A Long-Run Analysis of Salary Inflation

The contracts that baseball players sign are some of the longest contracts in business — not just sports. When handing out nine- or ten-year deals, projecting salary inflation is critical, and yet getting an accurate forecast is nearly impossible.

Prince Fielder and Albert Pujols signed huge deals this offseason, which pays the sluggers into 2020 and 2021, respectively. The rate of salary inflation over the next decade will play a big part in how those contracts work out. The most common assumptions are $5 million per WAR and a 5% inflation rate, which would mean Fielder would have to produce 36.9 WAR over the life of the contract in order for the Tigers to “break even.” The break-even number for Pujols’ back-loaded contract is 37.8 WAR:

Year Fielder Money Pujols Money $/WAR (5% inflation) Fielder Break-Even WAR Pujols Break Even WAR
2012 $24 $12 $5.0 4.80 2.40
2013 $24 $16 $5.3 4.57 3.05
2014 $25 $23 $5.5 4.54 4.17
2015 $25 $24 $5.8 4.32 4.15
2016 $25 $25 $6.1 4.11 4.11
2017 $25 $26 $6.4 3.92 4.07
2018 $25 $27 $6.7 3.73 4.03
2019 $25 $28 $7.0 3.55 3.98
2020 $25 $29 $7.4 3.38 3.93
2021 $30 $7.8 3.87
Totals $223 $240 36.9 37.8

If we instead assume inflation slows to 3% per year, those break-even levels change to 39.7 and 41.4. It might not sound like much — 3.7 WAR over ten years in Pujols’ case — but that translates to $18 million of contract value.

Everyone knows that baseball players are well compensated for their skills. The gap between baseball salaries and the rest of us massive, and it’s increasing over time. In 1992, the average MLB salary was 32-times the annual income of an American household. By 2010, baseball salaries were 66-times higher.

Is it possible for this trend to continue? Yes and no.

The difference in salaries between Major-Leaguers and the rest of us will continue to increase, but not at its current rate. In the long run, U.S. households will make about 3% more per year, but baseball salaries will increase at about 4% per year, meaning the gap between the average baseball player and the average American will widen by about 1% per year.

On a couple of occasions, I attempted to create a model which would predict salary inflation in baseball for the upcoming season. Both of those attempts looked at short-term predictions — the expected salary inflation given the current state of baseball revenue and the current economic health of the country.

If we pull back and look at the issue in the long term, the new perspective leads to the new 4% conclusion.

The first step is to trace where the money for players’ salaries comes from:

The above chart does not, under any circumstance, represent every nuance of this very complicated system. However, hopefully it shows the source of most of the money behind teams’ payroll.

Let’s start at the top. Households make money. Part of that money is taxed, some money is saved, and everything else is used for consumption. If you multiply the amount of household consumption by the number of households, you get the total pool of money being spent on goods and services.

Some percentage of that pool is spent directly on purchases from a team. That money is revenue for the team. Most of the total consumption money goes for non-baseball goods and services, but some of that trickles down to baseball as well.

As corporations earn money, they have money to spend on advertising, which is a significant income stream for both individual teams and the MLB as a whole. Advertising dollars are what drive the MLB’s contracts with television networks, and thus revenue which is split between the 30 clubs.

So, teams get money from direct sales, team-specific ads and licensing contracts, and a cut of the MLB pie. Yes, this is an over-simplification, but I am trying to keep it easy.

When teams earn revenue, it gets allocated to various parts of the organization, with a large hunk going to pay for player contracts. When teams have more money to spend on players, the supply of players is (presumably) fixed, and thus $/WAR inflation.

What does all of that mean, and how does that lead to a 4% per year conclusion? Well, the two boxes that start the whole process are always increasing. Household income increases by about 3% per year — which leads to 3% more consumption per household. The U.S. population increases by about 1% per year. Therefore, 1% more people and 3% more money yield a 4.03% increase in the pool of total consumption dollars available. As long as all of channels maintain a constant flow to the proceeding step, on a percentage basis, that 4.03% makes it way all the way to the bottom.

But wait a second, if the above is true, then why have baseball salaries risen by 5% on average, and by double digits in some years? Well, the flow chart assumes that the percentage going from one level to the next remains constant. For example, the percentage of income used for consumption or the percentage of team revenue dedicated to payrolls. If any of those increase, then everything below increases, and the amount of salary inflation increases.

For that reason, this is a long-term analysis. I would not be surprised to see salary inflation continue on its current 5% path for a number of years. Eventually, whatever channel that extra inflation is coming from will have to hit a ceiling. Things like the share of income spent on tickets or the percentage of advertising dollars spend on baseball simply cannot increase without bounds.

