October 8, 2009
Relative League Quality
I Told You, Don't Call Me Junior!
The American League is kicking the National League's butt, and has been for a number of years. Check out some of the numbers from the last five seasons, courtesy of Jay Jaffe's article last month: a 714-546 record in interleague games, an even better run differential, and not a single year with an interleague winning percentage under .540. The AL is Danny Almonte to the NL's 11-year-old little leaguer.
So how did this happen? Why, all of a sudden, is the American League so dominant? Is it just a cyclical shift, or is there something deeper that will keep the AL on top for the foreseeable future?
First, we need to actually quantify it. You might remember a couple years back when Nate Silver and David Gassko had a back-and-forth over historical league quality. Long story short, David felt that Nate and Clay hadn't regressed to the mean enough, while Nate said that David was regressing too much. Regardless, in terms of year-to-year AL versus NL quality ratings, both sets of numbers should tell us the same thing. So a huge thanks to David, who lent me his numbers from the beginning of time up through 2004:
This is coming from the AL's perspective, so anything above zero means the AL is better, and anything below zero means the NL is better. (I'm also using a three-year rolling average, instead of the single-year numbers.) Clearly this is pretty cyclical, with the NL having a few more total years on top, the AL being a bit more dominant in the years in which it's ahead, and a lot of back and forth in between. Subjectively, we can guess that the AL was dominant in the '20s because of the Ruth/Gehrig Yankees, while the NL was dominant for most of the '50s because AL teams were so slow to infiltrate the Negro Leagues. But we're not going to be able to explain every dip and jump, and those anecdotes don't tell us why the AL is better today.
So if we're looking for objective reasons, the most obvious explanation, of course, is money-based on the numbers we have, this does seem to have a pretty significant effect. The Retrosheet salary database only goes back to 1985, but since then, the r-squared coefficient between relative league quality and relative league spending is .46-in other words, 46 percent of the disparity can be explained by how much the teams in each league are spending. Here's the chart:
Not a very complicated picture. One axis is relative league quality, the other is relative payroll spend (on a per-team basis). As teams in a given league spend more money, the league becomes better. So, no surprise, AL teams have spent a lot more money than their NL counterparts over the past five years. This season is pretty indicative: coming into 2009, the average AL team had a $93 million payroll, according to USA Today's salary database. The average NL team, meanwhile, was just under $84 million. Even if you take out the Yankees, the AL still has a slight edge, coming in at a bit over $85 million.
This trend holds pretty well going all the way back to the mid-'80s. From 1985 to 1991, NL teams were actually outspending AL teams, and did so by about the same margin that AL teams are outspending NL teams today. In turn, the NL was the better league in every one of those seasons, except for a very slight hiccup in '85.
But there's another pretty significant trend taking hold here, and this one gives us an even higher r-squared figure (.48) than league payrolls: the number of teams in each league. From 1977 to 1992, when the AL fielded 14 teams and the NL had just 12, the NL was better in twelve of sixteen seasons. (And the AL just barely squeaked by in the years it was ahead.) In 1993, the Rockies and Marlins evened things up, and the AL played slightly better until the next round of expansion in 1998.
That round, of course, resulted in the current structure, with 16 teams in the NL and 14 in the AL. And from that point on, the junior circuit has dominated, winning six of the last seven years in David's database (which ends in 2004), and crushing the National League in every measurable category since then. (In the chart above, you'll see that the NL hasn't led the three-year rolling average since the 1993 expansion round.)
Those two spans of time (from 1977-92 and 1998 to the present day) are the only extended periods in major league history where the two leagues fielded differing numbers of teams, so it's tough to say for sure whether this is really a significant trend, or just a massive coincidence. But it does make sense: all other things being equal, the league that has to employ more players should inevitably have a longer tail of talent. No matter how you look at it, it's harder to field sixteen teams than fourteen. Add in that the American League teams have access to a slightly better talent pool (thanks to greater financial resources), and it actually makes a lot of sense that the AL has been so dominant.
There's also one other interesting dynamic that should come from having an unequal number of teams in each league, but is hard to empirically prove: each team in the smaller league has a greater chance at making the playoffs, so a perfectly average AL team should expect higher revenue than a perfectly average NL team in a similar market. In turn, that means that the average AL team should also be willing to spend more on players. It makes sense, then, that NL teams were spending more in the '80s than AL teams-the markets haven't drastically changed, just the number of teams in each league.
Could this all change going forward? Maybe, but the National League seems to be starting out with a clear disadvantage, so its teams may simply have to become richer in order for there to be a significant shift. The other question to ask is this: Does it matter? And is it better to be in a smaller but more competitive league, or a larger one that is a bit more diluted? Perhaps these will be issues for another day.