Regulatory Stupidity - an update

A while ago, I commented on the idiocy of spending $3.2 billion a year to save the lives of 103 people. At the end, though, I noted a problem:

Yes, I know my math ignores the injured people. I thought about this for a while and couldn't figure a reasonable way to allow for it. "Injuries" is extremely vague, and could include anything from a cut requiring stitches to full-on, almost-fatal intensive care treatment.

Well, after further thought, I figured it was reasonable to divide the total $3.2 billion cost evenly in half, assigning half of it to the deaths, and the other half of it to the injuries. That would mean deaths would rather arbitrarily be assigned about 70 times the value of injuries. It's impossible to avoid a level of arbitrariness in this, but a 70-fold difference seems fairly reasonable to me.

Put that way, then, you divide the original $3.2 billion in two and then calculate as before for each half. Under that model, cost per injury comes to about $207,000, and cost per death comes to roughly $15 million. The $15 million figure isn't the $31 million of the original estimate, but it's still a huge amount of money to spend on saving the life of one person, especially when having people buy these cameras voluntarily would probably accomplish a significant fraction of the same goal.

More interesting is the $207,000 figure for each injury. Obviously if there was some data available on the number and types of injuries, it might be possible to say that X amount was being spent to prevent (for example) cuts and bumps requiring stitches, Y amount to prevent broken bones and concussions, and Z amount to prevent more serious injuries, perhaps requiring intensive care. But since that data isn't available, I've settled for the blanket amount. It is actually quite large, if you make the very reasonable assumption that most of the injuries would be less severe, that is, cuts and bumps requiring stitches, rather than intensive care type injuries.

I also remembered one other thing since writing the original article: A lot of car models in recent years are designed in a way that results in very poor rear visibility. There is the stylistic tendency to make side windows swoop upwards toward the rear of the car in 5-door models, for instance. Big SUVs shaped similarly to pickup trucks are inevitably going to have poor rear visibility. Even current sedan designs seem to feature trunks in back which stick up quite high compared to older models, which certainly would not help. There are also a lot of models where the rear window is actually pretty small, enough so I can't imagine rear visibility would be all that great.

In other words, better design could alleviate a lot of this problem, with no necessity for cameras.