10 Years of Data on Baseball Teams Shows When Pay Transparency Backfires

Aaron D. HillFederico AimeJason W. Ridge

May 09, 2017

You ve probably taken a guess as to how much money your coworkers and others

make, compared with you. Evidence suggests you probably aren t very accurate.

In one PayScale survey of 71,000 people, for example, 64% of those paid the

average market rate thought they were paid less than average. At the same time,

35% who were paid above market rates also thought they were paid less than

average.

Because our perceptions about pay are often wrong, pay transparency has started

to gain popularity. Why not simply inform workers of what everyone in the

organization makes, in order to stave off speculation?

There are pros and cons to this line of thinking. On one hand, such a policy

can guard against discriminating along racial, ethnic, or gender lines while

giving people a firm grasp of where they stand in an organization. On the

other, firms naturally have differences in pay in some form. When people are

made aware of pay inequality (or dispersion ), it can lead to feelings of

inequity that affect satisfaction and motivation, increasing the likelihood

that people will quit. Because of this, some question the wisdom of openly

informing people that pay differences do exist.

We agree with both arguments. Pay transparency can be a good thing, but it can

also be a bad thing if executed poorly. So how does a firm correctly execute a

pay transparency plan? By making sure the inherent differences in pay are

justified by differences in workers performance on the job. Our research,

appearing in Strategic Management Journal, shows that if people know how much

they make relative to others, and if differences in pay can be clearly tied to

how their performance stacks up against coworkers , harmful effects of

differences in compensation can be negated. Pay transparency must go

hand-in-hand with performance transparency something that may seem obvious

but, at least in our experience, is lacking in most organizations.

To analyze this, we used data from an industry where individuals pay and

performance are quite transparent across firms and also, at least for some, are

quite unequal: Major League Baseball. We tested how differences in pay and in

players performance affected the winning percentages of MLB teams from 1990 to

2000, controlling for a host of factors including how much teams spend on

players salaries, relative to other teams; the ability of the manager; and the

overall team talent, which we measured using a comprehensive and comparative

player performance metric created by noted baseball statistician Bill James.

(While sports are different from other industries in many respects, baseball

makes a good test case for this kind of study for several reasons. First, there

s lots of transparency about individual performance and compensation. Second,

the format of the game makes it possible to separate, to a higher degree than

in other contexts, individual performance from team performance. And third, MLB

players compensation is less regulated than in other sports, so we can more

readily observe the impact of pay dispersion.)

We first show that, as in prior studies, pay dispersion is negatively related

to winning performance above and beyond the effects of the other factors. In

other words, the greater the inequality in pay, the worse the performance.

But when we looked more closely at performance, we found that teams actually

perform better (that is, they have a higher winning percentage) when there is a

match between pay and performance. In essence, a team could have a high

dispersion in pay between players, but if the pay corresponds with their

performance, the negative aspects of inequality go away, at least in the form

of a team s winning percentage.

Consider the Oakland A s as an example. In the late 1980s and early 1990s, they

were one of the top-spending and most successful teams, playing in three

straight World Series, from 1988 to 1990, and winning nearly 60% of their games

in 1992. New owners took over in 1995 and changed strategies, slashing payroll

and shifting their emphasis to using younger, cheaper players (as opposed to a

roster of expensive veterans), while also focusing their smaller payroll on

rewarding a handful of key performers. This approach, famously described in

Michael Lewis s book Moneyball, took time, as the team cycled out existing

contracts, so that by 1997 the A s had a very low match between pay and

individual performance.

But then things began to change: The match improved to about league average in

1998, and increased further in 1999. Most important, the winning percentage of

the A s improved along with the improved match in pay and individual

performance; the team went from winning 40% of games in 1997, to 46% in 1998,

to 54% in 1999. In 2000 they won 57% of their games and were back in the

playoffs, with a high match between their dispersion in pay and players

performance.

