Adaptive Markets and Quarterback Contracts
A look at how QB contracts got so large and at potential blueprints for building contending teams with a high percentage of the salary cap devoted to one player
One of my favorite professors at MIT Sloan was Andrew Lo, a legend in the field of finance. Perhaps his greatest contribution to the industry was the Adaptive Markets Hypothesis, which argues that markets are not purely efficient but rather governed in part by the laws of evolutionary biology. The prevailing theory in financial markets even today is the Efficient Markets Hypothesis, which tells us that participants are all rational and acting on the same information, thus resulting in a natural equilibrium when it comes to stock prices for example. This means in the long-run, it is not possible to “beat the market.” Professor Lo, however, argues that market participants model their behavior on their past experiences, constantly learning and adapting based on their successes and mistakes. Evolutionary biology, he argues, can easily explain “irrational” behavior such as fear, overconfidence, loss aversion, and more. Investors, he argues, are more likely to double down on strategies that work out, and change course when they do not, which in turn drives markets. I believe the same theory holds true for football contracts.
Over the past few years, we’ve seen several teams sign their young star Quarterbacks to $50m+ AAV extensions - such as the Kansas City Chiefs with Patrick Mahomes, Cincinnati Bengals with Joe Burrow, Buffalo Bills with Josh Allen, Baltimore Ravens with Lamar Jackson, Los Angeles Chargers with Justin Herbert, and the Cleveland Browns with Deshaun Watson. Given the structures of these deals, however, the large cap hits have not yet come into play. Next season, however, Watson ($64m) and Mahomes’ ($57.4m) cap hits will account for 23.2% and 23.4% of their team’s Salary Cap respectively according to OverTheCap1. These kind of figures are unprecedented, at least in recent history. I believe Professor Lo’s theory does a good job of explaining how we might have gotten here. Over the past 20 years in the NFL, the teams that were most often making deep playoff runs had great Quarterbacks - such as the New England Patriots and Tampa Bay Buccaneers with Tom Brady, Pittsburgh Steelers with Ben Roethlisberger, Indianapolis Colts and Denver Broncos with Peyton Manning, New Orleans Saints with Drew Brees, and Green Bay Packers with Aaron Rodgers. It is therefore natural based on the current NFL environment that decision-makers would come to the conclusion that great QB play is integral to building a winning football team. Furthermore, good QBs are hard to find. They rarely hit Unrestricted Free Agency, and the Draft is a 50-50 proposition at best. Thus it is also natural that Front Offices that had already acquired a Franchise QB would did not want to lose them. The decision to lock up a Franchise QB to a record setting contract can therefore be explained by behavioral science - the humans that make up NFL front offices feared letting someone like Mahomes walk and thus facing the alternative of having to identify and develop an new Franchise Quarterback, and were thus willing to make unprecedented offers to retain their guys. This, in turn, drove up the price of doing so to record high levels. NFL teams are not alone in this kind of thinking. Many sophisticated investors lost a lot of money because they jumped into the housing market before the Financial Crisis - fearing that they would miss out on a bull market and confident that prices would not drop since it had not happened in the decades leading up to the housing bubble.
What teams may have overlooked in the process of locking up their star QBs is the danger of dedicating so much cap space to one player in the ultimate team sport. From 2013-22, the teams that had one of the two QBs in the NFL with the highest cap hits largely did not do well. 13 of these 20 teams (65%) finished with a record of .500 or below. Only one of these teams won the Super Bowl, the 2022 Chiefs, when Mahomes cap hit accounted for 17.0% of the team’s cap - 6.4% less than he’s expected to take up next year. As a whole, these 20 QBs accounted for an average of 15.6% of the Salary Cap for their respective teams. A common thread between the teams that spent a lot on their starting QB and proceeded to have a losing season was poor play from their highly paid Quarterback. Prime examples include the 2018 San Francisco 49ers and 2019 Detroit Lions, who saw Jimmy Garoppolo and Matthew Stafford miss significant time to injuries respectively. Another example is Eli Manning and the 2013 New York Giants, when the Giants made Manning the highest paid QB in the league that year, only for him to lead the NFL in interceptions. In a league where one dollar of cap space that goes to one player cannot go to another, an underperforming or injured player with a large guaranteed cap hit puts a major strain on team performance that is uniquely difficult to alleviate relative to other professional sports. Even in the NHL, the only other of the big-4 American sports leagues with a hard cap, teams can get some cap relief by placing injured players on Long Term Injured Reserve.
