The Evolution of Positional Differences in Free-Agent Costs

This is Matt Swartz’ sixth piece as part of his July residency at FanGraphs. A former contributor to FanGraphs and the Hardball Times — and current contributor to MLB Trade Rumors — Swartz also works as consultant to a Major League team. You can find him on Twitter here. Read the work of all our residents here.

Unlike findings about statistical persistency or the physics of batted balls, any discovery about Major League teams’ propensities to spend is based on something less than an inviolable law. As I showed in my previous article about the decline in OPP Premium, teams wising up to an inefficient spending pattern can adjust their behavior in a way that collectively eliminates it.

A related finding from my earlier work on cost per WAR is that players get paid very different amounts per WAR by position. I remained agnostic about whether this was evidence of irrational spending patterns, so much as a feature of competitiveness.

Because teams have become smarter about their free-agent contracts, I decided to review this pattern to see if any changes had occurred. To do so, I looked at positional cost per WAR figures from 2006 to -11 (which was pre-discovery) and then also 2012-16 (the post-discovery era), roughly lining up with my first public work on this topic.

Although there’s some evidence of teams spending on free agents based on outdated valuation methods, there’s also some notable evidence that competitiveness for different positions in free agency plays a role in spending on those positions. When evaluating the numbers, I isolated “defense-first” positions, which included catcher, second base, third base, and shortstop, from “offense-first” positions of first base, outfield, and designated hitter. The key feature of the “offense-first” positions is that many players can easily move between those positions and often do, so teams with a player under contract at the same position as a potential free agent could still safely bid on that player, knowing that one of the two could be shifted to another position. The high cost per WAR of center fielders contradicts the idea that teams were undervaluing players at important defensive positions, because center field certainly is a crucial spot on the diamond. But the inferior center fielder can easily move to left or right field if a team wants two of them under contract. The common thread in high cost per WAR positions is positional flexibility rather than defensive importance.

Pitchers can also be moved around as needed. A great ace can easily be moved to the No. 2 slot in a rotation if another great ace is available as a free agent. A solid closer can become a setup man. The price per WAR for pitchers is definitely higher than defense-first positions, for whom the market is often more likely to be limited.

However, the price of relief pitchers is simply too high to be explained by competition alone. Depending on whether one looks at FanGraphs WAR (fWAR) or Baseball Reference’s version (rWAR), the gap between relievers and other free agents is either gigantic or simply very large. Either way, there’s a notable gap between what teams have paid per WAR for relievers, as compared to what they’ve paid for other free agents.

The table below shows each position’s cost per WAR in the pre-period (2006-11) and the post-period (2012-16). The cost per WAR for all positions did go up, but note that relief pitchers cost per WAR did not increase as much as other positions.

Proceeding with the categories I used in my initial articles five years ago, the table below illustrates the gap in cost per WAR for each group in 2006-11 and then again in 2012-16. Surprisingly, these gaps have remained persistent for all groups except relievers.

From 2006 to -11, the average reliever on the free-agent market cost 117% more per WAR than the average free agent overall. This fell to 68% in 2012-16. However, defense-first positions’ cost per WAR remained roughly 25% below the average cost per WAR for all free agents, and offense-first positions’ cost per WAR actually increased from 14% to 28% above average. Starting pitchers remained about 10% below the average cost per WAR.

The WAR statistic is calculated differently at FanGraphs and Baseball Reference (again, conventionally referred to as fWAR and rWAR, respectively) — and, specifically, in terms of how much WAR is attributed to starting pitchers and relief pitchers. We can actually see in the table below that, when using rWAR, the premium for relief pitchers nearly disappears in the post-period, as teams clearly became more careful about spending on them.

Teams’ propensities to pay different $/WAR at different positions is probably not due to irrationality. Instead, it is probably due to competition, as best evidenced by the fact that teams continue to spend more on positions where there are more likely to be several bidders. When a catcher reaches free agency, there may not be many large-market or win-now bidders who have a vacancy at catcher and wish to spend. However, nearly all teams looking to upgrade via free agency have a way to make room for an ace or an outfielder.

The prescription here for teams is to make sure that they infrequently find themselves in a situation where they need to add a free agent in a competitively bid position as often. That may be easier said than done, especially because teams risk being unable to find a free agent of merit at a defense-first position if they focus on filling their outfields and pitching staffs in advance, as well. However, it does help teams plan to know such things.

These last two articles in this series have found that teams have gotten wiser about spending on free agents from other teams and about spending on relief pitchers. The last thing to check from my earlier work is to see if players with certain skill sets are still over- or underpaid as I found in my 2013 Hardball Times Annual article.