Only heartless trolls worry about costs

Wednesday, October 30th, 2019

Michael Munger looks back at the joyful contrarianism of Gordon Tullock, starting with his insights into safety regulations:

Should governments mandate more safety in products? The usual terms of debate weigh reduced injuries — the “human toll” — against increased cost, with only heartless “rational choice” trolls actually worrying much about costs. The idea that perfect safety is morally undesirable, because such policies have enormous opportunity costs, is obviously, annoyingly important — and a big part of the reason economists often end up standing alone at parties, studying the wallpaper pattern.

Safety is valuable, of course. But economists pitch their arguments “at the margin,” meaning for the last increment. The first improvements in safety are cheap and uncontroversial: reliable brakes, turn signals, seat belts, safety glass in windshields. The next increment — airbags, anti-lock braking systems — comes at much greater cost and with a smaller associated reduction in injuries. Ultimately, the only way to make cars completely safe is to park them and throw away the keys. Driving is dangerous.

Tullock’s contribution was to ask, “And then what?” The problem is worse, actually much worse, than the increasing marginal cost of safety improvements. The safety of the car, after all, is just one factor; drivers and their attitudes toward danger are the key missing variable. The state can only mandate the safety of the car. Ultimately, the driver’s behavior determines the risk of driving.

This observation is now sometimes called the “Peltzman Effect,” after the University of Chicago economist Sam Peltzman, but Tullock had argued some versions of it for decades. As the University of California, Irvine economist Richard McKenzie recalls it, Tullock noticed that safer cars reduced the costs of accidents for drivers. If the government “subsidizes” accidents by mandating airbags, there will be more accidents. Worse, because of increased automobile speed and recklessness, there will be more pedestrian injuries and deaths. Safer cars mean more injuries.

Tullock’s famous counterproposal was to place a long, sharp dagger firmly in the center of the steering column. His earliest notion of this was to have the tip pointing back and locked one inch from the driver’s chest. By the time I talked to him about it, in the 1990s, the idea had evolved to work more like an airbag, so that the dagger would be hidden but would deploy with explosive force in the event of an accident.

Calling this conclusion counterintuitive is an understatement — but there is an important insight underlying Tullock’s drollery. The risk of injury is jointly determined by the behavior of all the people who are driving in a particular area. If I’m aggressive and cause a wreck, I’ve imposed additional risk on you. If safety “improvements” subsidize risk-taking, and some — not all, necessarily, just some — people drive more aggressively, then the observed reductions in injuries from safer cars will be much less than regulators expect. Worse, no individual driver can, by behaving safely, escape these bad effects. Safety regulations have negative externalities.

The most annoying thing about Tullock was that he was usually right. He was even right about car safety regulation (though maybe not about the dagger!): According to the American Automobile Association, there have been substantial increases in driver aggressiveness since 2000, with eight out of 10 drivers admitting to having intentionally tailgated another car and nearly half saying they have bumped, rammed, or gotten out of their cars to threaten the occupants of another vehicle. While the direct causal mechanism is complex, this increase in aggressiveness tracks the imposition of universal requirements for airbags and anti-lock brake systems in 1998.

This problem is borne out in the real “national sport” of America, NASCAR. Starting in 1988, the racing entity imposed “restrictor plates” as a safety measure, limiting the airflow into an engine (and therefore the horsepower, and speed, of cars). Restrictors were required in response to the horrific “going airborne” May 1987 accident of Bobby Allison at Talladega, where the car flew into the upper restraining fence and disintegrated, injuring five spectators, including one who lost his eye. But two refereed journal articles, one in the Southern Economic Journal in 2004 by J.B. O’Roark and W.C. Wood, and one in 2010 in Public Choice by A.T. Pope and R.D. Tollison, concluded that safety improvements had increased the number of crashes and multi-car pileups in the sport (though they had not affected the total number of deaths).

That is what you would expect. If speeds are suppressed and safety equipment is improved, the risks of death and serious injury are lowered. The result should be increases in risky behavior by drivers, including close drafting and “trading paint,” the euphemism NASCAR uses for high-speed bumping.

In February 2018, NASCAR switched from restrictor plates to the more precise and consistent “tapered spacers,” which have the same effect and the same “safety” rationale. The 2018 NASCAR “Cup Series” champion, Joey Logano, was clear about the likely outcome: “I totally expect to crash more cars,” the Associated Press quoted him saying. “As cars are closer and drivers are more aggressive, a mistake will create a bigger crash. We can’t get away from it.”

Of course, Tullock would ask, “And then what?” NASCAR is not stupid; it may not be an accident (sorry) that there are more crashes with the same level of driver safety. That may very well be the point: NASCAR fans come for the racin’, but they stay for the wreckin’. Using a “safety” rationale — particularly one that really does reduce injuries slightly — as a means of increasing the number of wrecks makes a lot of economic sense. If the authorities really wanted to prevent accidents, NASCAR would put big daggers in steering columns, not little tapered spacers in carburetors.


  1. Graham says:

    I’m torn, and some of it’s generational.

