Let’s Fix MLB’s Salary Arbitration System: The Arbitrators

In the last installment of this series, we explored the issues posed by the form the arbitration system takes, as well as the constraints a requirement to make an either/or decision when assessing player and team salary figures puts on arbitrators. Today we’ll take a look at the arbitrators themselves, and how they go about their work. To begin, we know that salary arbitrators are typically labor lawyers.

Salary arbitration cases are presented before a panel of three arbitrators, all of whom are among the top labor arbitrators in the country. Why labor? Because the relationship between the Players Association and the Clubs is grounded in labor law and governed by a collective bargaining agreement. When not hearing salary arbitration cases over the first three weeks of February, the panel arbitrators are presiding over arbitrations in the service industry, the building trades and in various other private and public unionized sectors.

Against that backdrop, it makes some sense that the information that helps determine the outcome of an arbitration hearing is typically more in line with “baseball card” statistics than advanced metrics. Lawyers aren’t supposed to be baseball experts, right?

Hitters are typically evaluated using batting average, home runs, runs batted in, stolen bases and plate appearances. There are some positional adjustments, but typically the added defensive value of a shortstop relative to a first baseman is not as important in arbitration hearings as it is on the free agent market. Hitters also can receive larger arbitration awards if they have unique accomplishments, such as winning an MVP award. Pitchers typically are evaluated using innings pitched and earned run average. Starting pitchers are rewarded for wins, and relievers are rewarded for saves and holds. Unique accomplishments, such as Cy Young Awards, matter for pitchers as well.

At the same time, however, it’s unfair – and inaccurate – to say that home runs and runs batted in are all that’s presented in an arbitration. As Jeff Passan relates:

The arguments throughout a case run the gamut. Arbitrators have long rewarded home runs and saves, so they are featured prominently among the players with them, like Oakland’s Khris Davis, who could seek a raise from $10.5 million into the $18 million range. At the same time, the arbitration system is not the antediluvian, abacus-using Luddite-fest it has been portrayed as. The wins above replacement metric is used extensively. So are fielding independent pitching for starters and leverage index for relievers. Statcast data is not allowed in cases, mainly because the league has a far greater plethora of it than the union; and in 2016, when the CBA was signed, the accuracy of spin-rate and launch-angle metrics so vital to modern baseball was not tested out over a large enough sample to warrant their inclusion.

So advanced metrics are making their way into hearing rooms, but are they swaying case outcomes? It doesn’t seem so. MLB Trade Rumors’ arbitration model, which is based on those “baseball card” numbers, remains remarkably accurate – suggesting that advanced metrics, to the extent they’re used, aren’t yet carrying as much weight as they perhaps should.

How can we account for advanced stats taking a backseat to more traditional metrics? Different theories abound. To some legal observers, the consistency in arbitration outcomes, and the usage of the same basic statistics year over year, is a sign of the arbitrators engaging in a form of self-preservation. Jeff Fannell, himself a former agent and assistant general counsel to the Major League Baseball Players’ Association, summarized that view thusly:

Clubs hold the historical edge in salary arbitration hearings. According to the MLBPA website, through 2012, arbitrators have ruled for clubs in 286 cases and for players in 214. This is generally reflective of labor arbitration, where employers win more than they lose. More than one cynic has suggested that such is [sic] way arbitrators ensure they continue to get cases.

What’s interesting, however, is that Fannell seems to have successfully translated at least some advanced metrics into his cases, as a way of undermining the traditional baseball card analysis used by teams. Take this case he highlights, which he worked on while employed by the MLBPA:

JFA represented a Major League starting pitcher who was seeking to more than double his previous year’s salary, despite coming off of a season in which he lost more games than he won. The pitcher was seeking a salary of $8 million, while the club had countered with $6.9 million … JFA crafted a strategy that focused on the quality of the pitcher’s individual performance, while highlighting the poor offensive production of the club. The pitcher’s subpar won-loss record, JFA argued, was primarily a product of poor run support. Based on historical data, JFA demonstrated that the pitcher’s individual performance would have translated into a superior won-loss record had his team provided better run production.  JFA also showed that pitchers with lesser performance in the current market who had superior offensive support, posted winning records – further demonstrating the importance of run production.

