I went to a fight and a hockey game broke out

Another of my favourite sports economics papers – which I didn’t have time to mention in this week’s AFR oped – is this one, on hockey violence. An ambitious economics student could easily replicate it for rugby league, I would’ve thought.

Blood Money: Incentives for Violence in NHL Hockey
John P. Haisken-DeNew and Matthias Vorell
The level of violence in the National Hockey League (NHL) reached its highest point in 1987 and has reduced somewhat since then, although to levels much larger than before the first team expansions in 1967. Using publicly available information from several databases 1996-2007, the incentives for violence in North American ice hockey are analyzed. We examine the role of penalty minutes and more specifically, fighting, during the regular season in determining wages for professional hockey players and team-level success indicators. There are substantial returns paid not only to goal scoring skills but also to fighting ability, helping teams move higher in the playoffs and showing up as positive wage premia for otherwise observed low-skill wing players. These estimated per-fight premia, depending on fight success ($10,000 to $18,000), are even higher than those for an additional point made. By introducing a “fight fine” of twice the maximum potential gain ($36,000) and adding this amount to salaries paid for the team salary cap (fines would be 6.7% of the team salary cap or the average wage of 2 players), then all involved would have either little or no incentives to allow fighting to continue.

Careful readers will notice footnote 2.

We would especially like to thank David M. Singer from hockeyfights.com for his help in the preliminary analysis.

This entry was posted in Sport. Bookmark the permalink.

9 Responses to I went to a fight and a hockey game broke out

  1. Kevin Cox says:

    While these investigations are interesting and fun I would have thought they were more about social science and psychology than economics. It worries me when economists seem to spend an inordinate amount of time and effort on using standard statistical techniques to investigate social phenomena. The reason it worries me is that economists seem to put most of the emphasis on money (price) and we are left with the impression that fighting in rugby is all about money incentives whereas fighting in sport was around well before we had professional sports people and probably occurs more when there is no money involved.

    One way to counter this emphasis on what may well be a peripheral factor is for social scientists to do the investigations and to call on economists for their input in the same way that they now call on statisticians for advice. In other words I would submit that fighting is more a status and social phenomena than an economic phenomena and that papers by economists tend to over emphasise the price aspects of things they investigate – probably because it is easily measured not because it is intrinsically important.

    Why is this important? Well it is important because government policy seems to be dominated by statistical economists and the emphasis on money and particularly price leads to the nonsense we have just witnessed with the petrol price “debate” whereas the more important issues (global warming, distribution of profits from price increases) are suppressed.

  2. conrad says:

    I think one problem with promoting violence in sports is that use lose out on a larger market in the long term. Anecdotally, if I compare AFL across the decades, what I see is something which was marketed as a family game (and hence became very strict on violence). Now I believe around half the spectators are women, and you also see lots of kids at the games, which means lots of spectators. Alternatively, if I look at rugby, all I see is a bunch of blokes in the stands in a game that is going nowhere, despite the far greater appeal it could have had due to it being played in more than one country.

  3. Jennifer says:

    @Kevin: contrary to popular belief, economics very much IS a social science, and it is NOT about money. Economists are interested in incentives, and what you call ‘social phenomena’ ARE also ‘economic phenomena’ because we’re talking about people responding to incentives, whether it’s money or status. Of course, money is always a strong incentive but any good economist will also identify other things that could be acting as incentives in a given situation. However, one reason for focusing on money as an incentive in sports, other than the obvious fact that it is easy to observe and measure, is that it is also a variable that can actually be controlled. The take-away here shouldn’t be that athletes fight because there are monetary rewards (and by the way, those monetary rewards likely arise precisely because of the status you mention); it’s that we could reduce fighting by given athletes a financial incentive (or dis-incentive, in this case).

  4. Pingback: Josh Reviews Everything

  5. christine says:

    An investigation of the effects on hockey violence of the recent spate of assault cases launched for on-ice fights (McSorley, Bertuzzi) would also be helpful.

    Of course, the effects of legal penalties have long been studied by economists, and are, surprise surprise, not technically financial. Kevin’s got a bee in his bonnet. And I really don’t think he should worry about economists spending a lot of time on sports analysis. They don’t.

  6. Kevin Cox says:


    It is the emphasis that concerns me. No matter what we do we are connected to others and a climatologist is interested in the effect of climate on humans and on the way that humans are changing climate etc. Of course economists are concerned about incentives but too often the only incentive that seems to matter is money not value because as you say money is something economists can measure. I think economists are moving back to a broader view of incentives and choice but they have both been represented by others and act as though money is the most important – indeed the only factor. While I think Andrew tries I believe his work on merit pay for teachers illustrates what I am talking about.

    Our occupations and lives are not mainly about money and people work better as a group when for whatever reason they work together for the greater good as a social cooperative group. The difficulty for economists is that they quite rightly think that you get further when you measure things but it is when the measure becomes the substitute for the goal and worse becomes to tool to achieve the goal that we run into trouble – and that is my complaint.

  7. Kevin Cox says:


    I like what economists try to do but I want them to start to question the way they do things because I think it is leading down wrong paths. My chief complaint is that their methodologies are stuck back in the late 1800’s. Most economists try to model complex systems using Newtonian techniques based on stable physical systems. Human economic activity is dynamic, evolving, and like quantum mechanics the mere fact of measuring changes it.

    I would not mind except the views of economists are taken seriously and their models are being used to determine our fate. The modelling that is being done to determine our response to climate change is based on a steady state model of the economy that even the economists do not believe will be able to predict let alone help us control the effects of what is coming down the track.

    We can’t even control the economists main measuring tool of money and we let it change its meaning and even encourage it through setting a target for how much we will corrupt it. There is something wrong with a science when we encourage our main measuring tool to randomly vary:)

    I have a few more bees buzzing around but that will do for the moment.

  8. conrad says:


    you need learn about parsimony. Simple models that explain a lot are generally better than very complex models that explain only slightly more, at least in the social sciences. It’s also true of other areas incidentally — Newtonian physics works fine for many many problems, and we don’t need to worry about the speed of light etc. even though it would be more accurate if we did.

  9. Kevin Cox says:


    I agree with merits of parsimony and the “evolutionary” models of economics are much much simpler – just more calculations and less algebra.

    Of course Newtonian physics works well for non dynamic systems but economic systems are not so simple.

    I was appalled when I found out the modelling techniques being used to model the effect of emissions permits. They have little chance of giving an accurate picture. Why not try modelling techniques similar to the people who try to forecast weather or simulate atomic bombs?

    Even better why not build into existing economic systems ways to directly influence multiple goals and then you can test your models as you introduce new functions like emissions trading. The problem with indirect methods such as emissions trading is that it will take years before you know if it is going to work. It is the same with most things economic. If at all possible you want your system to evolve and you to direct the evolution. That is what evolutionary modelling promises.

Comments are closed.