October 12, 2017
The Season in Inequality
Earlier this summer, I introduced the Gini Coefficient. Well, I introduced it to some of you; others were already familiar with it. The Gini Coefficient, invented by Italian statistician Corrado Gini, measures inequality over a distribution. Gini created the measure to gauge inequality of income or wealth. Here’s a standard application, from the Organization for Economic Cooperation and Development, a club of 35 mostly wealthy countries. It shows the Gini Coefficient of disposable income:
The Gini Coefficient ranges from 0 (perfect equality) to 1 (perfect inequality). As you can see in the chart, Iceland, Norway, and Denmark were the most egalitarian OECD countries in 2014. Chile, Mexico, and the US were the least.
I previously used the Gini Coefficient to measure inequality among MLB players hitting home runs and stealing bases. I’m going to guess that this is not what Gini had in mind when he came up with his metric in 1912. But hey, it worked, even if the application was a bit of a stretch.
Here’s an application that’s less of a stretch: Won-lost records. Yes, they’re not the same as wealth or income. But we can compare teams’ winning percentages in a given season on an equal footing—since they all play more or less the same number of games—easier than we can compare home run totals for one player with 700 plate appearances and another with 70.
I calculated Gini Coefficients for every season—separately by league—from 1901 (the year the American League was established) to 2017. Before I show you the results, let’s get our frame of reference. In a 12-team league, if every team finishes 81-81, the Gini Coefficient is 0.00—perfect equality. If one team wins 92, one team wins 90, one wins 88, etc., down to one that wins 70, Gini still scores that as pretty equal: 0.05. If half the teams are 162-0 and the other half are 0-162 (I know, that’s impossible, play along with me here), that’s a Gini of 0.50—pretty unequal, but not at the level of total inequality.
Over the past 117 years, the Gini Coefficient for the leagues ranges from a low of 0.047 to a high of .168. (Note: I calculated Gini Coefficients based on winning percentages rather than team wins; the differences are negligible).
Let’s look at the most unequal season in MLB history, the 1909 American League.
That’s how you wind up with a Gini Coefficient of 0.168, the highest in MLB history. The Tigers, Athletics, and Red Sox accounted for nearly half of all the league’s victories. Nobody else was good, and the Senators were cover-your-eyes bad.
This sort of thing happened a lot in baseball’s early years. Of the 20 most unequal leagues since 1901—that’s among 234 league seasons—the only ones post-World War II are the 1954 AL (fourth-most unequal; Cleveland won 111, New York 103, Chicago 94, and no other team won even 70), the 1950 AL (17th-most unequal; four teams at or above .600 and three below .400), and the 1962 National League (18th-most unequal; NL expansion: Meet the Mets/greet the Mets).
In contrast, the only pre-war year among the most equal seasons, per Gini, was the 1915 National League, which was depleted by the Federal League (which folded after the season).
The most equal season in history was just two years ago:
Of course, that’s not how it worked in 2015. There were three divisions, and the Royals won theirs by 12 games and the Blue Jays won theirs by six. But the league as a whole? Yeah, the teams were pretty close. The difference between the first-place and 15th-place teams in 2015 was less than the distance between the first-place and sixth-place team in the all-time most unequal season of 1909. You can just imagine the entire 2015 American League sitting around a campfire singing “Kumbaya.” Well, maybe not the A’s. But the rest.
The Gini Coefficient for the 2015 American League was 0.04765, just below 0.04839 for the 1974 American League (Baltimore and Oakland only teams with 90 or more wins, California only team with fewer than 72), 0.04840 for the 1968 American League (Detroit won 103 and Baltimore 91 but half the league won between 79 and 86), 0.04870 for the 1983 National League (only two teams won or lost more than 90 games), and 0.04881 for the aforementioned 1915 National League. Those are the only league seasons with a Gini Coefficient lower than 0.05.
So where does the just-completed 2017 season rank? There was a dichotomy between the leagues. The American League had, until the very end, four teams that were clearly playoff bound, three that weren’t, and a massive scrum for the second Wild Card spot. In the National League, only the Cubs had a race in their division, and there wasn’t much of a Wild Card race either. Which is more equal?
Well, as you might expect, it’s the American League, and it isn’t close. The 2017 American League's Gini Coefficient was 0.072. That’s the 53rd-lowest among the 234 league seasons to date. The National League’s was 0.084. That’s the 100th-lowest.
In the 98 league seasons since divisional play began in 1969, this year’s American League was the 39th-most equal, almost squeezing into the top third in terms of equality. The National League was the 28th-most unequal. In percentile terms, the AL was in the 61st percentile for equality; the NL was in the 28th. One league was pretty equal, one wasn’t. The American League was Sweden, the National League was Portugal.
And that continues a pattern. In 2015, the American League's Gini Coefficient was that all-time low of 0.048; the National League’s was 0.090. In 2016, the American League’s was 0.069 and the National’s was 0.075. This year, the American League’s was lower, 0.072 to 0.084.
And that’s consistent with our perception, isn’t it? In the AL, the White Sox and Tigers are clearly rebuilding. But no other team, not even the A’s, really is, and every team in the East was in contention for the Wild Card. In the National League, we’ve heard the T-word applied to the Braves, Phillies, Reds, Padres and, from time to time, the Marlins. The Giants were surprisingly dreadful, as were the Mets. And their foibles created a window for the Nationals (47-29 record in their division) and Dodgers (24-10 against the NL East) to put up gaudy won-lost records.
The American League’s egalitarianism may mean it had fewer super teams, but it had more clubs that could plausibly hope to play baseball this month as well.