    Not unlike disease-tolerance. I’m young enough to have been vaccinated for measles, among other things, so I think it was nuts that people took something so potentially fatal relatively in stride, as a quasi-normal childhood experience, just a few years earlier. Whereas my generation took chicken pox that way, but younger people got a jab against that. I never thought shingles a good idea, but the pox itself was a routine thing most of us had as kids. Only latterly did I learn it could be dangerous in some cases and that some folks now regard my attitude as nuts. I mean, I support the use of a vaccine that has become available, but still find it hard to take chicken pox all that seriously. Go figure. Maybe I resent that I might get shingles. But there’s now a vaccine for that we can get at 50, so I’ll white knuckle it til then.

    With cars, I was a kid when belts became mandatory in all seats in Ontario so I am used to them. My dad’s earliest cars may not have had them for every position. Oddly, I ride in cabs a lot and quite often neglect the belt even at highway speeds. Never forget when driving or passenger in a private car. I also find shoulder belts in rear seats poorly designed and the strap wants to cross my throat. One day in any minor accident I am going to be throttled by the seatbelt. So I guess I have mixed habits and view of them but they are certainly normal features.

    Naturally, I marvel at the high ages at which kids now are expected to have specialized seats. No doubt of the science, or the improvements in them over time, just marveling. I think it’s up to 10 now. [I never had any.] I suppose I marvel that we never had them in the 70s, in an age when accidents were more common and the cars less structurally safe, and now kids are strapped in at 10 when accidents seem less likely and the cars safer overall. Maybe it’s the adult size belts that imperil the kids. That I believe.

    One innovation tried in the Netherlands a while back demonstrated the same point made by Tullock- they removed all intersection controls in some town with the reported result that drivers were extremely careful.

    I quite understand that, but I am not convinced that approach would ever have scaled, even with the most competent drivers.

    Ontarian drivers are morons, for the most part. I’m no saint, but some basic stuff that’s been in the manual 30+ years is getting forgotten out there. We used to have ordinary two-lane residential streets divided into “through” and “cross” streets, with the former having right of way. My father insists it worked fine. Then in my time we increased the number of 4-way stops, presumably to make streets equal…, which is not that difficult a concept but I could see some transitional difficulty. It has in fact been a decades long problem for some. Now we are trying on roundabouts, even at some rather odd locations where they are more decorative and virtue-signaling [they're "European"] than useful. One, at which the pedestrian crossing points are poorly situated, has become my daily heart stressor despite being not that heavily trafficked. The same location, long ago before a couple of rounds of rebuilding and relandscaping, had involved just sprinting illegally across two sets of two-lane road and a grass median, and I thought it less stressful.

    I suppose my only point is, when presented with three or more roughly equivalent solutions, keep the one already in place rather than experimenting for the sake of. That next tiny increment of safety or efficiency, still anyway short of perfection, is not worth my time and effort to habituate, and not worth my additional stress. Life where I am was sufficiently safe in all aspects, far above global or historical or even US average, in all ways when I was 25.

  2. Kirk says:

    Humans are perverse creatures that cannot assess risk at all well. We focus on the spectacular and unlikely risk, while ignoring the prosaic and likely. Very few people get killed in mass shootings, while lots and lots get killed in random retail violence that takes place every day in dysfunctional cities. Yet, which gets more headlines?

    It isn’t just the media, either–If you were to actively try to draw attention to the low-order endemic violence in a city like Chicago, that would just lead to apathy as it blended into the environmental background noise.

    I remember being shocked to hear about a convenience store armed robbery “down in the city” while I was living in Illinois. Someone went it, took two employees and three customers into the back of a White Hen, and shot them to death in the walk-in cooler. Now, if something like that happened in the Pacific Northwest, even in a big city like Seattle, Portland, or Spokane, that’d be a nine-day wonder, front-page news for a couple of weeks.

    It was on page 6 of the Chicago Tribune. Only reason I knew about it was that one of the customers that was killed was the aunt of an acquaintance.

    Now, you look at that, and you go “OK, so… What’s your point…?”, and I’m going to tell you to look at the difference that that endemic background violence makes to people’s ability to grasp the reality of things. You ask the average Chicagoan what the chances are that they will be a victim in something like that, and they’re just gonna shrug resignedly and say “Who knows? That’s just life…”, and then they don’t do a damn thing about it. The same politicians who negotiate with the gangs keep getting elected, the same prosecutors and judges that don’t do anything effective about the crime rate still have jobs, and the locals are just resigned to it all as being a “part of life in the big city”. You had that same sh*t happen out where I grew up, and you’d see people losing their minds over it, and a bunch of cops, prosecutors, judges, and all the rest would be looking for new jobs.

    Which all goes to show that people make drastically different assessments of risk and safety. What Chicagoans routinely accept would be unacceptable anywhere else in the country, and cause for politicians and law enforcement administrators to fear for their jobs. Yet, the risk of falling prey to the criminal element is a hell of a lot higher in Chicago than elsewhere, and the local people don’t do sh*t about it…

  3. Eli says:

    I have seen similar discussions and conclusions regarding personal protective gear for athletes

  4. Paul from Canada says:

    It is not just in safety related things that you get these paradoxes. For example, initially, the use of computers and fax machines and photocopiers increased the use of paper, rather than decreased it. Before, if you wanted copies, you had to get the typist to make carbon copies, and there was a limit to how many could be made. So you limited the number, and used things like circ slips. Once you can make as many copies as you like easily, you just make one for everyone.

    Likewise, solar powered lights are mostly used for frivolous mood and accent lighting in gardens, rather than making any real difference in domestic electricity usage.

Leave a Reply