Whenever the percentage of cash flow from one step to the next is constant, salary inflation will be driven solely by the increases in nominal income and population, which should yield about 4% market inflation.

As stated earlier, this means the gap between baseball players and average Americans will continue to increase. The reason this exists in baseball (and other sports) is that the number of jobs is a constant. Increasing payroll means either increasing the average wage or hiring new employees, and in baseball there are only 750 major league roster spots, so the average wage has to carry all of the inflation.

When will the short-term become the long-term? One can only speculate, but on a 10-year contract, using the long-term estimate seems like a good educated guess.




Print This Post



Jesse has been writing for FanGraphs since 2010. He is the director of Consumer Insights at GroupM Next, the innovation unit of GroupM, the world’s largest global media investment management operation. Follow him on Twitter @jesseberger.


40 Responses to “A Long-Run Analysis of Salary Inflation”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. John says:

    HAHAHAHAHAH

    Vote -1 Vote +1

    • Nick says:

      +1 to John

      While I appreciate the author’s responses to a number of comments/questions, John pretty much sums up how seriously this post should be taken. Jesse’s admitted lack of scope on a number of salient issues should have disqualified this article from seeing the light of day.

      My humble suggestion is that Jesse gets out from behind his desk, work/live in the real world of finance for a few years/decades, and then come back to the grown-ups with an informed opinion.

      Vote -1 Vote +1

  2. Xeifrank says:

    Do you think that these local cable deals that teams are signing or are rumored to be signing soon (Yankees, Rangers, Angels, Dodgers) are or will be driving up salaries/payrolls at an increased rate over what we have seen in the near past?

    Vote -1 Vote +1

  3. Tasintango says:

    hahahha

    Vote -1 Vote +1

  4. Paul SF says:

    Given the enormous disparities opening up between the wealthiest percentiles (of which baseball owners are certainly a part) and “the rest of us” in other aspects of society, I’m not sure why we would expect this model to hold up. For one thing, there’s investment income the owners can and do receive (unless they’re Madoffed). Not only does that bear no relationship to the overall trends of household income and population growth, but it’s subject to lower taxes than traditional income, which arguably would make it the preferred source of revenue for an owner and therefore an increasing piece of a team’s pie.

    I could be way off, but it seems this oversimplifies too much if it’s not looking at other sources of revenue that ownership could use to supplement a team’s payroll.

    Vote -1 Vote +1

    • Agreed. There is some percentage of payroll that comes from investments, which is not included in the flow chart. But unless that percentage changes over time, then the model holds up. Will it change over time? Certainly, but not without limits.

      Vote -1 Vote +1

  5. Craig says:

    I think the trend is going to continue for a short time, but then level off.

    A big part of the player salary inflation is coming from teams opening new ballparks. The great disparity in player vs household income started in 1998 by your line graph. By my count there have been 15 new stadiums built since then (16 if you include the Marlins opening one this year). That’s huge new revenue for each team to spend on big contracts. In fact there are really only 4 “old” stadiums left that have not been built or majorly renovated in the last 25 years – Fenway, Wrigley, Dodger, and Oakland-Alameda-whateveritscalledthisyear.(someone correct me if I am mistaken)

    There are also the big new TV contracts which will probably continue to proliferate for a few more years. That’s like free money to a ballclub so it will certainly fuel more spending, as it has with Anaheim-not-in-L.A.

    So my guess is 5 more years of continued high salary inflation and then we will see it level off for quite some time.

    Vote -1 Vote +1

    • corvelay says:

      More specifically I would say that it is the large government subsidies they receive to build those ballparks that contribute to salary inflation. It is no coincidence that Marlins spent $190 million on Reyes, Buehrle, and Bell this offseason as they prepare to open a park paid for with $500 million in public funds after spending around $200 million total on payroll the previous five years. The public financing shifts some money to the payroll pool that would otherwise be spent on infrastructure, or in the case of parks that would not be built without public funding, generates increased revenue from the new park while the team does not have to cover all or even most of the capital costs.

      Vote -1 Vote +1

      • Craig says:

        Well there may be some shifting, but teams are probably borrowing the money for their share of the construction cost anyway. I doubt the capital funds mix much with operational funds. The revenue increase comes from increased ticket sales, premium seating and concessions.

        Vote -1 Vote +1

  6. MikeS says:

    Occupy MLB!

