Yesterday was Christmas for a certain breed of MLS need. The MLS Players Union released 2014 player salaries for public consumption. The release does not reflect the most labyrinthine aspects of MLS wage rules, like retention funds, allocation money, etc., but I don’t know of another soccer league anywhere that has player wage data like this available mid-season.
I have delved into all the previous MLSPU salary releases before, and I strongly advise that people resist the urge to focus on specifics here. The odds that a player’s salary listed here it’s their exact salary cap cost fall somewhere on a spectrum between unlikely and impossible. Only the first 20 players on the roster count toward the cap, and the MLSPU release does not reflect allocation money, retention funds, former/lending clubs continuing to pay some wages, or special player statuses like Homegrown, Generation Adidas, or especially Designated Player. Almost feels appropriate that on the MLSPU website this file is listed as up to date through April Fools Day. However, we can get a good sketch of which clubs spend most, and of the wage disparities between the top and bottom of each club’s roster.
According to this release, yes, the guaranteed compensations of Clint Dempsey, Michael Bradley, Jermain Defoe, Landon Donovan, Robbie Keane, and Thierry Henry are higher than those for the full rosters of art least 12 of the 19 clubs in MLS. Those six (1.1% of players on club rosters) make 28.5% of the league’s full player wages. Also, the lowest salary reported, $36,500 made by 54 different players, is 0.5% of the highest, Dempsey’s $6,695,189.00. Of course, those players all bring in merchandising, ticket sales, and headlines that the grunts don’t. It would not be surprising if MLS and club accountants file some portion of the big player expenditures under marketing, instead of wages.
About a month ago I showed that total salaries have been a very poor predictor of league points going all the way back to the first MLSPU release in 2007. That’s not to say they are irrelevant, but their influence is overwhelmingly more subtle on the field than in big European leagues, which some have joked might as well be played on a balance sheet.
Figures like this are sure to be a major topic of conversation in upcoming collective bargaining negotiations between MLS and its Players Union. The current CBA expires after this season, and most of the expected points of contention are related in small or total ways to salary disparity. The players will want higher minimums, free agency, and a big boost to the salary cap, while the league will likely seek to maintain as much of the status quo as they can in the name of profitability and stability. Fans of the league would be well-served by becoming at least passingly familiar with the wage dynamics at play as MLS heads toward this critical juncture. Hopefully the above visual will clarify the issues for some.
I’d like to see this in comparison to the overall numbers in Liga MX or some other leagues, just for comparison.
But otherwise, this is incredible work!
I’d happily do so if I could find the data. Like I said in the article, MLS is about the only place where you get this kind of information in a timely manner. 2012/13 data for the Championship and Premiership was recently released, but all I’ve found is summaries, none of which mention single players.
I will be pleasantly surprised if anyone can point me to data from any non-MLS league that divulges salaries for every single player in that competition.
On the other hand, I definitely could do so with other US sports, and might do just that at some point.
So Toronto has 2 players that are making more than 15 complete teams? That seems fair.
Interesting visualization. I do research on how people understand data visualizations, and I’d like to suggest a couple things to consider in the future. First, you’re using a two-tone, diverging color scheme to represent, as far as I can tell, a single variable, that is, salary. If you’re trying to show that some are above and some below some threshhold (represented in the neutral white/tan), you haven’t labeled that on your legend. Is it the $387,500 cap figure (doesn’t seem like this would make sense to me). If not, it could be more effective to just use a single hue increasing steadily from 36,500 to 500,000. Maybe green to associate with money (in the U.S. at least).
Even if you stick with the two-color scheme, I’m not sure if you mean to give a significance to the middle value on the scale, $268,250, but you seem to by virtue of making it the center value. If you did mean to give it significance, I don’t see it explained in your accompanying text.