While good Quarterback play is important to building a contender, the data shows that Super Bowl teams do not tend to be among the biggest spenders at the position. From 2013-22, I found that the starting QBs on the teams playing in the Super Bowl accounted for an average of 7.9% of their team’s cap. This number is brought down by Quarterbacks on their rookie contracts, but even the QBs on veteran deals only accounted for an average of 11.3% of the cap - less than half what Mahomes will likely take up in 2024. Prior to Mahomes, no QB playing in the Super Bowl accounted for more than 15.0% of the cap in the time period I studied - with that player being the Atlanta Falcons’ Matt Ryan in 2016. From a football standpoint, this makes logical sense as teams need much more than just a good QB to play in the Super Bowl and thus have to spread that money around to acquire good players at other positions.
In 2024, the three highest paid QBs are expected to be Watson, the Dallas Cowboys’ Dak Prescott, and Mahomes. Watson’s underperformance to date makes his contract probably the worst in the NFL, and one that the Browns will have a hard time building around. The Cowboys can probably get some immediate cap relief by signing Prescott to an extension. Mahomes performance has declined this season, likely in part due to the lower quality team the Chiefs are fielding as a result of the constraints created by his deal, but given that they and so many other teams already made the decision to sign their Franchise QB, I sought to find a blueprint for how these teams can succeed in the future despite dedicating so much cap space to one player. To do so, I looked at four teams that made deep playoff runs while also having among the highest amount of cap space dedicated to their Starting QB that season. Those teams were the 2016 Steelers, 2016 Falcons, 2019 Packers, and the 2022 Chiefs.
Ben Roethlisberger played well on the 2016 Steelers while accounting for 15.3% of the Steelers’ salary cap, but the team also got great performance from players still on their rookie deals. Despite only playing 12 games, Running Back Le’Veon Bell provided the team with 1,884 yards from scrimmage while accounting for just 0.7% of the Cap. On defense, young players like Defensive End Stephon Tuitt, Defensive Tackle Javon Hargrave, and Linebacker Ryan Shazier were productive starters on rookie deals, with Shazier making the Pro Bowl. Many of the other players that the Steelers chose to pay highly, such as Wide Receiver Antonio Brown, played at a level consummate with their cap hit. Mahomes’ 2024 cap hit accounts for 8.1% more of the cap than Roethlisberger’s did in 2016, and for context, Bell and Brown’s combined cap hits took up 8.3% of the Salary Cap. Mahomes is of course, the far superior player compared to Roethlisberger but this just goes to show the kind of talent a team can get with ~8% of their cap. The Steelers would have found other players besides Bell and Brown to cut if they had to accommodate a cap hit like Mahomes, but it is still likely that they would have to have made painful choices in releasing players to be cap compliant.
In 2016, Matt Ryan won the MVP and Offensive Player of the Year award en route to leading the Falcons to the Super Bowl. That season, he accounted for 15% of the team’s cap. The team also handsomely paid WR Julio Jones (10.1% of the cap), but were rewarded as he led the league in receiving yards per game. But the Falcons also got a lot of contributions from players on affordable deals. For example, their entire starting Offensive Line combined to account for 12.1% of the cap due to Tackles Jake Matthews and Ryan Schraeder being on rookie deals, Guards Andy Levitre and Chris Chester having affordable veteran deals, and star Center Alex Mack being in the first year of his relatively lucrative Free Agent contract when his cap hit was the lowest. Other key players on rookie deals on that team included Pro Bowl Running Back Devonta Freeman and key starters on defense such as DT Grady Jarrett and LBs Vic Beasley and Deion Jones. Ryan’s cap hit in 2016 accounted for 8.4% less of the cap than Mahomes is projected to in 2024, and as evidenced by all the accolades he received that season, it is hard for anyone to play much better than Ryan did in 2016. For context, Matthews, Schraeder, Chester, and Mack accounted for 8.7% of the cap that season, so paying Mahomes (or any QB) nearly a quarter of the salary cap will likely come at a great cost to the quality of the rest of the team.
