Hockey visors and moral hazard

In its Feb. 1 issue, The Hockey News makes the case for mandatory visors in the NHL. Currently, visors are optional, but the magazine argues that making them mandatory would cut down on eye injuries.

"The players would benefit from improved safety in the workplace. Mandatory visors wouldn't guarantee the elimination of catastrophic eye injuries, but the evidence suggests it'd be pretty darn close," writes Editor-in-Chief Jason Kay.

Writer John Grigg brings up the very scary eye injury suffered by then-Leafs defenseman Bryan Berard in 2000. In analyzing the current trend of more players wearing visors, Grigg argues: "[I]f current trends continue, the NHL will have fewer and fewer reminders of [Berard's injury]."

Reading the magazine, I wondered if anyone had ever considered that visors might create a moral hazard problem. Moral hazard occurs when people engage in more risky behaviour when they are partially protected from a risk. The textbook example is bike helmets — cyclists take more risks on the road and motorists drive more recklessly near cyclists when the cyclist is wearing a helmet. The logic is that you feel safer wearing a helmet, so you take more risks to compensate.

We can imagine that visors might have a moral hazard effect for hockey players. If you wear a visor, you feel more safe, and might make more plays that put your body at risk. Or other players might wave their stick around more recklessly in your presence since you're protected by a visor.

The important question is whether or not the moral hazard effect is big enough to outweigh the safety benefits. If visors make players play way more recklessly, this could cause more injuries than the visors prevent. Going back to the bike helmet example, many studies have shown bike helmets do decrease accidents, but some studies (notably a study of a 1992 Australian helmet law) found bike helmets don't reduce deaths, and in some cases increase the chance of death.

Since The Hockey News' Jason Kay said evidence suggests visors would all but eliminate eye injuries, I decided to investigate. Unfortunately, much of the research is outdated, and very little is from the NHL. All of it seems to overlook the moral hazard problem.

Although most studies find players who don't wear a visor experience more frequent and severe injuries than visor wearers, it's not necessarily the case. One study looking at data from five ECHL seasons found that facial injuries tended to be more severe for visor wearers than for non-visor wearers, since they were generally caused by high sticks that went under the helmet.

The most applicable paper to today's NHL visor debate, however, was a study by four neuropsychologists looking at data from the 2001-02 NHL season. The study concluded that wearing a visor "significantly reduces the number of eye injuries in the NHL." But their finding is questionable for a couple reasons.

First, the researchers found that no players who wore a visor that year suffered an eye injury. But only 181 players wore a visor that year. Compare that with the 606 players who didn't wear visors: only 1.7% suffered an eye injury. Since the frequency of eye injuries is low in both cases, it's possible it is simply a fluke that no visor wearers suffered an eye injury that year. To put it another way, it would only have taken three eye injuries for the visor wearers to suffer the same percentage of eye injuries as a non-visor wearer.

The second issue is that visor wearers may behave differently than non-visor wearers. It's quite possible that the players who choose to wear a visor are more cautious. The difference in injuries could be because of this cautiousness rather than because of the visor. The real question is whether forcing a visor onto a player who currently doesn't wear one would decrease that player's chance of injury, and the study provides no evidence on that front.

In fact, the authors revisit their analysis in a 2008 article. They find that players who wear visors take fewer penalties than players who don't, supporting my theory that players who wear visors may be more cautious. Curiously, however, the authors shrug this off: "… it is certainly possible that less aggressive hockey players choose to wear visors and that this is the cause of the reduced number of aggressive penalties. Future research should focus on this possibility. However, given the reduction in eye injuries at the NHL level, it is surely prudent to suggest that the visor be seriously considered as protection at all levels."

The moral hazard effect of wearing a visor in the NHL may be tiny, and the safety benefits may be large. The Hockey News may be bang-on that mandating visors would make eye injuries history. But it would be nice if the moral hazard problem is examined more closely first.