DRI-219 for week of 11-23-14: A Columnist’s Dawning Recognition of Deadly Auto-Safety Regulation

An Access Advertising EconBrief:

A Columnist’s Dawning Recognition of Deadly Auto-Safety Regulation

We are familiar with investigative reports by reporters in print and broadcast media and, in recent years, online. We view these as the mechanism for regulating institutions not subject to the constraints of the marketplace. Government is chief among these.

This routine has accustomed us to casting the news media in the role of cynical watchdog, always looking for wrongdoing and too prone to suspect the motives of those it covers. Of course, we may suspect the press of pre-existing bias – in favor of Democrats, for instance. But for the most part, we believe that their interests are served by finding scandal, wrongdoing and malfeasance, because these things are news.

The possibility that the press itself may be naïve and complacent is the last one we consider. It should not be overlooked.

Air Bag Safety 

Wall Street Journal columnist Holman Jenkins has written a series of columns about auto safety and regulation. Many of them followed the regulatory travails of Toyota, which endured a prolonged crucifixion when its vehicles were ostensibly subject to a problem of “unintended acceleration.” Although it was all too clear that the problem was caused by drivers unwittingly depressing the accelerator instead of the brake pedal, the company was beset by the fable that a bug in the car’s computer code was causing cars to accelerate when they should be slowing. Despite the conspicuous lack of scientific evidence for this hypothesis – not surprising in view of its impossibility – Toyota eventually was forced to pay out hundreds of millions of dollars in settlement money to make the issue go away.

Having set a tone of skepticism toward regulators, Jenkins turned next to the recent disclosure by Takata that their air bags have displayed defects. Toyota and Honda have recalled over 8 million vehicles to replace the air bags. The defect (apparently caused by moisture entering the ammonium nitrate air filter of the air bag) breaks down the explosive tablets in the air bags, causing them to burn quicker and explode more violently than normal. In turn, this shreds the metal housing surrounding the tablets and sends a shower of shrapnel into the driver and front-seat passenger (if any).

Jenkins noted that the demand by federal automobile regulators that the companies recall millions more vehicles is suspiciously timed to coincide with the end of hearings on the response by Japanese automakers to the finding of defective air bags. He reserved his strongest note of skepticism, though, for the use of air bags as safety devices.

“The faulty Takata air bags are connected to five deaths in 13 years, which is a tiny fraction of the deaths known to be caused by air bags working as designed [emphasis added]. When the Takata mess is cleaned up, we’ll still be left with a highly problematic safety technology.”

What’s this? Air bags themselves cause automobile-occupant deaths? They’re supposed to prevent deaths, not cause them. This is surely news to the general public, which is why Jenkins continues with a brief chronology of air bags’ journey from industrial infancy to ubiquity. “Washington began pushing automakers to install air bags in the 1980s, and ever since Washington has been responsible for research that confirms that air bags save hundreds of lives a year. These studies, though, credit air bags with saving people who were also wearing seat belts, when considerable evidence indicates seat belts alone do the job.”

“These studies also assume that deaths in collisions where air bags deployed are always attributable to the collision, never [to] the air bag.”

Jenkins does not mention that the push for air bags coincides with a federal push for mandatory wearing of seat belts. The first state law that required the wearing of a seat belt meeting federal specifications – essentially, a three-point seat belt buckling over the lap but also including a shoulder restraint – was passed in 1983. States received seven-figure federal bounties for passing a mandatory seat-belt law and achieving an estimated compliance beyond a specified rate.

“A 2005 study by Mary C. Mayer and Tremika Finney published by the American Statistical Association tried to correct for these errors and found that the clearest effect of air bags was an increased risk of death for unbelted occupants in low-speed crashes. Likewise, a 2002 study of 51,000 fatal accidents by University of Washington epidemiologists found that air bags (unlike seat belts) contributed little to crash survivability.”

“[Thus] air bags began as simple bombs buried in the dashboard designed to protect the typical non-seat belt wearing accident victim – the typical unbelted victim being a 170-pound teenage male. In 1997 came the reckoning: Air bags designed to meet the government’s criteria were shown to be responsible for the deaths of dozens of children and small adults in otherwise survivable accidents.”

This is the key point in Jenkins’ chronology, the point at which the reader’s eyebrows shoot up and he shakes his head in disbelief. Dozens of deaths? Adults and children? How did I miss the public furor over this? After all, when one or two people die owing to an automobile defect that a company knew about or should have known about, all hell breaks loose. In fact, the history of government suppression of unfavorable air-bag performance goes back decades. But Jenkins makes no mention of this; instead, he moves on.

