Comparing free agent deals with estimates points to a more complicated off-season story


Jays From the Couch looks into the slow off-season for the free agent market, which is more complicated than you might think


Embed from Getty Images


The 2017-18 MLB off-season has been notable for its extremely slow pace. Over the previous four off-seasons, an average of about $1.2 billion was spent on free agents listed on Fangraphs’ Top 50 by January 1st. This off-season, the total was only $426 million. A number of explanations have been offered for this slowness, the most serious one being outright collusion by MLB front offices.


Now, my intention today isn’t primarily to jump into that hornet’s nest. Collusion is bad and should be punished when evidence is found that it has occurred. That said, the appearance of collusion is much more common than collusion itself. Here’s the standard explanation I give to my students. When three large corporations overcharge their customers on, say, cell phone plans and avoid competing via lower prices, there is the appearance of collusion. Now, collusion may certainly be at play (I have no interest in going to bat for giant telecom corporations). However, the three companies may simply be fearful that if they unilaterally lower their prices, their competitors will do the same, resulting in lower prices, the same market share and, ultimately, decreased revenues. So, they don’t bother competing via prices, instead using non-price competition like advertisements and sponsorships (items with much more predictable costs). Over time, as production costs increase, the companies slowly and steadily increase prices together. It can look like everything is coordinated, and it is in some cases, but usually there are more innocent explanations.


Shifting focus to the MLB specifically, I think that the appearance of collusion is more likely the case right now than actual collusion itself. While many of the non-collusion explanations can seem like (and might indeed be) management spin, some of them are very compelling reasons for the slow off-season. Ross Atkins has discussed the fact (as have many others) that front offices value players very similarly nowadays. Free agent signings often result from one or more teams pushing past their initial maximum offer to move beyond the pack and get a deal done. Everyone being on the same page may make those scenarios less likely.


Other explanations make sense as well. A number of them speak to the idea that front offices aren’t as motivated to spend on free agents as they previously were. Fans have grown to perceive full-on rebuilds as a good thing (just look at the clamouring among some Blue Jays fans for two tank years before 2020). Small-market teams have worn down fans over many years to accept low payrolls. A higher luxury tax on repeat offenders combined with an upcoming monster free agent class has motivated the big spenders (who help set the market in most off-seasons) to focus on getting under the luxury tax threshold. The tighter international signing rules create more incentive to build via those cheaper signings than expensive veteran signings. Super-teams are omnipresent (each division has one) and are motivating the mid-level teams to take a more long-term approach—the just-regular-good teams may think that if they make a splashy signing to shrink the gap a little bit, their division’s super-team will just make their own big signing to pull away again. So, instead, they may feel it’s more prudent to develop their own super-team via player development. Fans are more engaged and analytical than in the past, which leads them to look down upon the big-money, back-loaded, might-become-a-boat-anchor kind of move and support the smart value plays. [Many other plausible explanations abound.]


I haven’t sat in on meetings between owners, so I couldn’t possibly say that collusion is or isn’t happening. Moreover, I have a lot of empathy for professional athletes and the fact that they don’t receive nearly as much of the profits generated by their activities as they deserve. Sure, it seems weird to say that someone earning tens of millions of dollars a year is “underpaid”. But, given that MLB players help generate upwards of $10 billion per year in revenue for owners, it really isn’t that weird a thing to say or think. Especially, when the owners aren’t particularly risking anything themselves, often having stadiums (which should be their primary overhead cost) given to them for free (or close to it), by setting governments against each other.


Increasing the players’ share of revenue will likely require a few solutions. One interesting idea is restricted free agency. Instead of arbitration, players in their fourth to sixth seasons would be eligible for restricted free agency. Other teams can sign the player, but their current team has the right to match any offer and receive compensation for a signed player. The compensation would be tied to the quality of the player lost. This system would see the market determine the player’s salary, rather than an arbitrator, which should increase salaries for players in their mid-20s, even before they enter unrestricted free agency.


In thinking about the existence of collusion, a big unknown for us outsiders are the actual offers being given to the biggest unsigned free agents. Yu Darvish may be unsigned because he’s only received four-year, sub-$100 million offers (in which case, yeah, collusion is probably happening). Or, it’s because he is waiting on the Dodgers or Yankees to clear some payroll space for him. Or, one of a hundred other explanations.


In that vein, the top remaining free agents all seem to have some potential reasons for being unsigned. It sounds as though J.D. Martinez has offers (in the range of $120 to $150 million) that surpass his Fangraphs’ median crowdsource estimate (five-year, $110 million). Moreover, Fangraphs uses comps to project that he’d be fair value for a six-year, $150 million contract. Nevertheless, Martinez (and his agent, Scott Boras) are apparently waiting out for an offer in the $180 to $210 million range. Whenever he signs, he’ll be well above his median crowdsource estimate.


Jake Arrieta probably wants to get paid like the Cy Young winner he was in 2015. Unfortunately, more recently he’s pitched closer in quality to a #3 starter, maintaining a 4.23 FIP over 262 IP since July 2016.  Among the 164 starting pitchers who pitched 100+ innings since July 2016, Arrieta ranks 73rd in FIP. To be super clear, Arrieta remains the second best starter available (and an above-average pitcher). However, a disconnect between the player’s self-valuation and the valuation of front offices would lead to a drawn out negotiation process.


