The Baseball Economist: The Real Game Exposed (28 page)

The NL Cy Young winner, Chris Carpenter, was the sixth most valuable player in all of baseball in 2005. The writers got this one right, despite blowing the AL award on Bartolo Colon of the Los Angeles Angels of Anaheim—the 49th most valuable player. Despite the fact that starting pitchers only take the field once every fifth day, they still make up nearly half (44 percent) of the top 50. While the number of games in which starting pitchers play is less than hitters, when they do play, they play a lot more. In 2005, pitchers who appeared only as starters and pitched in more than twenty games faced an average of a little more than twenty-six batters a game for thirty games— approximately 780 batter-pitcher contests in a season. A good batter normally steps to the plate a little more than 4 times a game in 150 games—approximately 640 batter-pitcher contests in a season. This gives good pitchers more opportunities to contribute to winning than good hitters, who bat only when their lineup position comes up. Of the pitchers in the top 50, players threw an average of 15 percent of their teams’ innings, while batters took an average of 10.5 percent of their teams’ plate appearances.

Another interesting finding for pitchers is that the top relievers do not tend to be as valuable as you might think. Mariano Rivera had the highest MRP among pure closers, generating $5.89 million for the New York Yankees. That is a mere $39,000 more than his rookie teammate, Chien-Ming Wang, who played part of the 2005 season in Triple-A. The reason for the closeness in their values, despite Wang’s decent but inferior stats, is that he pitched more innings. Do we put too much weight on the value of closers? It’s not clear. The runs that Rivera prevents are very likely important to winning a game. As a closer, Rivera enters the game when giving up runs likely leads to a loss and preventing runs preserves a win. The runs Rivera saved actually had more value than the average run scored against the Yankees, because they occurred in specific spots that would tilt the game one way or the other. However, Rivera pitched less than 6 percent of the Yankees’ innings, nearly a third of what the Yankees’ top starter, Randy Johnson, provided. Still, I suspect the top closers are more valuable than their estimates. Unfortunately, weighting the MRPs for this additional value is quite difficult, and requires some very specific information about performance, so I will not do so here. Just be aware that closers may be a little more valuable than the MRPs estimates.

A complete list of players and their estimated dollar contributions to hitting and pitching in 2005 and 2006 is available in Appendix D. In a free market for baseball talent, teams ought to bid up the wage of a player until it equals his MRP. How out of line are the estimates with what the market is actually paying for talent?

Figure 18 shows the correlations between a player’s actual dollar salary in 2005 and his MRP, for hitters and pitchers separately. It only includes players with more than seven years of major-league “service”—to guarantee their free agent status—and who played more than 2 percent of their teams’ total playing time. The names of players with salaries or MRPs of greater than $10 million are listed. The regression lines map the positive relationships, which are statistically significant. For hitters, differences in player MRPs explain about 27 percent of the difference in actual salaries. For pitchers, MRPs explain about 23 percent. Given some of the limitations we face, that MRPs explain only 23–27 percent of player salaries is not that surprising.

First, even among free agents, valuing what a contract is worth to a player in any given year is difficult, because structures of long-term
contracts vary. A player’s salary in any one year may not be consistent with his performance, but it might be over the life of the entire contract. Contracts often involve back-loading or front-loading of the overall value of the deal. Reasons for such agreements range from tax advantages to payroll flexibility to meet the needs of existing and future teammates. For example, a player might sign a five-year deal that pays out $5 million a season for the first three years and $15 million a season for the final two years. If he performs at a level equal to the average yearly value of the total contract of $9 million ($45/5), then the GM accurately valued the player, even though yearly estimates will miss high and low every year.

Second, predicting what any one player will do from one year to the next is quite difficult. Signing a contract with a team does not bind a player to a certain level of performance; it only requires the player to show up and play when healthy. Both GM and player expectations may differ from what actually transpires over the course of a contract. Few would have predicted Derrek Lee’s $19 million or Carlos Beltran’s $5.75 million performances in 2005, when they earned approximately $8 million and $17 million respectively. It’s not that these performances were out of the realm of possibility, especially given the possibility of injuries, but these deviations in performances, which are quite common, limit the predictability of salaries as a function of performance for any season. In the previous chapter we saw that a pitcher’s previous season performance explains only about 30 percent of his performance in the following season. Similarly, in their book
The Wages of Wins,
economists David Berri, Martin Schmidt, and Stacey Brook report that a hitter’s previous season’s performance explains only about a third of his performance the following year.
⁸¹
A team or player may have been justified in signing a contract that pays a value far different from his added value, based on expecting his past performance to continue in a predictable manner, but the performance of players varies widely.