“Since then, air bags have become ‘smarter,’ with computers modulating their deployment depending on type of crash, passenger characteristics and whether seat belts are being worn. Undoubtedly the technology has improved but still debatable is whether the benefits outweigh the risks and costs. Air bags remain one of the biggest reasons for vehicle recalls – and no wonder, given that these devices, which are dangerous to those who manufacture them and to those who repair vehicles, are expected to go years without maintenance or testing and then work perfectly.”

“Because, in the minds of the public, not to mention in the slow-motion videos on the evening news, air bags are seen as gentle, billowy clouds of perfect safety, yet another problem is the potential encouragement they give motorists to drive more aggressively or forgo the hassle of buckling up.” Now Jenkins has driven himself and his readers into water over their heads and is stalled. His column will drown unless it is rescued promptly. He has made a strong case that air bags are inherently unsafe, but is now suggesting something else – a different source of harm from their use. The fatal stretch of water was entered with the words “…yet another problem is the potential encouragement they [air bags] give motorists to drive more aggressively or forgo the hassle of buckling up.” This requires the services of a professional economist.

The Economic Principle of “Risk Compensation”


People tend to increase their indulgence in risky activities when they perceive that the safety of those activities has been enhanced. Risk should be treated just like any other consumption good – when the price of risk goes down, we should expect people to purchase more of it.

The first of the previous two sentences would meet with the approval of most people. The second would not. Yet from the economist’s perspective they might be interpreted as saying the same thing. A space shuttle is currently being readied for commercial use by tourists; wouldn’t we expect tourists to be more enthusiastic about it when improvements in launch and flight safety reduce the risk of death and serious injury for passengers? Still, we would expect there to be an appreciable risk of space travel for the indefinite future, wouldn’t we? When improvements in contraception result in better prophylactics, don’t we expect people to have sex more often, despite the fact that they still run a risk of contracting a sexually transmitted disease?

In 1975, economist Sam Peltzman published a seminal article in the Journal of Political Economy. He analyzed the effects of a series of government-mandated safety devices introduced beginning in the mid-1960s. His analysis suggested that the net effect on safety was approximately zero. Peltzman offered two explanations for this surprising result, the most plausible of which was that requiring the use of seat belts by drivers causes some people to take more driving risk than they would have if they had been driving beltless. This additional driving risk produced more accidents. While the increased use of safety devices tended to produce fewer injuries and fatalities among automobile occupants, the increased number of accidents also implicated non-automobile occupants such as pedestrians, motorcyclists and bicyclists. These additional injuries and fatalities tended to offset the injuries and fatalities saved by the use of seat belts, so that the comparative end result in driving statistics such as “fatalities per million miles driven” was a wash.

Over the succeeding forty years, this kind of outcome became proverbial throughout the social sciences, not just economics. In 2006, Smithsonian Magazine published an article summarizing the powerful effect that Peltzman’s work has had on the world. His ideas are grouped under the heading of “risk compensation,” an evocative term that implies that we satisfy our appetite for risk by compensating for added safety by “purchasing” more risk.  The principle has been observed in nations around the world, among adults and children, in activities ranging from driving to playground behavior to sports. Famous economist N. Gregory Mankiw, former Chairman of the President’s Council of Economic Advisors under President George W. Bush, blogged about “Sam Peltzman, who taught us all that mandatory seat-belt laws cause drivers to drive more recklessly.” Mankiw dubbed the relevant principle the “Peltzman Effect,” making Peltzman one of a select group to have a scientific principle named after him.

Despite the scientific status of risk compensation and the Peltzman Effect, Holman Jenkins shows no sign of having heard of it. He speaks of the “potential encouragement” offered by air bags to more aggressive driving by motorists as if he had just exhumed a Stone Age cave and stumbled upon a rectangular version of the wheel therein. And he applies the principle to air bags with no apparent awareness of its equal applicability to seat belts.

The Perils of Mandatory Safety


“By now,” Jenkins laments, “those of a certain age remember that Detroit was the villain that opposed putting explosive devices in their vehicles, plumping instead for mandatory seat-belt laws (which, amazingly, certain safety groups opposed).” No economist is surprised that Detroit opposed the idea of being forced to increase the cost of production by providing a safety benefit that (a) consumers didn’t want and (b) didn’t work, which would (c) expose them to endless litigation as well as threaten them personally. Seat belts were several orders of magnitude less expensive and the loss of freedom to consumers did not represent a business loss to automobile companies.