    Vote -1 Vote +1

  7. newsense says:

    You’re making two very questionable assumptions. The first is that the propensity to consume baseball is directly proportional to income. If I make twice as much money, am I likely to spend twice as much on baseball? Could be less, could be more. The second assumption is that everyone’s income increases at the same rate. Over the last 10-20 years, the incomes of higher income groups (the “1%” or even the “10%”) have increased ayt a faster rate than lower income groups. If the high income groups are baseball-crazy,baseball revenue will grow much faster than the economy as a whole.

    Vote -1 Vote +1

    • Jason says:

      He’s making an even bigger assumption than that. In any other free market wage inflation is driven by the size of the available talent pool relative to the demand for that talent. Software engineers make more money than janitorial staff becuase the talent is more skilled an harder to acquire. It has nothing to do with how much money ACME sanitation services grossed the previous quarter.

      I think the author needs to explain why he is going though all these contortions rather than looking at the size of the potential labor pool for elite baseball players relative to the number of roster spots?

      Vote -1 Vote +1

      • LarryM says:

        I don’t know whether the author’s assumptions will hold up, but I think it’s interesting that a lot of the comments that disagree are basically comments by people who THINK they have a good understanding of economics, but don’t.

        This is a prime example. The market for baseball players is a peculiar one in many respects (see below). I don’t think that anyone who (a) understands baseball, and (b) understands markets, would assert that baseball saleries in the long run have anything at all to do with “the size of the labor pool for elite baseball players.” I’ll touch briefly on the theoretical reasons for that below, but let’s first note that empirically there has been no such relationship (if anything, an inverse relationship, but that’s purely coincidence).

        Now, as for the reasons: I could point to all the ways in which baseball salaries are not set by classic market mechanisms (a plethora of contractual rules regarding salaries of players subjected to the draft, now international signings as well, in addition of course to pre FA players). But that’s not even the biggest factor. Ultimately the competetive structure of baseball is such that the “pool” of players is roughly fixed (for a given number of professional teams & a given roster size). That’s because the “pool” does not consist of all players capable in some abstract sense of playing “major league” baseball. It is a pool fixed by definition as the very BEST players available to play professional baseball in the United States. Key point: “elite” is defined in relative terms. If the talent pool doubles, that doesn’t mean twice as many “elite” players. That means the definition of what it takes to be an elite player shifts over time based on the available talent pool.

        Vote -1 Vote +1

      • LarryM says:

        I do think that the author’s assumptions may be off, but not for reasons touched upon here. The biggest problem (and IMO the reason why salary growth has been at 5% rather than 4%) is that baseball has been very successful at broadening it’s “share” of the entertainment dollar through better exploitation of revenue sources. Of course this is self limiting, so maybe more an explanation of the past disparity rather than a flaw of the author’s model goingforward. Perhaps a bigger issue going forward will be the ability of teams to tap international revenue sources.

        The more I think about it, though, this is actually a pretty elegant and probably pretty accurate LONG TERM model. Ironically, some of the critiques in comments if anything bolster the model, as they focus on “flaws” which do not exist.

        Vote -1 Vote +1

      • Jason says:

        Larry – this is wrong: If the talent pool doubles, that doesn’t mean twice as many “elite” players. That means the definition of what it takes to be an elite player shifts over time based on the available talent pool.

        As the talent pool doubles the variation in performance (even among the elite) is reduced… thereby depressing the value of the marginal delta in performance between players.

        A hypothetical I like to illustrate the point. If a cruise ship crashed on an uninhabited island and started a baseball league, their would be wide variations of performance among the best of the best. Guys that played high school or college ball would put up monster numbers against players of minimal athletic ability needed to round out rosters. This is roughly the state of baseball in the 1920’s or so.

        As the talent pool expands, and new sources of talent are discovered, the variation is dimished. Even if we limit our scope to the US, ignoring for the moment that in the last 60 years the game has found many new sources of talent among African Americans, South America, and the Pacific Rim… the growing population of the US alone will produce more “elite” baseball players than are being produced today.

        Vote -1 Vote +1

      • LarryM says:

        I’m going to have to punt to some extent, not because I agree with you and not because I don’t have an answer (well, several answers), but because we’re getting into territory that can’t be fairly dealt with in a comment thread. I’ll make a few points going the other way, but will have to leave the more sophisticated arguments out for lack of time and space.

        (1) The empirics of the past 20 years go the other way. The trend toward decreased variation you cite from the 20s ended and reversed around 1990 (though is now perhaps once again going the other way). That despite large increases in the pool of talent. I’d add here that there are convincing alternative explanations for the earlier trend toward decreased variation (see Gould, Stephen J.).

        (2) The fact that we are talking about the very rightward tip of the Bell curve invalidates your argument for reasons that are hard to set forth in a comment thread. Yes, I’m punting, but I don’t want to write a short novel.