On top of a Pro Bowl season from Aaron Rodgers, who accounted for 14.9% of the cap, the Packers benefitted from productive seasons from several players on rookie contracts. Some examples of such players include DT Kenny Clark, RB Aaron Jones, CB Jaire Alexander, and LB Blake Martinez. The team also signed productive players like Edge Rusher Za'Darius Smith, LB Preston Smith, and Guard Billy Turner in the off-season, and all three were in the first years of their backloaded deals and thus had reasonable cap hits that season - each accounting for less than 4% of the Packers’ cap. Lastly, LT David Bakhtiari and WR Davante Adams only accounted for 7.2% and 5.5% of the cap respectively, highly reasonable figures for players of their caliber at their positions. Had Rodgers’ cap hit been closer to that of Mahomes in 2024, it is possible the Packers could not have afforded to sign the Smiths and Turner, who all combined to account for 8.9% of the cap - only slightly more than the 8.5% difference in 2019 Rodgers’ and 2024 Mahomes’ cap hits.
The 2022 Chiefs won the Super Bowl with Mahomes accounting for 17.0% of the cap, a year in which he won the MVP award. But Mahomes did not win the Super Bowl alone. DT Chris Jones also carried a significant cap hit, but finished 3rd in Defensive Player of the Year voting. The team also had starting LT Orlando Brown Jr on the Franchise Tag, making his a bargain for a high-end starter at that position. The Chiefs also got good value from WR Juju Smith-Schuster and TE Travis Kelce, who combined to rack up 2,271 receiving yards while only accounting for 5.8% of the cap - roughly half of what many top veteran WRs cost. Again, like with Ryan in 2016, it is hard to have a better season than Mahomes did in 2022. Even if he plays at that level in 2024, which is no guarantee, putting together the same caliber of a team is going to be much more difficult when he accounts for 6.4% more of the cap than in 2022.
Interestingly, I did find in my research that Peyton Manning accounted for 20.5% of the Indianapolis Colts’ cap in 2003, the highest figure by far in his career. WR Marvin Harrison also took up 10.3% of the cap that year. I was not able to find robust cap data for most other players from 2003, but I did find that the Colts also had Hall of Fame caliber players like RB Edgerrin James, WR Reggie Wayne, and Edge Rusher Dwight Freeney on rookie contracts - which helps explain how they were able to allocate a combined 30.8% of the cap to Manning and Harrison. Manning won the MVP award that year and Harrison finished 6th in the NFL in receiving yards en route to another Pro Bowl appearance, making them worth their contracts. However, having three current or likely future Hall of Famers on rookie contracts at the same time is a low probability event that no team should expect to happen in the future.
What all these teams have in common is that they got outsized production from several key players relative to their cap hits - whether it was because they were on a rookie contract, bargain free agent deal, or early on in a backloaded contract. Furthermore, in the case of teams like the 2016 Falcons and 2022 Chiefs, other players that were on lucrative contracts played at an All-Pro level. The 2024 Chiefs will have less leeway than the other teams I looked at due to Mahomes’ cap hit being so high. Further complicating matters is that Right Tackle Jawaan Taylor is projected to take up 10.1% of the cap next season (with no easy way of alleviating the cap hit) and has been one of the worst players at his position in the NFL this season. On top of Mahomes returning to form, the 2024 Chiefs will have to have a lot of players producing at a level higher than their cap hit. Some of this could happen through continued growth from young players already on their roster like Edge Rusher George Karlaftis III, CB Trent McDuffie, RB Isiah Pacheco, and WR Rashee Rice. The Chiefs can also structure contracts in Free Agency to have low cap hits in Year 1, and push the high cap hits to later when they can get more relief from a rising cap and getting out of contracts like Taylor’s. Trades are another way for the team to find more impact players on affordable contracts. Through a combination of trading down in the first couple rounds of the draft to get more picks and dealing selections in the 3rd-5th rounds to rebuilding teams for players in their 3rd or 4th years of their rookie deals, the Chiefs can maximize their chances of fielding a roster rich with players outperforming their cap hits.
