“If the government wanted people to drive safely, they’d mandate a spike in the middle of each steering wheel.”  — Gordon Tullock

We call economics “the dismal science.” Economist Gordon Tullock certainly reinforced that idea when he applied the economic phenomena of risk compensation to driving safety and suggested that we would all drive more safely if we knew that death was certain if we were in an accident.

That’s not how it is, though. Our efforts to improve safe behavior focus exclusively on making activities less dangerous, not more dangerous. Whether we’re prepared to admit it, our approach to safety is to achieve it despite behaviors, not because of behaviors.  We talk about behavior-based safety, but that’s not what we really do.

The Peltzman Effect

Tullock’s thought experiment is fascinating, but it is the work of economist Sam Peltzman that really grabs my attention. In the 1970’s, when seatbelts first became required in the United States, Peltzman studied the impact.  On the one hand, he expected that seatbelts would reduce the probability of a crash being fatal.  On the other hand, he expected that drivers, feeling safer in their seatbelts, would drive more recklessly, increasing the probability of a crash. His question was about which effect would be greater. His conclusion was that the two effects cancelled each other out.  The Peltzman Effect.

His epiphany was that bystanders—pedestrians, bicyclists, people pushing baby strollers—would receive no safety benefit from the seatbelts but would suffer as a result of increased recklessness. The data showed it was true.

A cottage industry has grown up in the field of economics, looking for examples of the Peltzman Effect in a range of activities.  Seatbelts continue to be a subject of study, with researchers looking to validate or repudiate Peltzman’s original conclusions.  Any more, though, the question is not whether risk compensation is a thing, but its extent.

There Ought to Be a Law

For those of us concerned about the safety of others, whether in the home, on the road, or in the workplace, our first instinct is to figure out what a safe behavior is and then make that behavior a rule. What parent has handed the keys over to a new teenage driver without making a very detailed list of rules for them to follow? In fact, the further removed we are from the situation, the more we want to make rules to substitute for being there.  This is just as true in the workplace as it is at home.

OSHA, created in 1970 under Richard Nixon when the workplace fatality rate in the United States was 18 fatalities per 100,000 workers per year, has promulgated hundreds of regulations.  Almost 50 years later, the fatality rate in the United States has fallen to 3½  fatalities per 100,000 full time equivalents (FTEs) per year. It would be hard to argue that OSHA regulations have not made a difference.

Many have said, “Safety rules are written in blood.” We don’t make safety rules about things that don’t matter.  We make safety rules because something horrible has happened and we want to prevent it from happening again.  First, though, the horrible thing has to happen.  Next, we look around to see why it doesn’t happen all the time. We find those practices and behaviors that seem most likely to reduce the likelihood of the horrible thing from happening again. Then someone decides, “There ought to be a law,” and a new rule or regulation is born.

Voluntary vs. Mandatory

Safety rules are not created out of thin air.  Safety rules are a codification of practices that have worked. Before there was the rule, there was the practice, and because there was no rule, the practice was voluntary. People who were committed to the practice before the rule followed the practice because they wanted to, because they saw the value in it. After a practice becomes a rule, anyone else following the practice does it because they have to, not because they want to.

That means that a practice performed voluntarily will always be more effective than the same practice performed because it is mandatory.  Consider PPE.  Let’s say that the injury rate for people who wear gloves and safety glasses with side shields, in addition to their safety shoes and hardhat, regardless of whether they are required, is 12. Additionally, let’s say those people are 25% of the work population. At the same time, let’s say that the injury rate for the other 75%, the people who never voluntarily wear PPE, is 120.  Overall, the injury rate is 93.

Now, imagine someone implements a safety rule that requires everyone to wear the additional PPE.  Will the overall injury rate drop to 12? No. The people who think the rule is unnecessary will do other things to compensate for the reduced risk.  The overall injury rate will drop, but not nearly as much as imagined.  Whatever the cost-benefit analysis was, it will have overstated the actual benefit. Still, there will be a benefit.

You Can’t Make Me

The safest people are the people that want to be safe. These are the very people for whom safety rules and regulations are not mandates, but instructions on how to be safe. Likewise, the safest workplaces will be those where everyone wants to be safe. The role of safety professionals in such a workplace is not to enforce safety rules, but to educate people on how to be safe.  The rest will take care of itself.

Peltzman’s work led him to be a proponent of deregulation.  He was convinced that mandatory safety rules induced as much harm as they prevented, with no net benefit, but at the cost of implementing the regulation. His argument suggests that people are simply as safe as they want to be and that they cannot be made to be any safer than that.

The evidence of 50 years of OSHA regulations belies this notion. With increased regulation, the workplace fatality rate has fallen. For those that want to argue that the fatality rate would have fallen anyway, because that is simply the direction society was moving, there is another statistic to consider.

OSHA only has jurisdiction over employers, over companies with employees.  OSHA has no jurisdiction over the self-employed.  While the overall fatality rate for workers is 3.5  fatalities per 100,000 FTEs per year, the fatality rate for employees—those who work for someone else, so, in accordance to OSHA’s mandates—is 2.9 fatalities per 100,000 FTEs per year.  On the other hand, the fatality rate for the self-employed, who are not subject to any of OSHA’s regulations, is 13.1 fatalities per 100,000 FTEs per year. While there are clearly ways to be safer, this is as safe as they want to be, which is not much safer than workers were in 1970.

As Safe as You Want to Be

Our best efforts to improve safety in the workplace is in recognizing safety concerns, knowing what to do about them, convincing our colleagues that they want to be safe, and showing them how.  However, we should not let fear of the Peltzman Effect prevent us from imposing safety mandates when voluntary efforts are not enough.  Although never as effective, mandates still make a difference.

Which brings us back to the Tullock spike. The government is never going to mandate a spike in the middle of each steering wheel.  Yet knowing that a spike coming out of our steering wheel will induce safer driving, there is nothing that prevents any of us from installing our own. Instead, parents continue to give their teenagers driving rules, highway departments continue to post speed limits, and workplaces continue to implement safe work practices.

When it can be done, it is always better to instill a desire to be safer than to install a safeguard.  When it cannot be done, when “safe as I want to be” is not safe enough, there is no other choice: implement and enforce safety rules. Although rules should be the last resort, we can make the workplace safer and we must.


This blog is based on an earlier version, “You Can’t Make Me: Mandatory Safety and the Peltzman Effect”, posted on 8-Mar-2017 by Elsevier in Chemicals & Materials Now!