Jenkins’ failure to understand the opposition to mandatory seat-belt laws is astonishing, though, since it is based on the very same principle that he just invoked to oppose air bags. There is a lot to be said for seat belts when provided as a voluntary option for consumers. There is everything wrong with mandatory seat-belt laws because they encourage (force?) risk-loving drivers to obey the law by buckling up, then to fulfill their love of risk by driving more aggressively – and to do this as a substitute for going unbuckled in the first place. An unbuckled risk-lover is a driver who is himself bearing the risk he chooses to run. A buckled-up aggressive driver is a risk lover who is imposing the risk he chooses to run on other drivers – and pedestrians and cyclists – who may be more risk averse. This is bad theoretically because it is economically inefficient. Economic inefficiency is bad in the practical sense because it misaligns cost and benefit. In this case, it allows risk lovers to benefit from the risks they run but imposes some of the costs on other people who don’t benefit because they are risk-averse individuals who didn’t want to run those risks in the first place.

The practical side of all this has been seen is various ways. New Hampshire is the only state that hasn’t passed a mandatory seat-belt law in the interval since the mid-1980s. It has poorer-than-average weather and topography, so we would expect to it to have worse-than-average traffic-fatality results, all other things equal. Since traffic-safety expert predicted that mandatory seat-belt laws would usher in traffic-safety nirvana by reducing fatalities hugely, we would expect to find that New Hampshire highways had become a veritable slaughterhouse – if mandatory seat belts were the predicted panacea, that is.

Instead, New Hampshire traffic statistics have improved to near the top of the national rankings despite its singular lack of a mandatory seat-belt law. New Hampshire should be the poster-state for mandatory seat-belt laws; instead, it is the smoking gun that points to their guilt. This fact has gone completely unremarked in the national news media, which is probably why Holman Jenkins hasn’t noticed it.

But there is no excuse for Jenkins’ failure to notice the slowing improvement in nationwide traffic statistics that occurred along with the installation of air bags and mandatory seat-belt laws. The rate of fatalities per million vehicle miles driven has been falling since the 1930s and the growth of modern automobiles, highways, safety methods, signage and improved quality control in production and repair. The federal highway safety bureaucracy makes a point of announcing the yearly fatality data because it usually represents a recent low point. What they fail to announce is the slowing rate of decline. Indeed, recently fatalities have actually risen in spite of the poor economy and less auto travel.

Jenkins apparently considers himself daring for suggesting that air bags are counterproductive and should be eliminated. He cites the myriad of safety innovations that have come on line in the last few years: automatic lane-violation warning devices, automatic skid-correction devices, automatic collision avoidance and braking sensors, automatic stabilizers and design features that direct crash energy away from passengers. “One imponderable is how much faster progress might have been without the bureaucracy’s forced diversion of industry capital to air bags … Each stride tilts the calculation away from having an IED in the dashboard as a net benefit to motorists, bringing closer the day when a new safety innovation will be announced: an air bag-free vehicle.”

Actually, Jenkins’ repudiation of air bags and reaffirmation of mandatory seat belts puts him about 45 years behind the times – about where we stood before Sam Peltzman wrote in 1975. Jenkins deserves credit for daring to break out of the regulatory mindset by opposing air bags, something other journalists have failed to do. That indicates the intellectual depths to which the downward market spiral of journalism has taken us.

What Jenkins should be doing is calling for abolition of the Department of Transportation, not air bags. Consider this: According to Jenkins’ own logic, the DOT mounted a nationwide campaign for mandatory seat-belt laws while also insisting upon mandatory air-bag installation in vehicles. But this is crazy. Using the language of game theory, we would say that the presence of air bags “dominates” seat-belt use, making it superfluous. With an air bag, one of two things happens: the air bag deploys as intended – in which case the passenger is protected in the accident – or the air bag explodes – in which case the passenger is maimed or killed. Either way the seat belt is superfluous. Wearing a seat belt doesn’t add protection if the air bag works and doesn’t protect against shrapnel if the air bag explodes. There is also a third possibility: the air bag might explode prematurely, killing or maiming the passenger even though there is no accident. And the seat belt is superfluous in this case as well.

Although shouldn’t be too hard on Holman Jenkins, we shouldn’t feel bound by his intellectual limitations or his inhibitions. Now that we know that both mandatory air bags and mandatory seat-belt laws are abominations enacted in the name of automobile safety, what are we to make of a federal-government safety bureaucracy that insists upon them even after their demonstrated failures? And with the technology of self-driving cars a demonstrated reality, what are we to think when that same bureaucracy is distinctly reluctant to allow it to proceed?

Why Does DOT Tend to Hinder Rather Than Promote Automobile Safety?

