The Relationship Between Spending Efficiency and Labor Markets

This is Matt Swartz’ first 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.

I’m excited to begin my FanGraphs Residency this month, during which I’ll present an updated analysis of the Dollars per WAR estimates that I’ve used for a long time. I’ve written about the Dollars per WAR framework for analyzing the free-agent market for nearly a decade now, most recently in a three–part series at Hardball Times using data through the 2013 season. In that collection of posts, I established the important definition of Dollars per WAR that I will use throughout this series of articles — namely, the average cost of acquiring one win above replacement on the free-agent market.

Since I’ve written about this, however, there has been a progressively minded, labor-sympathetic pushback against this framework that I felt it was important to address, because if the criticism were fair it would cast a long shadow across all of the analysis in the coming articles. Fortunately, I believe that this criticism is misguided, even if you accept the value system that proponents of this line of criticism generally espouse.

From my perspective, I will remain agnostic on the value system itself in these criticisms, but simply explain why I think this type of analysis does not line up with an anti-labor view at all. I will admit up front that I consult to a Major League team and therefore, when working for them, I do represent the interests of that employer. What I say in these articles, however, will represent only my own views — and, in general, I’m writing this from my perspective as a frequent contributor on this topic predating this good fortune, and as an economist — but neither as a team employee representing ownership nor as a former Department of Labor employee, either.

I’d like to address two well-written and well-argued articles here that I believe characterize some of the labor-related concerns. One by Mike Bates asks if statheads are pro-ownership and another by Michael Baumann reframes a series of team-friendly contracts as inherently bad and unfair. What I’d like to consider here is the implicit suggestion made by both authors that, when teams individually target lower cost-per-WAR players, that this doesn’t affect the prices of these lower cost-per-WAR players and drive them up, but rather that it serves only to drive down the price of higher cost-per-WAR players. This seems very unlikely to be true according to some of the increased prices for lower cost-per-WAR categories of players I find in later pieces in this series.

Naturally, if an individual team attempted to take advantage of lower-cost players by trying to add a fixed amount of WAR to their rosters cheaply, that could lower total salary expenditures for that team. However, if multiple teams were to target these undervalued players, that would drive up their prices — indeed, it appears as though this has happened in reality. Baumann says in his piece that “at some point, ‘finding market inefficiencies’ turns into ‘exploiting your workers,’” but this misses the point that the increased demand for such players will drive up their prices. “Exploiting your workers” is what an employer with market power does, but a market of competing employers targeting certain players together does not exploit them — it bids up their earnings.

Furthermore, teams that can add WAR at a cheaper cost will probably also try to add more WAR via free agency than it otherwise would, and therefore would not even necessarily have a lower payroll in the first place. It’s important to keep in mind that if the goal is a labor-sided belief that players should be making a larger share of the revenue — a goal to the end of which I’ve provided some public guidance — that means actually going after ways to increase total compensation.

In Bates’ article, he turns his criticism inward, acknowledging regretfully that he referred to Edwin Encarnacion’s $65 million deal as “good” because it was team-friendly and Mike Dunn’s $18 million deal as “bad” because it an inefficient use of resources. However, I would assert that Encarnacion’s deal was a better deal for Encarnacion himself, as well as the player’s union, in that it got him and the players collectively more money than Dunn’s deal. The argument on behalf of players should always center on getting a larger collective slice of the revenue, rather than focusing on individual team strategies and assuming that the effect on players’ compensation levels would be unchanged if 30 teams simultaneously behaved the same way.

Here’s another way of looking at this: if one team outbids all others for a free agent, that team raises the total labor expenditure relative to the absence of their actions. That is true whether the player will add a lot of value relative to his salary or a little.

One of Baumann’s critiques of the current system concerns contracts that buy out the arbitration and free-agent years of young players, which I think hits a little closer to the mark. However, I actually think that this, too, fails to alter or suppress total labor compensation in the way he suggests. Firstly, he exaggerates the frequency with which such deals are team friendly. Players and their agents are not fools — they frequently do benefit ex-post when that player gets injured or his skill level otherwise deteriorates. If this were always untrue, they would not be making such deals.

