What the Peltzman Effect Is and Isn't
And a personal story.
Peltzman’s second major contribution to the understanding of the unintended effects of regulation was his 1975 study of the effects on traffic safety of a slew of US National Highway Traffic Safety Administration regulations on the design of cars. In the mid to late 1960s, the federal government made a number of safety features mandatory. These included seat belts for all occupants, an energy-absorbing steering column, a penetration-resistant windshield, a dual braking system, and a padded instrument panel. In his study, Peltzman stated that the goal of the mandates was to reduce traffic fatalities and serious injuries sustained as a consequence of vehicle accidents. But he found something different. Fatalities were not reduced at all. Instead, deaths of vehicle occupants fell but those of pedestrians and motorcycle drivers rose. Peltzman’s tentative explanation was that by reducing the probability of being killed in a given accident, the mandates caused drivers to drive more “intensely.” His finding became so well known that economists started referring to the “Peltzman effect.” Later studies found that drivers with anti-lock brakes tended to follow the cars in front of them more closely. A 2010 study of NASCAR accidents found that the “mandated use of a head-and-neck- restraint system has almost completely eliminated serious driver injury, while simultaneously increasing the number of accidents per race” (Pope and Robert D. Tollison, 2010).
This is from David R. Henderson and Steven Globerman, The Essential UCLA School of Economics, Fraser Institute, 2021.
Here’s a good definition of the “Peltzman effect:”
According to the Peltzman Effect, when safety measures are implemented, people’s perception of risk decreases, and so people may feel that they can now afford to make riskier decisions. As a result, the phenomenon predicts that mandatory safety measures actually experience a lower benefit than we would expect, because the safety benefits brought about by these measures are offset to some extent by increases in risky behavior.
This is from “The Peltzman Effect,” The Decision Lab.
By the way, I attended the UCLA workshop where Sam first presented his results. I don’t recall Sam ever saying that the benefits were lower than we would expect. I actually tried, in questioning him, to address that. My argument at the time was that if you get to drive more intensely, that is a benefit. For one thing, it saves time. We don’t measure costs and benefits just by looking at fatalities and injuries. Sam was focused on whether the legislation achieved its goal; he was not doing a cost/benefit analysis.
Notice what the definition above doesn’t say. It doesn’t say that the offsetting behavior will always be large enough that it will completely leave injuries and/or fatalities unaffected. It says simply that there will be offsetting behavior. It’s actually an application of the law of demand: when the price of something falls, not due to a decline in demand, we buy more of it. To take an imperfect analogy, nothing in the law of demand says that when the price of something falls, we would buy so much more of it that we would spend more; that would require that our demand be elastic at the pre-existing price.
I’m writing this to respond to some of the skepticism that commenters expressed on a recent blog post by my EconLog colleague Benjamin Seevers. Fortunately, commenters were not skeptical of the idea that there was any offsetting effect, but of the idea that the effect would be as strong as the one that Peltzman found. [Note: I rewrote this in response to Dylan’s comment below; I was too careless in reading the comments and had though that that skepticism seemed to carry over into even questioning whether there was any offsetting behavior.]
Here’s a true story I told my Naval Postgraduate School students years ago when I covered this issue and met with skepticism.
In the early 1990s, my wife, daughter, and I were driving at the start of a vacation. I was driving my wife’s 1990 Camry wagon at about 65 to 70 mph in a 65 mph zone. My daughter, who was about 7, was sitting in the back. She realized that she had packed a book she wanted to read in a suitcase in the trunk. If she were to unlock her seatbelt, she could reach back and get it. She asked permission to do so. My wife and I consulted briefly and decided that she could but that I should drive “less intensely.” So I slowed to 55 mph and looked around even more hawkishly than was my wont.
Many of the students had an “aha” moment. They could imagine themselves doing the same thing. With less internal security, I compensated with my driving behavior.
Personal note: I was lucky enough to take Sam Peltzman’s 2-quarter sequence in Industrial Organization at UCLA in my first year there and Sam’s last year there. It was one of the 5 or 6 best experiences of my time there.
