Hockey's CBA, salaries, and competitive balance

Here’s a nice paper on the NHL collective bargaining agreement.

It’s titled "Sharing the Wealth in the 'New NHL'", by Taylor Brinkman, and it taught me a bunch of stuff I didn’t know about the CBA.

Brinkman starts out by explaining the revenue sharing process, by which small-market teams with low revenues get monetary transfers from the other teams. The algorithm involves a bunch of different slush funds from different sources (television revenues, playoff tickets, etc.) that get allocated in certain ways. One thing that surprised me is that playoff teams are taxed from 30 to 50 percent of the total value of tickets for playoff games in their arena. The bottom teams get some of that back in transfers, of course, but it still seems to me that taxing teams for succeeding in the playoffs is the wrong kind of incentive.

There’s also a bit of a disincentive to improve your team in the off-season. Because teams must pay players at least 54 percent of revenues, there’s effectively a 54% tax on hiring better players. To quote Brinkman’s example, suppose the Carolina Hurricanes figure that signing Erik Cole will gain them $2 million in additional revenues. Since they will eventually have to pay an extra $1.08 million in salaries because of that additional revenue, that means their gain will be only $920,000, and they can’t offer Cole any more than that. Presumably this doesn’t bother the Hurricanes too much, because the overall effect is to depress salaries in general, which can only help their bottom line.

As for competitive balance, Brinkman concludes that the salary cap won’t necessarily help. Teams receiving transfer payments aren’t required to spend that extra money on salaries – all they need to do is sell 13,000 tickets per game and grow revenues faster than the league average. One way to do that is to improve the caliber of play by signing players, but if a team doesn’t find that necessary to meet the conditions, it can simply pocket the cash.

The conclusion, therefore, is that if competitive balance does increase, it will be due to the salary cap acting on the richest teams, rather than because of transfer payments to the rest. However, there’s always the possibility of an owner who doesn’t care as much about losing money. “It is impossible to predict the behavior of owners in the real world,” Brinkman writes. “[Some] surely prefer to maximize wins [instead of profits].”

There’s a bit of sabermetrics in the paper, and, although the numbers only indirectly impact the argument, they’re useful results.

First, there’s a regression on salary versus standings points for the years 1994-2003. It turns out that the correlation is .388, which is almost exactly what “The Wages of Wins” reported for baseball. Like the “Wages” authors, Brinkman concentrates on the r-squared of .150, pronounces it small, and argues that “a team’s ability to compete on the ice had relatively little to do with how liberal they were with the pocketbook.”

But as I repeatedly and tiringly argue (example), the .388 means that 38.8% of a team’s salary spending goes directly to the win column -- and that's actually quite significant.

Second, there’s a chart of NHL competitive balance from 1994-95 to 2003-04. The standard deviation of winning percentage seems fairly constant in the 0.100 range, and there doesn’t seem to be much of a trend. Here are the numbers so you can judge for yourself:

.121, .111, .074, .097, .092, .108, .111, .095, .097, .099

If all teams were exactly equal in ability, Brinkman says, the SD would be .060.

If you’re interested in hockey at all, this a great paper – it’s very well written and not too full of economic jargon. You can tell by the way Brinkman explains his reasoning that he knows hockey and understands the issues. His explanations make sense, and the conclusions seem credible to me.