Arbitrators found the case persuasive; the pitcher ultimately won his case. A search of the 2012 arbitration results shows that the pitcher in question was likely Anibal Sanchez, who in 2011 posted a 95 ERA-, 88 FIP-, and 84 xFIP- with good strikeout (24.3%) and walk (7.7%) numbers despite an 8-9 win-loss record.

If data like this works, why don’t agents use it more often? There are a few possible reasons. First, the arbitrators might not be well versed in it. Remember, they’re labor lawyers, not front office analysts. As Fannell explains:

As to the latter, while sabermetrics are increasingly being used by teams in evaluating and predicting player performance, the use of such measures in arbitration has proceeded more slowly. So the use of advanced metrics must be carefully considered. If you have to spend valuable time explaining an advanced metric to the arbitration panel, and if the explanation is not an easy grasp, then it may not be worth introducing that metric.

This might explain why players who derive much of their on-field value from their defense have often fared poorly arbitration, at least so far. It’s a lot easier to explain home runs and runs batted in than it is UZR and FRAA. A bigger issue, however, is that a player’s agent might spend their time explaining this data and not know if it even made a difference. To see why, let’s take a look at how an arbitration panel makes its decision:

The next day [after the hearing], the panel chairperson will call the designated Union and MLB representative and report that the panel ruled either for the player or the club. That’s it. No explanation. No rationale. Nothing. Just a stated winner of the case. The Union and MLB then call the player and club, respectively, and report the result. If an outside advocate is used, he or she will also get a call.

On the one hand, this isn’t that unusual for arbitration generally. But then, arbitration generally, unlike salary arbitration, doesn’t have precedential effect. Every ruling an arbitration panel issues has the possibility of being used in the future: “Players are compared to recent players who went through the arbitration process, who played similar positions and who had similar MLB service time.” So when an arbitration panel issues a decision that merely announces the hearing’s outcome, they aren’t telling the parties which arguments they found persuasive. Players’ advocates just have to guess for next time. And the teams, who can coordinate their figures, have a lot more to work with than the players do.

So how can we fix this? There are three fairly straightforward changes to the salary arbitration system that would likely address these problems. First, the arbitration panel could be required to issue a written memorandum opinion for each decision it reaches, and to make those memoranda available to the MLBPA, player reps, and team front offices. We don’t need to make it public, for obvious reasons; a player’s value doesn’t need to be judicially determined in a public forum. And we don’t need thousand-page tomes. But a few pages explaining what the arbitrators found persuasive, how they reached the decision they did, and why they rejected the losing arguments, would go a long way towards creating a measure of predictability and accountability in the process that is, right now, lacking.

Second, teams and player representation should both be required to disclose to the arbitrators, and the other side, their own proprietary measures of that player’s value. If a team is claiming at arbitration that a player is worth $5 million in arbitration and not the $8 million he is requesting, the player should know if the team’s own internal valuations show him as being worth $16 million. How those numbers were calculated can be kept confidential except as is necessary, but there is a lack of transparency in the present process that allows the parties to take contradictory positions that don’t serve the interests of the process.

And third, the arbitrators themselves should be signed for long-term, multi-year contracts, be trained in advanced metrics from a third party not affiliated with MLB or the MLBPA (like FanGraphs or Baseball Prospectus), receive pay by the hour instead of by the case, and be granted contractual assurances that their decisions will not be cause for the termination of their contract by MLB or the MLBPA. This insulates the arbitrators from the parties they are presiding over and adds a measure of independence. Even whispers that an arbitrator rules in the team’s favor for self-interest can be damaging to the process, whether or not those rumors are well-founded. Separating the arbitrator from the parties in this way ensures that the process is viewed as impartial.

Arbitrators face a difficult task, one that is likely to get harder as advanced stats continue to proliferate and the number of hearings over which they preside increases as teams employ the “file-and-trial” method. As labor lawyers, they are well equipped to engage the law; the hope is that these changes would better position them to engage baseball as it is now. Doing so can only improve arbitration outcomes.