The heading for this section will anger many readers. The conventional view of federal regulation has been described by the late Nobel laureate James Buchanan as the “romantic” view of government. Roughly speaking, it is that government regulators act nobly and altruistically in the public interest. Upon very close examination, the term “public interest” will be found so vague as to defy precise definition. However, this is advantageous in practice, as it allows each user to define it according to his or her individual desires – it makes the theory of government into a sort of fairy-tale, wish-fulfillment affair. No wonder this approach has survived so long with so little clear-eyed scrutiny! Everybody is afraid to look at it too hard for fear that their fondest dreams will go up in smoke. And indeed, that is what actually happens when we try to put this theory into practice.

Suppose we depart from this sentimental approach by inquiring into the incentives that confront bureaucrats in the Department of Transportation (DOT). First, ask what happens if DOT develops an innovative safety technology that saves the lives of consumers. Let’s say, for example, that they develop an improved seat belt, such as the three-point seat belt which turned the failed two-point lap belt into a viable safety device. Will the individual researcher(s) in DOT get a bonus? Will he or she (or they) patent the device and earn substantial royalties? Will they become famous? The answers are no, no and no, respectively. Thus, there are no positive incentives motivating DOT to improve automobile safety.

On the other hand, suppose DOT does just the opposite. Suppose it actually worsens auto safety. Indeed, suppose DOT does exactly what the political Left routinely accuses capitalist businessmen of doing; namely, kills its “customers” (in DOT’s case, this would be the consumers who are the ostensible beneficiaries of regulation).

What an irritating, outrageous question to pose! We all know that government regulatory agencies exist to protect the public, so it is unforgivably irresponsible to suggest that they would actually kill the people they are supposed to protect. But – let’s face it – that is exactly what Holman Jenkins is implying, isn’t it? He never has the cojones to blurt it out, but the statements that “Washington has been responsible for research” and “air bags designed to meet the government’s criteria were shown to be responsible for the deaths of dozens of children and small adults in otherwise survivable accidents” don’t leave much to the imagination, do they? As it happens, there is plenty more dirty linen in the government’s closet that Jenkins leaves unaired.

As long as everybody else is as deferential (or as cowardly) as Jenkins, the general public will not link government with the deaths in the way that private businesses are linked with the deaths of consumers. When more consumers die, what happens is this: government benefits. DOT uses this as the excuse to hire more people, beef up research and spend more money. Larger staffs and bigger budgets are the bureaucratic equivalent of higher profits, but this differs from the private-sector outcome in that higher profits are normally associated with better outcomes for consumers while, if anything, the reverse is true of bureaucratic expansion in government.

Suppose DOT were to recommend that we proceed at breakneck speed to adopt driverless cars in order to eliminate virtually all of our current 30,000+ annual highway fatalities. Suppose the agency even brings about this outcome within just a few years. There would be little or nothing left for the agency to do; it would have succeeded so well that it would have innovated itself out of existence. No wonder that DOT is dragging its feet to slow the acceptance of driverless cars!

In the private sector, there is an incentive to solve problems. In government, there is never an incentive to solve problems because that will usually leave government with no excuse to exist, to grow and expand. When a private firm solves a problem, it makes a big pile of money that it can use to expand or enter some new line of business – even if the solution to the problem leaves it with no reason to continue producing its current product. There is no government analogue to this reward and consequently no incentive for government to succeed, only incentives for it to fail. Indeed, there are even incentives for it to do harm. And in the arena of automobile safety, that is exactly what it has done.

Just to reinforce the point, let’s generalize it by broadening our evidentiary base beyond federal regulation. Earlier we cited various numerous safety improvements that are being incorporated piecemeal into automobiles by the major auto companies: lane-violation detection, automatic braking, collision avoidance and others. Driverless cars include all of these and more besides. The state of California has recently passed regulatory legislation forcing all driverless cars to allow a human driver to “take over in an emergency;” e.g., bypass the sensors that govern the driverless car’s actions. But every one of the safety improvements listed was designed expressly to produce mistake-proof behavior by the car in various emergency situations. In other words, the regulatory legislation has the effect of defeating the safety purpose of the driverless car. Oh, some nobler, more romantic rationale is advanced, but that is the effective result of the law.

The case of the DOT is not unique at the federal level either. Hundreds of thousands of corpses could attest to the harm the FDA has done by blocking the approval of new drugs. Many economists could, and have, detail the harm done to competition by application of the antitrust laws ostensibly designed to preserve and protect it.

Economists are the real investigative reporters. Most of the time, their tools consist of logic and arithmetic rather than confidential informants and leaked documents. But when it comes to exposes, their stories put those of journalists in the shade.