Sam Peltzman is an economist who popularized what became known as the Peltzman Effect – that safety regulations (may) induce riskier behavior. The canonical example is in cars. He showed (Journal of Political Economy, 1975, “The Effects of Automobile Safety Regulaton”) that additional safety devices (regulated or voluntary – he focused on regulation) caused drivers to engage in riskier behavior. In specific, he showed that the number of non-fatal car crashed increased as did the number of pedestrian deaths due to car crashes.
Why does this happen? As with anything, there is a cost-benefit trade-off. The benefit to riskier (read: faster, more aggressive) driving is getting to your destination faster – most people view time as a valuable resource. The costs are somewhat obvious - higher risk of accident. Specifically, there are the costs of personal injury to you, the driver, other passengers, and the cost of replacing your car. Of those, given car insurance, I think it is safe to say that the cost of personal and passenger injury is by far the major cost. With better crumple zones, airbags, and seatbelts, we have lowered the cost of an accident given that an accident happens. The problem is that by inducing riskier behavior, we increase the probability of an accident.
Here is a little math for the geeky probablists among us. Those who are math-phobic may want to skip this – it is intended to help people understand, not confuse. Let P(Injury | Accident) be the probability of Injury given an Accident. Let P(Accident) be the unconditional probability that an accident happens. Let’s ignore the cost of fixing your car or increased insurance payments, and let’s assume that the hospital bills and other personal costs (monetary or not) are constant. The, the probability that you are injured is P(Injury | Accident)*P(Accident). So, if additional safety features reduce P(Injury | Accident) then your cost has gone down. If your Overall Benefit = Benefit of Time Saved – Cost of Accident then you are better off! The problem is, accidents are rare, but you get benefits from time saved every time your drive. A whole bunch of small benefits (2-4 minutes?) of saving time is nice. So what do you do? You drive a bit faster, Overall Benefit stays the same because while you increase Benefit of Time Saved but you simultaneously increase Cost of Accident by increasing P(Accident).
Here is a personal experiment. If you are someone who typically drives with your seatbelt on (I do, I hope you do also), then experiment with driving without it for a while. I know I’ve occasionally skipped it for a quick trip to the store – or at least I started to; after a while, I felt a bit naked without it. As soon as I put it on, I felt more secure. That – that ‘secure’ feeling – is a little glimpse of what causes this problem. You feel more secure than you should. Gordon Tullock, another economist, has tongue-in-cheek proposed an “airbag” which is a 10″ spike that comes out of your steering column in the case of an accident. While nobody is proposing that in reality, it sure would make for safer drivers! By making P(Injury | Accident) = 1 (you face certain death in case of accident) you are fully incentivized to make P(Accident) as low as possible. And everyone is safer. That is because there are Externalities here as well: the pedestrians. They gain nothing from you driving faster, they only bear the cost of more reckless driving. So, we could say that increased automotive safety has increased the risk to pedestrians due to the externality. Safer drivers also mean safer pedestrians.
OK, that was sort of long. Why did I go through all of that? Because it relates to Financial Regulation. Do we want a “safer” financial system? I’m not sure we do. One of the things we found out for sure is that Wall Street bankers took enormous risks because they had at least the perception of safety. While some certainly may have knowingly sold dodgy securities, most of the purchasers (current Justice Department investigations nonwithstanding) were themselves sophisticated investors who should have known better. Everyone felt safe – markets were rising, business was good, no need to read all that fine print – look at how great this CDO has performed over the past 6 months! Investors bought all manner of securities they didn’t really understand because it seemed fine at the time; they also made a bit more money off of a AAA CDO than they did off of a AAA Government Bond. They take the small benefit now (saving a few minutes driving) while increasing the probability that something bad would happen – housing bubble and then collapse. Imagine each Wall Street bank as a bunch of drivers on the highway who, over time, find that they can drive faster and faster and closer and closer with quicker lane changes and fewer signals. There were a few bust ups along the way – Long Term Capital Management in 1998, and then a bit closer to home Bear Sterns collapse in March 2008. But each time, the creditors lost nothing – i.e the drivers got new cars. Until we had the Big One in Sept 2008 – a massive, multi-car pileup with injuries.
Our capitalist system requires that failure is an option. If it is not, all we can do is prepare for more and more risky behavior. If failure is not so bad, then why not take bigger and bigger gambles? Why did bank profits come right back; and at record levels? They got bailed out – and with fewer competitors to split the pie than before. We need a system that allows both success and failure. Right now, once big enough, that doesn’t apply any more; and I’m not sure it is limited to big banks. Passing legislation to make our financial system “safer” to me is the wrong direction. We need a system that is riskier, but at a smaller scale. We need a system with smaller banks, any of which can fail at any time.
However, I’m skeptical that this will happen. The current financial regulation bill about to be passed does nothing really, which is chronicled well at Baseline Scenario by Simon Johnson and James Kwak. I’m a bit afraid that we’re headed toward bigger business and bigger government, which to me simply says we’ll have fewer blowups, but they will be bigger in scale each time. I’d rather have more frequent, but smaller crises.