Comparing salaries to MRPs by salary class is interesting. Because teams have the sole right to employ their own players during any player’s fist six years of service, we should expect salaries to be below MRPs during this indenture. For his first three years in the league, a player makes what the team says he makes, subject only to some minimum salary constraint. After a player’s third year of service, he becomes eligible to negotiate his salary subject to the opinion of an arbitrator. While the player’s negotiating leverage is somewhat limited, we should expect his salary to tick upward. Finally, upon becoming a free agent, a player should earn his MRP as teams bid for his services.

Table 34 lists the median salary differences from estimated MRPs as a percentage by salary class. As expected, reserved players earn less of their MRPs than the other classes. The extracted value for purely reserved players, approximately 90 percent less, is nearly the same level that Scully estimated before the introduction of free agency.

Arbitration-eligible players earn a little more than the reserved class, but still earn far less than their MRPs. While arbitration gives some power to players, the rules of the process require the arbitrator to judge appropriate compensation based on comparisons to players with similar levels of service. Because low-service players tend to earn less, arbitrators continue to favor salary proposals that are far lower than a player’s MRP.

If young players are generating more value than they are being paid, then why do players continue to allow this “exploitation” of their abilities? As puzzling as this might seem, I don’t think it is all that exploitive. In order to become a good player, you need plenty of training. Most players make stops at five levels of the minor-league system during their careers—rookie ball, Low-A, High-A, Double-A, and Triple-A. Along the way, the players do much more than give fans in towns without major-league teams some live baseball to watch. Players receive
heavy monitoring and instruction. Each organization has personnel in place to evaluate and guide players through the minors. Most of the players won’t see a day on a major-league roster. And as good as organizations are at identifying talent, many first-round draft picks flame out, and some undrafted roster fillers become All-Stars. But in order to train those who will succeed, you have to train all players to succeed, for many years. The cost of producing a major-league player is not cheap.

If all players became free agents after every season, teams would have very little incentive to train players to succeed over their careers. Why spend resources to improve a player who will become more expensive to hire the following year—the better he gets, the more money he will command on the free agent market—and possibly employ his skills against you if he’s hired away?

This is similar to the problem private companies face with research and development of new technologies. To develop a cure for cancer, which would be a very valuable thing to discover, drug companies spend billions of dollars every year to get close. All of these companies are willing to spend resources searching for a cure, because a cure will earn them a patent—a time-limited legal monopoly on the cure—that should generate more than enough revenue to cover the costs of development. What would happen if other companies could copy a successfully designed cure and sell it for a fraction of the cost? Then no company would expend the effort to develop the cure in the first place. We would be stuck with a bunch of companies waiting to copy a cure that would never come.

This is the same dilemma faced by teams. Players need teams to spend resources to develop their raw talent into major-league ballplayers. If players are free to bolt upon development, teams aren’t going to be willing to supply the needed grooming. Players and organizations could agree on other payment plans to get them to the majors, such as each player paying to join a private development league, but the players and owners settled on the reserved method through the collective bargaining process. It’s likely that both sides favor the limited reserved-time arrangement because players have unique skills that MLB teams want to groom, and teams have access to specific training that future players want. The arrangement still might be exploitive, but it is not necessarily the case.

While the results for reserved and arbitration classes are expected, the result for free agents seems a little puzzling. In a free labor market, shouldn’t players earn their MRPs? Well yes, but not necessarily the same number we just estimated. The MRP we estimated is a gross estimate, which includes the total value of what the player contributed without taking into account any additional costs, borne by the team, that allow the player to do the things that generate wins. If a team spends $200,000 on a player to produce $1 million in value on the field, his value is equal to $800,000 in additional revenue to the team, which is the net MRP that his salary should equal. That $200,000 may include things such as coaching, weight training, medical attention, video equipment, advanced scouting reports, etc., which teams often provide to players. Without the help of these inputs, for which the team often picks up the tab, players would not be effective on the field. These items represent additional costs of hiring a player beyond his salary. If teams are willing to pay player salaries equal to the value produced on the field minus other training costs, then our gross MRP estimates should be above actual player salaries.