Newsletter Sign-up

"Seat belts save lives" is a phrase often heard in TV advertisements supporting seat-belt laws. It is also often heard in safety meetings as support for a variety of measures meant to prevent accidents and injuries. It has been stated so often that it has become a truism.

That seat belts save lives is, however, completely false. Seat belts, by almost any measure, do not save lives. At best they do nothing. At worst, they actually cause more accidents and fatalities. That more people are not aware of this is a testament to the power of group-think, advertising and statistical illiteracy. As risk managers, we can't afford to succumb to any of these if we are to provide value.

What Adams discovered in his exploration is that there is significant empirical data supporting the theory of "risk compensation." This theory states that people have a certain level of risk that they find to be preferable. And, like a thermostat in a house, they will adjust constantly to retain that preferred level. So, taking the seat-belt example, if people feel that they are less at risk when wearing a seat belt, they will increase their risk levels in other areas to compensate. For example, they might follow the car in front of them more closely, take turns faster or drive at higher speeds. Over a large number of drivers, these will have the net effect of negating the safety benefits of wearing seat belts. In fact, there are studies that prove that drivers tend to drive faster if they are wearing seat belts.

But one does not have to perform lengthy studies to prove this point. It can be done with a simple experiment. Adams suggests that the reader close his eyes and imagine driving a car with poor brakes, no seat belts and a sharp iron spike protruding from the center of the steering column. How would one drive in that car compared to a modern car with seat belts, airbags and anti-lock brakes? Clearly one would drive much more cautiously in the iron-spike-equipped vehicle. If all vehicles were so equipped, there would probably be very few accidents, and the fatality rate, though high as a percentage, would be lower in the total number of deaths.

Risk compensation has much broader implications. It suggests that motorcycle helmets might not save lives (they don't), that child safety seats might cause more infant deaths (they probably do) and that safety devices at work might not actually decrease injuries (they probably don't). It is quite possible that, while wearing steel-toed boots and hardhats, workers might feel sufficiently protected and therefore decrease their level of vigilance to workplace hazards.

Risk compensation flies in the face of current safety theory. Safety devices cause accidents, and dangerous situations drive cautious behavior. For years we have been attempting to make the workplace safer. Perhaps it is time that we question our hackneyed assumptions. At the absolute least, these assumptions should be tested through empirical tests and statistical analysis. Iron spikes might be a better safety device than hard hats.

BEAUMONT VANCE manages risk for Sun Microsystems Inc.

April 15, 2007

Copyright 2007© LRP Publications