Taxes, discrimination, and sports

I found two interesting studies for you sports fans:

1.  Mark Perry links to a Drew Johnson post on the relationship between state income tax rates and NBA team performance.  I linked to the Perry post because it has a nice graph.  Teams in states with no income taxes (Florida, Texas, etc) tend to draw better athletes in the free agent market, and thus post better winning percentages.  Is the relationship real?  They negative correlation is significant at the 99% level.  Most economists consider that to be highly significant; because of the problem of data mining and publication bias, I don’t.  But if the same result holds 10 years from today, I’ll change my mind.

BTW, progressives like Yglesias often point out that no matter what they say, the GOP is devoted single-mindedly to one goal, and one goal only—lower tax rates for the rich.  And I have to agree that that is an obsession of many GOP economists.  But then why the strange pattern of state income taxes around the country?  Income taxes are often higher in conservative Republican states in the South, than in liberal Massachusetts.  Even more surprisingly, the rich in the South are especially likely to be Republicans (as compared to the rich in Boston, NY and LA.)  Yes, there are some GOP states with no income tax (Texas, Tennessee and South Dakota.)  But there are also swing states (Nevada, New Hampshire and Florida) and even one liberal state (Washington.)  Why don’t the southern and Rocky Mountain GOP states at least cut the top rate down to Massachusetts levels (5.3%)?

Let’s see if this study causes (NBA fan) Matt Yglesias to call for abolition of state income taxes in the DC area.

2.  A fascinating study by Christopher Parsons, Johan Sulaeman, Michael Yates and Daniel Hamermesh looked at 3,524,624 pitches and found that umpires in baseball discriminate in favor of pitchers of the same ethnicity.  Even more fascinating is the fact that pitchers seem to know this (at least subconsciously—they might think the umpires of the same ethnicity like them for some other reason) and throw more pitchers where the umpire’s favoritism would help.  Those include more breaking balls and pitches on the edges of the plate, as in both cases umpire discretion is more feasible. 

And here’s the most interesting finding of all:

Roughly one third of ball parks we studied contained a system of computerized cameras (QuesTec) used to evaluate the umpires, comparing their ball/strike calls to a less subjective standard.  Umpires have strong incentives to suppress any bias in such situations, as the QuesTec evaluations are important for their own career outcomes.  With such explicit monitoring, evidence of any race or ethnicity preference vanishes entirely.

On a more serious note, this is a problem for minority pitchers, as 90% of umpires are white.  It also suggests the performance evaluations can’t always be relied upon when looking for discrimination. 

I’m no good at econometrics, so I’d be interested if a reader was able to ascertain whether their empirical results are also consistent with minority pitchers throwing slightly more balls, and minority umpires being much more biased than white umpires.  You’d want to look at the QuesTec part of the study, but I had trouble interpreting those results with a quick skim.  I’m not saying this alternative interpretation is particularly plausible, just that I’d like to nail this down.  With 3.5 million pitches there ought to be a high level of statistical significance.  That’s why interpretation is so important.  Can we rely on this study?  (I already know that we can’t rely on the AER, as they published a paper on stock prices and rainy days with no statistical significance.)



10 Responses to “Taxes, discrimination, and sports”

  1. Gravatar of Brett Brett
    9. July 2011 at 08:15

    Let’s see if this study causes (NBA fan) Matt Yglesias to call for abolition of state income taxes in the DC area.

    Lower D.C. taxes aren’t a bad idea. It would be an incentive to gentrify some of its more poverty-stricken neighborhoods.

  2. Gravatar of John Thacker John Thacker
    9. July 2011 at 08:22

    Income taxes are often higher in conservative Republican states in the South, than in liberal Massachusetts.

    Don’t forget that many of those states haven’t really been Republican at the state level; they’ve had socially conservative fiscally moderate Democrats running things a lot. For example, North Carolina has a relatively high 7.75% top rate, but it was only in the 2010 election that the Republicans took the State Senate for the first time since 1894. (And that was after the NC Supreme Court forced a non-gerrymandered map.)

  3. Gravatar of Matt Waters Matt Waters
    9. July 2011 at 11:31

    I can think of two reasons for the state tax contradiction:

    1. Many GOP states have balanced budget amendments. I know Georgia does at least and probably most other GOP states. Even in GOP states, there is a lot of pressure to not cut education and other spending and small tax increases can be allowed.

    2. Compared to, say, South Carolina, Massachusetts has a much higher per-capita tax base. It also has many more people in the top bracket I would guess. Massachusetts probably has many more people who make high six-figures or more. California, New York and New Jersey also certainly have a huge population of people with very high incomes.

  4. Gravatar of Scott Sumner Scott Sumner
    9. July 2011 at 13:04

    Brett, I agree.

    John, Yes, but there are many Republican states that have higher income tax rates–in the Great Plains and Rocky Mountains, for instance. Places like Idaho.

    Matt, Maybe that’s part of it, but it doesn’t explain states like South Dakota and Tennessee (with no income tax) and states like NY and New Jersey and California, with very high income taxes.

    Any comments on the baseball paper?

  5. Gravatar of Lorenzo from Oz Lorenzo from Oz
    10. July 2011 at 00:51

    On the baseball study, I would be interested in distribution of the effect within umpire racial groups. Is it a consistent tendency or is it concentrated in particular umpires?

    Also, what is the effect of technological checks on the error rate in general? If umpires take more care in general when technological checks are operating, this would reduce the bias effect on its own, since the bias effect is a form of error.

  6. Gravatar of Jeff Jeff
    10. July 2011 at 05:14

    If (i) umpires are 90 percent white, (ii) white umpires unfairly disfavor non-white pitcher, (iii) except in parks where the QuesTech system is installed, then non-white pitchers are more valuable to teams with QuesTech at their home parks than they are to non-QuesTech teams. Do the home park QuesTech teams employ a higher percentage of non-white pitchers?

  7. Gravatar of Scott Sumner Scott Sumner
    10. July 2011 at 06:20

    Lorenzo and Jeff, Those are good questions, and I don’t have the answer.

  8. Gravatar of Philo Philo
    12. July 2011 at 12:41

    For the baseball study, it would be important to set up the “ethnicity” categories the way the umpires themselves see them. Does a Puerto Rican umpire see a Venezuelan pitcher as a “fellow Hispanic,” or as a South American foreigner? Would a Mexican umpire of European descent favor a Mexican pitcher of Indian descent? Do Anglo umpires discriminate in favor of or against pitchers with Polish or Italian names? Would a Siciliano umpire favor a pitcher whose surname suggested that his ancestors hailed from northern Italy?

    The difficulty of setting up the categories properly would tend to produce an underestimate of the discrimination that was going on.

  9. Gravatar of Quote Of The Day at Hispanic Pundit Quote Of The Day at Hispanic Pundit
    12. July 2011 at 23:27

    […] Mountain GOP states at least cut the top rate down to Massachusetts levels (5.3%)?” –Scott Sumner, Economist Share and Enjoy: These icons link to social bookmarking sites where readers can share […]

  10. Gravatar of ssumner ssumner
    13. July 2011 at 11:47

    Philo, Yes, but for the black/white groups I’m not sure that’s a big problem. It depends how many errors you have, and how biased the errors are. For hispanics it is certainly trickier.

Leave a Reply