists believe we can attach dollar values to human lives. A typical value of life, estimated by this method, might run in the neighborhood of $7 million. That would mean Volkswagen has been destroying perhaps around $100 million in value a year. To put that number in context, a single Picasso painting can cost that much, or a Hollywood studio might spend (waste?) that much money marketing a single blockbuster movie.
Admittedly, the individuals and families who lose those lives don’t view the matter in such abstract, impersonal terms, but still if we had an extra $100 million we could save at least 5 to 27 lives through safety investments in other areas; in that sense the use of the figure is meaningful.
We need to ensure that deliberate corporate fraud does not spread, but in the meantime let’s not forget that is not always the biggest problem. It just bugs us more.
The real problem is that everyone dies?
There are two messages here. The first is that the VW's fraud is not a big deal for two reasons. First, after putting the numbers of deaths in the appropriate context, they are just a rounding error. I disagree with this approach because there is nothing in economics, that I'm aware of, to suggest that the proportion of deaths matters whatsoever. If it did the most important discussion would be about the most appropriate denominator (is it the number of deaths from air pollution worldwide or just in the US?).* The larger the denominator the greater the chance we'll give VW a pass due to their fraud being simply a "rounding error."
The second reason this isn't a big deal is because we all cause air pollution every day, am I right or what? Through our own inability to stop killing people by our daily driving and cooking, we are just as guilty as VW! I disagree with this argument because air quality regulations are designed to reduce deaths when the benefits of doing so are less than the costs. Cowen is confusing the positive and negative portions of the net benefit curve. Here is a picture that illustrates a situation where lots of people die (almost everyone eventually) and it is economically efficient to avoid those from segment a to b on the net benefit line. This segment contains those deaths that VW was supposed to help avoid by installing working pollution reduction equipment on the cars that they sold. The line segment b to c, where the costs of avoiding deaths exceed the benefits, is what Cowen, for some reason, is trying to get you to focus on ("the real enemy is us").
The second message is that there is an objective way to measure the costs of VW's fraud. This is the approach that the economic community should be taking, instead of distracting people with phony arguments, and that I took in this post. Cowen and my estimates are similar. I use a range of VSL estimates with a midpoint of $7.5 million, Cowen uses $7 million. The death estimates are different but Cowen points to the estimate that I use in his follow-up post at Marginal Revolution. My bottom line is that the total cost of VW's fraud is about $500 million (vs. Cowen's annual cost of $100 million). Unfortunately, Cowen also pooh poohs this estimate by noting that, to be the number in context, the annual cost of premature death is worth about a Picasso. But that context is not appropriate. The appropriate comparison is benefit to cost. Also, the comparison is distractingly cute and that's the point. So what if we lose a Picasso, there are still are large number of masterpieces left! Again, we're confusing the marginal with the total.
VW tried to cheat to get a competitive advantage. "Let’s not forget that is [snip] the biggest problem."
*There might be something in behavioral economics that finds that small proportions of anything have little value. This might be an explanation for the infamous Exxon-financed birds study (even large changes in small proportions have little value differences ... ergo, no scope!). So, to frame the issue as a proportion makes total sense if you would like to make the problem seem minor. I'm not exactly saying that is what is going on here but it is consistent with an earlier MR post (the everybody does it defense of VW).