Secondly, Baumann doesn’t consider the effect that such deals have on the spending levels of their teams. When a small-market or mid-market team is able to obtain team-friendly deals for a handful of players, they are more likely to be competitive and more likely to compete on the free-agent market for other players, thus driving up the salaries of those free agents.

But even more generally, the greatest flaw in Baumann’s argument is that he continues to refer to the market value for young players ineligible for free agency as equivalent to the very high Dollar per WAR marks that free agents receive. This would not add up — and, indeed, misses a point that I attempted forcefully to make in the first of the above Hardball Times articles I wrote in 2013.

The Dollar per WAR for free agents in 2016 was about $10.5 million. If all 1000 WAR were compensated as such, that would amount to a $10.5 billion league-wide payroll for players alone, which would outstrip the entirety of league revenue, before accounting for a penny of overhead (including non-player labor). The truth is that free agents are compensated at such high levels because a handful of teams possess a sufficiently talented core of young players that the club’s marginal revenue per WAR far outstrips their average revenue per WAR.

The current labor agreement in Major League Baseball probably is not the system that maximizes labor compensation, but it does enable free agents to earn far more than the average revenue per WAR league wide. A free market for all players every year would lead to much lower salaries for those eligible for free agency under the current system. That is a fact — otherwise nearly every team would be hundreds of millions of dollars in the red each year. They would quickly go bankrupt if they paid all players on the scale they pay free agents now.

Even more generally, the notion that fans or analysts in particular are pro-ownership misses the concept of fandom in general. There will be 2,430 wins over the course of the MLB regular season this year across all 30 teams, regardless of how they’re distributed. Wishing that your Home Nine gets more than its share of these wins is not pro-ownership collectively. In fact, wanting your team to do what it takes to win the World Series over 29 other teams is actually 29 times more anti-ownership than it is pro-ownership. This is a zero-sum game in the most literal sense possible.

The goal of wanting your home team to get more wins per dollar has always centered on the concept of budgets, of wanting your team to win more games for a given amount of payroll. No fan is happy with their team being efficient simply to lower the payroll, and no analyst makes such recommendations for any other reason than to say that a team can pick up some extra wins with the money saved.

I do wish that the original metric presented had been “Wins per Dollar” instead of “Dollars per Win,” because I do think that might have made this point clearer. But the view being taken in all such articles is always about getting more wins for a given amount of dollars, rather than lowering the collective amount of dollars for a given number of wins (and hence going to labor in Major League Baseball overall).

In the subsequent articles in this series, I will present a series of findings about “smart” shifts in the free-agent market since I last wrote on this topic. Just to tip my hand a little bit, I find that teams are no longer paying “Other People’s Players” quite as much more per WAR as “Re-Signed Players” as I had found was occurring both in the 2012 Hardball Times Annual and in subsequent research. However, this is not because teams are paying less for “Other People’s Players” so much as they are spending significantly more on their own players when re-signing them.

I also found in earlier work (included the second piece in the Hardball Times series from early 2014) that relievers were paid far more per WAR than other positions. Teams have gotten “smarter” in this sense, too. They are no longer paying relievers quite as much more per WAR as other players. However, again, this is because they are paying other positions substantially more, not because they are just paying relievers less (and they are actually paying them more in absolute terms).

I have some other findings of interest in the coming articles that I won’t cover here. But the general point is that, during this period of time, labor share of revenue has stayed almost exactly the same. Teams are spending their budgets differently, but the drivers of their budgets are ticket revenues. (See above link on labor share of revenue for details on why I write “ticket revenue” rather than “total revenue.”)

Changing the structure of the institution such that players receive a larger share of revenue is certainly a goal for which labor-minded analysts are right to advocate. But the ability of some club to identify undervalued talent isn’t an impediment to that goal. Smarter front offices will drive the undervalued players’ compensation upward and overvalued players’ compensation downward collectively at a market level. The labor share will not change, however, and if your eye is on that prize, keep it there. Uninformed front offices won’t help the union, while something like indexing the league minimum to television revenue probably will — and, in general, wishing for ownership foolishness is not a logical prescription on which to pin the hopes of any kind of labor advocacy.