The purpose of this analysis is not to pick on the Chiefs, as they will not be the only team that has to face the challenge of winning with so much money tied up to one player in the future. Furthermore, it is possible that teams expected the cap to rise more rapidly when they signed their QBs to those large deals. Jason Fitzgerald at OverTheCap, for example, had expected the Salary Cap to be $256m next year until an NFL memo said that it would likely be closer to $242m. Still, even if the cap were to have been $256m, Mahomes would have accounted for ~22.4% of it for the Chiefs. It is also possible that the extensions will work out, and that allocating 20%+ of the cap to a Starting QB becomes the norm for winning teams, particularly if the price of players at other positions drop as a result. Given that football is a team sport, however, simple salary cap math tells us that the teams built around these highly paid QBs in the future will likely be weaker than those in the past like the 2016 Falcons and 2022 Chiefs. Mahomes is already likely a lock for Canton, but as we can see from looking at other teams, his deal will require the Chiefs to forgo signing or cutting some good players to bridge the difference between his cap allocation and that of almost all other highly paid QBs in the past. We are, in fact, already seeing the constraints that these large deals put on roster construction with the 2023 Chiefs. The only way teams can build as good a roster while carrying a player with such a high cap hit is to get significantly better at drafting and finding values in Free Agency than before, which is a very tall order. My prediction is that moving forward, top QB salaries will stagnate for a while and the market will again adapt such that they will fall somewhere between 12-17% again.
I believe that understanding the Adaptive Market Hypothesis can help NFL Front Offices build better football teams. Professor Lo argues that just being aware that emotion plays a role in decision-making can help us better understand financial markets. An application to football would be realizing that fear of the consequences of letting a top-QB walk could be inflating the price of these contracts beyond what the data - both quantitative and qualitative - suggests they should be. Much like the few traders who successfully saw that emotion rather than empirical data was leading to the housing crisis were able to profit, teams able to identify such situations might be better off in the long-run even if they face some short-term losses in the process. Michael Burry of The Big Short fame was considered crazy when he made his bet against the housing markets, especially when prices continued to rise. In the long-run, however, Burry and the investors in his Hedge Fund were better off than those who invested at or near the peak of the housing bubble. In football, a team that traded away a star QB for a massive package of picks and young players rather than committing nearly a quarter of the salary cap to that player might have seemed crazy in the moment. However, it is possible (though far from certain) that they would have been better off in the long-run, assuming they had the patience to stay the course and not panic even if short-term returns were not optimal.
In his interview with the CFA Institute (linked above), Professor Lo also talked about how technology could help mitigate and prevent mistakes caused by human emotion. He gave the example in a finance context of a risk management tool that could automatically manage portfolios during times of volatility instead of a human. In such an instance, technology is not taking over for humans but rather filling in where evolutionary biology may have created a weakness. In football, I believe that the tools that the team over at SumerSports are building will be of great use to the industry. To be clear, there is a huge human component that goes into building a cohesive team, and things like character assessment are integral parts of the process. I do not believe that team building should be outsourced to a computer, but technology can help decision-makers in instances where human emotion may create blind spots. Sumer’s technology can obviously help teams objectively value players but its more important use could be in idea generation through helping teams identify multiple ways to build winning football teams that are backed by empirical data. A General Manager simply knowing a few other paths to follow in roster construction should bidding for a certain player go too high could prevent them from panicking and overpaying, setting the team back in the long-run in a league where every dollar spent matters. Rather, they can chart forward on different path, guided by technology and data, that could make their team better.
Note: Teams can carry over cap space from one year to another so the Browns will likely have a higher adjusted cap than the Chiefs next year, which is why Watson is projected to account for less of the cap than Mahomes despite having a higher cap hit