Finally, “stephen”‘s point about adding comparison is a good one. People like to compare. Why not add in some general U.S. salaries for comparison, too, that might be more familiar to other people, like, what someone on minimum wage makes in a year of 40-hour-a-week work, etc.? For example, someone making the U.S. Federal minimum wage of $7.25/hr working 40 hours a week, 52 weeks a year, would make $15,080 (which hasn’t changed since at least 2009). Custodial workers in the Florida universities make in the neighborhood of $20,000-$30,000, while the president of UF makes about $525,000 in base pay. Median U.S. Individual income in 2012 was $26,989.
Thank you for the solid critique. Are you on Twitter, and/or do you write about data visualization somewhere online?
The color scheme was probably the aspect that I thought out least when building this, and I do see it as a bit of a weakness (though an improvement over my previous iterations, which drew the ire of a blog called Junk Charts, which I responded to here http://www.stathunting.com/2014/04/03/aiming-for-improvement-in-response-to-criticism/ ). Rather than maxing out the color at the arbitrary round number of $500,000, I should probably have done so at the individual salary cap maximum of $387,500, since every player above this figure equally hinders their team’s ability to spend on the rest of the roster.
Aesthetically, I’ve generally preferred two-tone color schemes, but you bring up a good point that it does highlight the middle value, which is pretty arbitrary. I could set the middle value as the median of the data set, or simply go with a single color, as you suggest.
Comparisons would be interesting, but also a bit beyond the scope of the core goal here, and it does seem to have been quite popular with it’s intended MLS audience already. I could pull some figures from other sports, leagues, and society in general for a secondary tab (and possibly article) later on, though.
The MLS Players Union usually updates this data a couple times each year, so I’ll keep all of this in mind when the next release occurs. For right now I’ve got some other things brewing, and I’ll let this iteration stand as is, despite it’s little flaws.
Thanks again.
Anyone else reading these comments, please don’t hesitate to offer constructive criticism.
Hi Steve, I’m actually a new professor at the University of Florida and haven’t really started a blog or twitter about my work yet; I know I should. I’m trying to find the right outlet to actually reach people who do this stuff. So far I’ve just been reaching out one on one to individual visualizers when I see elements on which I can suggest improvements (mostly through tableau).
Yeah, the call on aesthetics vs meaning-making or understanding is a tough one. If I had to guess in this case, the size of the boxes for individual salaries serves the same purpose as the color, though, right? If they are to scale. Avoiding color (or just using one that increases gradually) would make it more print-friendly, though, too.
If I can be of help thinking some of these things through in the future as you’re creating the next ones, please let me know. As an Extension agent (part of my appointment), it’s actually my job.
So, it’s part of your job to be kind of a data visualization ombudsman? That’s sounds like fun.
I would like to pick your brain on another project I’m working up. Could you email me? Steve@StatHunting.com
Do you think being printer-friendly is a worthwhile concern? Particularly with Tableau’s interactive features, it seems odd to me that anyone would want to print out a dashboard like this. Besides, even if they do want a physical copy, color printers are pretty common, right?
Hi Steve, just saw your comment (and emailed you, as requested).
Yes, I have a University Extension appointment (30%) which says I do outreach about my research. Since my research is on data visualization, I get to do outreach on that. It is fun, for me anyway
Gotta find what you love.
As for color printers, sure, many online readers won’t have a need to print, but many (some?) data journalists and educators are still putting things in print, so as this is really my first public effort to comment about data visualization, I thought I’d throw it in there. It definitely depends on who your audience is. However, there may be people with older monitors, too, and smartphones, etc., that may represent colors differently. For students, at least, color printing in the library usually still runs about 5 cents per page vs 1 for black and white. Some faculty are also reminded that color toner is significantly more expensive, too.
The dialog is just as informative as the Viz! Since the color and size seem to be one in the same, one possible enhancement would be have the color scale based on a ratio to the avg salary in the league?
Pingback: Visualization: 2014 MLS Player Salaries
Pingback: Joining the digital age … at least professionally | Dr. Katie Stofer