Feb 13 2023 at 5:42pm
As one of those who expressed skepticism to that post, but explicitly said:
I didn’t see any questioning of whether there is any offsetting behavior by me or any of the other commentators, just whether that offsetting behavior eliminated the safety benefit of the additional equipment. And, I think that is likely a case by case determination.
Benjamin overall wrote a great piece (certainly better than anything I wrote during my undergrad days, and probably better than anything I write now), but there were a couple of parts that didn’t quite add up. For instance:
I read the first part of the statement to mean something like, with the first piece of safety equipment you might be 50% more reckless, with the 2nd you’re another 25% more reckless, but at some point you kind of top out (like, how much more reckless can those Fast and Furious drivers really get after driving their cars out of an airplane??) But then, the second part doesn’t follow. If you’re at the point where the added safety equipment doesn’t add a recklessness penalty and does add a safety benefit, it doesn’t necessarily make sense that you would want to remove all protection.*
*I’m assuming that safety is the only thing we’re interested in for this example, but of course we could want to remove safety equipment because it makes the game more enjoyable to watch or more fun to play, etc…
Feb 13 2023 at 6:31pm
Thanks, Dylan. My error.
Feb 14 2023 at 12:55pm
Thanks David, appreciate the update.
Just for matters of completeness, I wanted to add that I think, in some instances, the offsetting behavior can be more than the safety benefit. Think of the driver that drives faster in icy conditions than they otherwise would because they have AWD. I just think that it is hard to know what the effect is going to be in any individual case a priori.
Feb 13 2023 at 5:45pm
IIRC, Peltzman said that if you want to get people to drive really safely, put a sharp spike in the middle of every steering wheel.
Feb 13 2023 at 5:48pm
I believe that was Gordon Tullock, not Peltzman
Feb 13 2023 at 6:02pm
You are correct, sir.
I sit here duly chastised, in my topple-proof double-cushioned safety seat.
Feb 13 2023 at 9:04pm
I know they call it the Tullock spike, but Steven Landsburg attributes the idea’s origin to Armen Alchian in his book, The Armchair Economist:
Feb 13 2023 at 9:36pm
Hyman Minsky’s financial instability hypothesis, stability begets instability, is the stock market version of the Peltzman effect, where he suggests prosperity begets recessions and bull markets beget crashes in an unending cycle of “Minsky moments” (a term coined by Paul McCulley to refer to the 1998 Russian Financial Crisis).
Feb 13 2023 at 5:59pm
Researching the origin of the “spike” idea, it seems to have originated not with Peltzman himself but with Gordon Tullock and Armen Alchian, independently of each other.
It’s commonly known as “Tullock’s Dagger”, with many photo illustrations of same in Google Images, (for instance). 🙂
Feb 13 2023 at 6:05pm
Regardless of who said it, it’s a powerful metaphor.
ALthough I have found since COVID it’s lost some impact. A lot of people are thinking the Spike is a good idea now
Knut P. Heen
Feb 14 2023 at 11:23am
I think this issue was discussed in financial economics before Peltzman. It is closely related to both the Modigliani-Miller theorem and the CAPM. The idea is that the investor has a preference for a combination of risk and return, and that changes in risk-return made by the corporations will lead the investor to change his portfolio in the opposite direction.
Applying the theory to car speed and car accidents gives exactly the Peltzman conclusion. Safer cars lead to more speed because the regulation doesn’t change the driver’s risk-return preference. A neglected point is that the wealth effect may very well go in the same direction. A more expensive and safer car make you poorer. The driver may want to work longer hours to compensate, and drive faster to recoup some of the lost leisure time. It is very similar to accepting more pollution the poorer you are.
Feb 14 2023 at 5:50pm
Perhaps I have a disconnected neuron or two, but I can’t see how anyone could doubt that this effect is a real thing, even based on a simple explanation. Sounds like a failure to appreciate basic human nature.
Feb 14 2023 at 9:13pm
A semi-related side note. Look at this wikipedia page on vehicle safety, and see if you can tell only from the graph itself when the government began regulating vehicle safety. Looks to me like the government has merely jumped on a bandwagon, leading from the rear.
Feb 14 2023 at 10:21pm
The Peltzman Effect would seem to be a kind of moral hazard*
*Moral hazard: the risk that an individual or organization will behave recklessly or immorally when protected from the consequences
Comments are closed.