        (3) Even were you right, that wouldn’t necessarily have the salary implications that you claim. The number of players in the “free market” of baseball players is essentially fixed (the number of free agent baseball players). If the variation in talent among those players declined, that might effect the distribution of their salaries but not the total spent on salaries. Why would it? It’s not as if a reasonable alternative (for a team with money to spend) is to dip into the pool of “freely available” talent. That talent is by definition replacement level. That won’t change even if the talent pool quadruples.

        (4) I’m not sure that this is a logically distinct argument from #1, #2 and #3, but if we look at the free agent era, free agent salaries have been driven by increased team revenues, not variations in the talent pool. The reasons for this are merely touched on above, but comes down to this: major league teams are interested only in playing the very best players available. Increase the talent pool all you want, you’re still bidding for the same number of “very best” players. Even if the variation is less, the difference between even a mediocre starting baseball player and an “AAAA” type ball player is huge. Decrease that gap a little, and teams will still pay the same premium to have the mediocre regular over the “AAAA” player.

        That’s maybe the tip of a very big iceberg of analysis as to why you are wrong.

        Vote -1 Vote +1

    • Great question. I’m making a ton of assumptions, including the two that you mention. Income in various for various groups has changed over time, but as long as there is some long-run percentage for each group, the model would hold up. I’m not suggesting that we’re at the long-run right now, whenever the growth rate of each group hits its long-term limit, the theory would hold.

      Vote -1 Vote +1

  8. Dekker says:

    With TV contract deals becoming more lucrative, I imagine MLB teams make more money. By adding two additional playoff spots, MLB teams will be willing to spend more on players, continuing the trend of escalating salaries.

    Vote -1 Vote +1

  9. Lex Logan says:

    As an upper limit, sure, baseball revenue and player salaries cannot outpace GDP growth indefinitely. We don’t know that GDP growth (which includes productivity gains and population growth) will continue at historical rates, but that’s a reasonable guess. However, as baseball accounts for only a very small fraction of GDP, and the potential market extends beyond the Unitied States, I don’t think the 4% has much relevance in our lifetime, or our granchildren’s. Baseball has vast room to expand or contract in popularity, if nothing else.

    Vote -1 Vote +1

    • I meant to mention something about international revenue, but ran out of space. Right now, I imagine international revenue to be a drop in a bucket, but if it can continue to grow, it is certainly a reason why salary inflation would rise at higher than 4%.

      Even with international revenue though, there is some long-term percentage of income from foreign countries that must exist. It can’t increase without bounds or else we’re saying that eventually foreign countries will be spending 100% of their income on U.S. baseball.

      Vote -1 Vote +1

  10. tz says:

    Salary inflation aside, I can still run holes through the $5M per win formula. Here’s my biggest – it doesn’t add up!! :

    Average team has ~30 WAR (if replacement level roughly a 50-win team) ->
    30 wins x $5M/win = $150M average payroll.

    (last I checked, average MLB payroll was still < $100M)

    Vote -1 Vote +1

    • Xeifrank says:

      Isn’t the $5M for a win on the free agent market? Lots of MLB players are still under team control.

      Vote -1 Vote +1

    • Craig says:

      I’m not sure of the answer, but does the $5M number reflect the average cost of all WAR (your argument) or the marginal cost of an additional win (different calculation)?

      Vote -1 Vote +1

    • Jason says:

      There are a lot of ways to run holes through the $5M WAR formula – but the way you identified is not one of them. The formular refers to the cost of labor on the free market… most major league talent isnt acquired through free agency.

      The problem I have with it is 1) teams understand that production will decline over the life of the contract but prefer to structure the agreements to have similar or increasing AAV… thus we should be looking at projected WAR over the life of the contract divided by the $… not WAR from the previous year divided by AAV. 2) Even a “market” deal gives the team surplus value. Thats basic economics. Its how corporations make money.

      Vote -1 Vote +1

    • Seattle Homer says:

      That math actually strengthens the $5M/Win point for me. If you were to build a team from scratch using only free agents at market value, it seems about right that it’d take something a fair bit north of $100MM to build a .500 team… As Xeifrank says, most teams have tens of millions of dollars of surplus secured in their younger players. (Think Brett Lawrie would get more than $400k on the free agent market?)

      Vote -1 Vote +1

  11. Anon says:

    This analsis also doesn’t consider growing international revenues, which add to the total pot of revenues available, and ultimately drive salaries. This could be the major growth driver over the next couple decades, that makes the revenue increases more on the order of the 5-7% that’s been observed over the past couple decades.