There are also examples of pay and performance mismatches, either by having

highly dispersed pay or highly similar pay (low dispersion) that is not

justified by performance. In 1998, just a year after winning the World Series,

ownership of the Florida Marlins (now named the Miami Marlins) engaged in a

fire sale, cutting the payroll to one-quarter of its previous total. As with

the A s, certain players and contracts were hard to trade, and the result was

that the Marlins had a highly dispersed payroll the highest in the league, in

fact. That same year, the Montreal Expos (now the Washington Nationals), which

had been cutting payroll for the previous few years, had a roster full of

inexpensive players all paid relatively the same, resulting in the lowest pay

dispersion in the league. Yet for both teams, dispersion in individual

performance was about average, creating high mismatches for the two teams but

in different ways: high dispersion in pay with average dispersion in individual

performance for the Marlins, versus low dispersion in pay with average

dispersion in individual performance for the Expos. Unsurprisingly, the Marlins

won only one-third of their games, while the Expos won 40%.

That same season, however, the New York Yankees set a then-record for wins in a

season, and the next winningest team, the Atlanta Braves, won nearly two-thirds

of their games. While the former was in the bottom quarter of the league for

payroll dispersion and the latter was in the top five, both rated highly in

terms of the match between the dispersions of pay and the dispersions of

individual performance.

In the end, our analysis points to two general conclusions. First, the negative

impact of high pay dispersion is not about equality but about equity. Or, more

specifically, group performance suffers when pay differences or similarities

are not justified by individuals performance. Relatively high or low

dispersion is not in and of itself bad; rather, it is the match (Yankees/

Braves) and mismatch (Marlins/Expos) between pay and individual performance

that creates problems.

While there are some limitations to the generalizability of our research,

including the head-to-head nature of competition in baseball and the openness

of pay and performance data, we suspect similar dynamics may be at play in the

workplace. Sports data is useful for studying pay and performance since it is

widely available, but studies identifying problems associated with pay

dispersion appear in diverse contexts, ranging from hospitals to trucking firms

and concrete pipe manufacturers, and from administrative professionals to

executives in S&P 500 firms.

As such, it is imperative that pay be allocated based on equity, where pay

matches performance. If not, firms that pay more or have better overall talent

may not perform as well as expected, if high performers are paid the same or

less than their lower-performing peers. Similarly, firms with employees who

perform similar tasks, and to a similar degree of quality, should all be paid

similarly to eliminate any harmful effects of pay dispersion. If not, high

performers may lack motivation to continue to outperform their peers who do

less but make the same, and low performers may lack motivation to perform

better when they can make the same as high performers by doing less work.

Our second conclusion is that it s important to understand pay transparency as

a complex issue. As calls for pay transparency increase, it s important for

companies to understand when it may and may not pay off. Sure, it can be a tool

for broader pay equity and for employee motivation. But if pay is transparent

and performance does not justify any discrepancies or similarities, pay

transparency can have disastrous effects.

Consider the infamous case of Seattle credit card processer Gravity Payments,

which, in 2015, set a minimum salary of $70,000 for all employees. Two

employees interviewed by the New York Times ultimately quit. One was frustrated

with people who were just clocking in and clocking out earning the same as he

was. Another was upset by the lack of fairness, noting the company gave raises

to people who have the least skills and are the least equipped to do the job,

and the ones who were taking on the most didn t get much of a bump.

The bottom line for companies is that people are going to make comparisons

about pay and, more often than not, will make them inaccurately. Rather than

hiding pay information or making it accessible without context, organizations

would be better off forming transparent performance metrics, matching pay to

those metrics, and having open conversations with employees about where they

stack up. That, more than anything, is what a truly transparent compensation

program would look like.

Aaron D. Hill is an Associate Professor and William S. Spears Chair in Business

Administration at Oklahoma State University. His research focuses primarily on

how key employees, and particularly executives and individuals in the political

arena like congresspersons and lobbyists, affect firms.

Federico Aime is the Spears Chair of Business Administration at Oklahoma State

University. He received his Ph.D. from Michigan State University and an

honorary doctorate from Moi University (Kenya). Before becoming an academic,

Dr. Aime occupied several senior management positions in foreign and

international organizations such as Citibank and Zurich.

Jason W. Ridge is an Assistant Professor of strategic management at the

University of Arkansas. He received his Ph.D. from Oklahoma State University.

His research focuses on the intersection of executive leadership, compensation,

and political strategy.