High Tax vs Low Tax - How the Taxman decides who wins the Super Bowl

I was watching the Super Bowl on Sunday night (or rather Monday morning). While I watched the New England Patriots defeat the Atlanta Falcons (btw: what a game!!!) I recognized that the two teams playing in the biggest game of the NFL season are located in somewhat low-tax states (Massachusetts and Atlanta). Massachusetts has a personal income tax rate of 5.15% and Atlanta of 6.18%. Compare that to California (where 4 NFL teams are located) with 14.10% or Minnesota with 10.15%. But there are also the 0% personal income tax states such as Florida (3 NFL teams), Texas (2 NFL teams), Washington (1 NFL team), Tennessee (1 NFL team).

When looking at the twelve teams that made the playoffs this year and the personal income tax rates of their home states you might think there is some relation:

The average personal income tax rate of players of teams that made the playoffs is 4.62% and thus 1.31 percentage points lower than that of non-playoff team players. The NFL has a salary cap, which means that all teams are allowed to spend the same grand total per season for their players (in 2016: $ 155.27 million). Additionally, the number of players a team could have is limited to 63 (53 on the roster plus 10 on the practice squad).

After state personal income taxes (without accounting for federal income taxes and other charges), every playoff team can spend on average for every of the 63 players (roster + practice squad) $ 2,350,738.51 and the non-playoff teams can spend on average $ 2,318,452.21. So, there is an advantage for every player of $ 32,286.30 per season.

When focusing on the highest personal income tax rates (California – 14.10%) and the lowest (Florida, Tennessee, Texas and Washington – 0.00%) the difference in average per player (after tax) spending amounts to $ 347,509.05 per season. For a player who can choose where to play (undrafted and/or unrestricted free agents) that difference is a great incentive to play for teams in lower tax rate states. Assuming there is a relation between payment and performance, one might be inclined to think there is a competitive advantage of teams in low tax states as they can effectively pay more to their players which means that these teams might be able to sign the better players.

To test this hypothesis I looked at all NFL seasons from 1996-2016 (1996 was my first year of football). This gives me 21 seasons and 663 team-year observations (the Cleveland Browns were reborn in 1999; the Houston Texans started to play in 2002). Every year 12 teams make the playoffs (=252 playoff teams). I ran a logistic regression with a binary variable (playoff) indicating whether a team made the playoffs (1) or not (0), the state tax as explanatory variable, a set of control variables as well as team fixed effects and year fixed effects.

The control variables are critical as you need to figure out what determines the success of a NFL team. Two things are mentioned regularly in the sports media: experience and consistency; experience and consistency at two positions: the head coach and the quarterback. So I use the number of years a quarterback had played in the NFL previous to the respective season (qbexp) and the number of starting quarterbacks the team had in that season (startqb). With respect to coaching, I use the number of years a coach had coached before (coachexp), whether there was a coaching change in the year (coachchange) and whether that new coach was a first time head coach (firstheadcoach). The variable divisionnr is an indicator variable of the team’s division, which is important since 6 of the 16 games (37.5%) in the regular season are played within the division. So the overall strength of the division has some effect on the chances of making the playoffs.  

The results of the regression show that there is a significant negative relation between the state personal income tax rate and the team’s probability of making the playoffs. An increase of 1 percentage point in tax rate relates to a decrease in playoff probability of 4.04 percentage points. To put this into some perspective, the difference between the highest and lowest state tax rates is 14.10 percentage points. This means that a team from Washington (Seattle), Texas (Dallas, Houston), Tennessee (Titans) or Florida (Jacksonville, Tampa Bay, Miami) has a 57% higher probability of making the playoffs than a team from California (Oakland, San Francisco, San Diego, Los Angeles). This year’s playoff teams had on average a 5.29 percentage points higher probability than this year’s non-playoff teams.

To add to this result, I looked at the number of wins and the winning percentage of every team during the sample period (1996-2016). I employ a fixed-effects panel regression looking at every teams’ number of wins in each season and relate it to the team’s state personal income tax rate (the control variables remain the same). The results show again a significant negative relation between the state tax rate and the number of wins. A 1-percentage-points increase in tax rate relates to an average reduction of team wins by 0.2295 per season. The difference between highest and lowest tax rates means a difference of 3.24 wins per season. 

I find similar results when using the winning percentage (1 percentage point increase in tax rates relates to a decrease in winning percentage by 1.4 percentage points). I also ran the regressions excluding the best (New England) and the worst (Cleveland) teams and both over the whole sample period. I also used the difference in the local tax rate to the average tax rate of all teams per season and re-ran the regressions. Every of these additional analyses showed the same trend: a statistically significant negative relation between state personal income tax rates and team success.

The only thing I could not find was a statistically significant relation between personal income tax rates and Super Bowl wins. (However, rather anecdotally, but on Sunday the low tax rate team won and out of the last 5 Super Bowls 4 were won by the low tax rate team.)

So getting into the playoffs is one thing – winning in the playoffs is another…