    Vote -1 Vote +1

  12. pft says:

    Also. one must consider the valuation of the teams, which is inflating even more than revenue growth. The Dodgers for example look to be sold for 1.2 billion-1.6 billion. McCourt bought the team for 430 million in 2004, and has seen his investment increase almost 200%-300% .

    During this time, Dodgers payroll increased 90 million to 120 million, or about 33%, while revenues increased 189 million in 2005 to 246 million in 2010, at about the same rate as team payroll.

    The valuation bubble, which is in part due to the TV rights bubble, may fuel players salary inflation once revenue growth slows. Profit gets taxed yearly, valuation growth does not. Losses can be carried forward to offset future tax liability, and teams will willingly except losses if it helps keep fans interested and the teams value continues to grow.

    However, once revenue growth slows, and if the valuation bubble bursts, look out below.

    The other consideration for salary growth projections is the possibility MLB collusion, and the MLBPA complacency. MLB managed to sneak in a defacto salary hard cap into the new CBA by linking revenue sharing in the form of rebates to the luxury tax. Teams like the Yankees will suffer effective tax rates of 150% unless they stay below the CBT threshold. How the players agreed to this is beyond me. As for collusion, it can be as simple as teams meeting and agreeing to cap FA salaries at certain positions, or agreeing on a model to use to value players that undervalues players actual contributions. MLB got caught in the 80’s, but these are very difficult to prove.

    Vote -1 Vote +1

  13. Justin says:

    Wages for middle class Americans are certainly not growing 3% every year

    Vote -1 Vote +1

  14. kid says:

    In a Free Market the only way we can exercise force and change the status quo is with out wallets. Of course one person deciding to not attend a baseball game because he/she disagrees with the ludicrous contracts won’t make a difference, but it’s fun to think about what would happen if fans said “Enough is enough – change or we won’t come.”

    Vote -1 Vote +1

  15. pkdryan13 says:

    While my question’s only semi-related, it’s still related; given the knowledge we have of aging curves, doesn’t it make more sense to frontload long-term contracts rather than backload them (from a team’s perspective)? Since almost all free-agents signing with a new team are, at the very earliest, in their prime years at the time of the deal (i.e. Prince), while many, (if not most) are past their prime. Seeing as how their output should decline as they progress through their thirties, shouldn’t teams want to pay them more in the years in which their output is greater? That way, when Player X is posting a 1.9 WAR in the 7th year of his deal, he’s not taking up 35% of the teams payroll, thus allowing the team more funds pay for production (meaning players) they would have otherwise been unable to. In other words, frontloaded contract structures reflect a more steady ratio of money per WAR through the entirety of the deal. I know inflation’s also a factor but if the contract were initially in real terms, then each year could be adjusted for the inflation rate of MLB salaries.

    Vote -1 Vote +1

  16. MG says:

    This post should after ended after what WAR value that Fielder/Pujols table. Stick to marketing.

    Vote -1 Vote +1

  17. JoeElPaso says:

    This analysis, while admirable in its intentions, suffers completely from what is called the ecological fallacy (nothing to do with the environment). In that fallacy, one infers from general conditions (total household consumption in this case) to specific conditions of a subgroup or individual. However, without knowing the internal distribution of the data and the placement of the subunit in that distribution, we cannot make that leap. Some commenters above hinted at this, by pointing out that household income is unequally distributed, and growth in total income and population are largely irrelevant to a market mainly made of prosperous people and, increasingly, corporate buyers of boxes. But this is not all. The various submarkets, revenue streams, and specific rates of inflation cannot easily be decomposed from national aggregates. Various industries rise and fall as shares of that national whole. Baseball, despite amazingly greedy and stupid owners, almost certainly has risen, seen in the faster-than-national average increase in gross industry revenues in recent decades. The reasons are complex, but include shameful public subsidies (to rich owners), increasingly unequal distribution of income, transformations in the media industry (cable/satellite sports channels), and the underlying wonderfulness of baseball (I’d like to think). I do want to close by saying, however, that some effort to either measure historical free agent inflation rates and possible future inflation rates is highly interesting.

    Vote -1 Vote +1

  18. Sean says:

    I keep hoping that sports franchises will invest new revenue into lower ticket prices, but there’s little financial incentive for that :( I hate that ticket prices for the teams in Boston are right at the point where you almost don’t buy the ticket out of disgust. I’d likely go to more games if I felt like ticket prices were appropriate.

    Vote -1 Vote +1

  19. Jeff says:

    Draw a similar chart for the amount of expendable income spent by consumers on media, entertainment, and sports fashion and you’ll probably see it matches up perfectly. We are to blame.

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>