The same goes for position players like Eric Hosmer and Mike Moustakas. Is Hosmer the guy who posted fWARs of 1.0, -1.7, 0.0 and -0.1? Or, is he the guy who posted fWARs of 3.2, 3.5 and 4.1? Again, regardless of those issues, Hosmer has an offer on the table from the Royals (seven year, $147 million) that would surpass his median crowdsource estimate (five-year, $95 million). Moustakas has produced 3 fWAR per 600 PA since 2015 and is projected by Steamer to do about the same in 2018. However, he posted the sixth lowest DRS (-8) among third basemen last season. Offensively, his 2017 power spike—he posted a .249 ISO after posting a .186 ISO in 2015, his previous full-season career high—may or may not be sustainable. On the one hand, he put the ball in the air more often (45.7 FB% vs. a 41.4 FB% in 2015). On the other hand, his HR/FB% spiked (from 11.2% to 17.8%), in spite of the fact that his rate of hard-hit balls barely budged (from 31.5% to 31.9%). Moreover, it appeared to come at the cost of plate discipline, with his BB/K falling from 0.57 to 0.36.


Again, just to be clear, my intention isn’t to be an apologist for owners or to say that these are all bad players. Instead, some of them are players who you can easily imagine value themselves more than even public analytics would suggest, which may be the key to drawn out negotiations.


Now, let’s get to the reason I started writing this post in the first place. When I dug into the deals that have been signed and compared them to expectations, I just didn’t find evidence that players were being signed at discounted rates, relative to recent off-seasons. The table below displays the 25 free agents who, at the time of writing, signed deals. Using Max Reiper’s methodology, I compared the term, average annual value (AAV) and guaranteed money of the actual deals with the median crowdsource estimates determined by Fangraphs. In total, those signed so far have received 4.6% less in guaranteed money compared to the median crowdsource estimate. Over the previous four off-seasons, free agents received 4.2% less than estimated.



This off-season, free agents have received both a bit less term and a bit less AAV than expected. In recent years, free agents received even fewer years than expected, but beat expectations in terms of AAV. The net result in both samples is total guaranteed money falling about 4% short of estimates.


It is worth mentioning that the last four off-seasons were not homogeneous in terms of how deals compared to estimates. In the 13-14 and 14-15 off-seasons, free agents actually received a bit more than the median crowdsource estimate, mainly by beating the AAV estimates. In 15-16, things started to change. While AAV estimates were surpassed by about 7% (just like in 13-14 and 14-15), free agents received significantly fewer years than expected. Last off-season, players saw significantly less term and AAV than estimated, which resulted in free agents receiving only about four-fifths of the guaranteed money the crowd had predicted.



Free agents have, so far, signed deals closer to estimates than they did over the previous two off-seasons. However, the bigger takeaway might be that the last three off-seasons (including this one) have seen players underpaid relative to estimates, whereas they were beating estimates the two off-seasons before that.


On a side note, last year’s big miss really stands out, and it could be argued that potential collusion this off-season started last off-season (or even the one before). One interesting observation is that out of last year’s Top 50 Fangraphs free agents, ten position players were responsible for the majority of the gap between estimates and reality ($202.5M of the total $294M gap). [To be clear, I didn’t just take the ten biggest misses, just ten of the big misses.] Take them out of the equation and free agents missed their total estimate by 8.8%. And, here’s the unfortunate truth: it really looks like front offices were prescient in underpaying those ten free agents. As a group, they produced only 1.5 fWAR over 1172 games. Only two (Edwin Encarnacion and Carlos Gomez) produced 0.5+ fWAR, while seven were below replacement level.



If collusion is indeed happening, it seems to me that one big sign would be players receiving deals that are less than expected. While the Fangraphs median crowdsource estimates are far from perfect, they are a standard benchmark to use in these situations. In total, the deals that have been signed so far are not much different than those estimates. In fact, the 4.6% gap I calculated is smaller than the gaps that existed by the end of the last two off-seasons.


So, while this off-season may be a lot slower than normal, at the end of day, players are still basically getting their money. Ultimately, we need to wait for the off-season to fully play out before we can be sure either way. If the remaining free agents end up taking on heavily discounted deals, then the gap will be much larger than -4.6% and the case for collusion will be much stronger. That said, given the massive offers on the table for Martinez and Hosmer, it’s just as likely that many of the top remaining free agents will beat their estimates—adding Hosmer at $147 million and Martinez at $120 million to the 25 players previously signed this off-season leads the group to beat their cumulative estimates by 3.8%. Perhaps this off-season will end up setting a template for future off-seasons, in which the pace is relatively slow, but the outcomes are still roughly the same.




*Featured Image Credit: kdemerly UNDER CC BY-SA 2.0







I’m an economics professor in the GTA whose lifelong love for the Jays was reignited by that magical August of 2015 and the amazing moments since.

Jeff Quattrociocchi

I'm an economics professor in the GTA whose lifelong love for the Jays was reignited by that magical August of 2015 and the amazing moments since.