Saturday night, Round 1, the Pittsburgh Steelers are on a video call with the wide receiver Makai Lemon. They hold pick 21. They are telling him, in front of cameras, that they intend to take him. Lemon, unflappable, accepts the news. Then his phone buzzes with another incoming call. The caller ID says Philadelphia. Why is Philly calling me? he asks, into a room of Pittsburgh personnel who are still on the line. The answer comes seconds later: the Eagles have traded up with the Cowboys to pick 20 and have selected him, one slot ahead of where Pittsburgh wanted him to be selected. Howie Roseman has, on live television, sniped a wide receiver out of a phone call.
There is an additional, slightly more painful detail. At the moment Pittsburgh’s general manager Omar Khan was on the line with Lemon, Pittsburgh was not on the clock. The Cowboys were. Khan called Lemon on the assumption that the Cowboys would not be taking a wide receiver, an assumption that was correct in spirit and catastrophically incorrect in fact, because the Cowboys had quietly traded the pick to the Eagles, who very much would. The NFL has now confirmed that it will review the call as part of its post-draft audit. There is a specific league rule that prohibits non-clock teams from interfering with teams that are on the clock, and former Eagles executive Jake Rosenberg flagged on social media that the Steelers may have run afoul of it. The consensus among observers, mostly arrived at within eight seconds of the trade being announced, is some variant of: never call a player before you are on the clock. A first-year law student would know better.
I want to argue that this is, in fact, exactly what a first-year law student would know better — and that the formal version of what they would know better is the actual subject of this post.
The standard sports-economics story about the National Football League draft begins, more or less without exception, with Cade Massey and Richard Thaler’s 2013 Management Science paper, “The Loser’s Curse.” Their finding — call it the canonical version of the result — is that NFL general managers systematically overvalue early picks. They show this by computing the “surplus value” each pick generates: roughly, the value of the player’s on-field production minus what the team is contractually required to pay him. By that measure, surplus is highest in the early second round, not at the top of the first, because top picks get paid like top picks but produce more like high-variance lottery tickets. The implication, drawn straight from Thaler’s behavioral playbook, is that GMs trade up because of overconfidence, anchoring on combine numbers, the availability of pre-draft narratives, and the rest of the warm-blooded portfolio of failures of expected-value rationality.
A recent paper by Ryan Brill and Abraham Wyner (arXiv:2411.10400) sharpens the question without quite refuting the answer. Their argument is that the loser’s curse depends entirely on assuming that surplus value is the right utility function for a general manager, and that this is a substantive assumption rather than an obvious one. A general manager building a roster around an elite quarterback is not in the same decision problem as one swapping interchangeable parts; right-tail outcomes can rationally weigh more heavily than expected-value calculations admit; positional scarcity does not reduce to a regression coefficient. The behavioral story, in their telling, is what you see when you assume away the production technology.
Both papers, however, treat each general manager’s decision as a separate single-bidder problem. Each team computes a value, possibly with the wrong utility function, and trades against a market price determined by aggregate demand. What is missing from this frame is the entire reason the draft exists in the first place: the league.
The proper formal home for what happened to the Steelers on Saturday night is a 1996 paper by Philippe Jehiel, Benny Moldovanu, and Ennio Stacchetti called, with the kind of title that makes economics worth doing, “How (Not) to Sell Nuclear Weapons.” The paper studies auctions in which a bidder who does not win the object suffers an externality whose size depends on the identity of the bidder who does. Selling a nuclear weapon to North Korea hurts every other state in the world; selling it to Belgium hurts no one in particular; the seller’s optimal mechanism therefore depends not just on bids but on the entire matrix of cross-buyer payoffs. Jehiel, Moldovanu, and Stacchetti establish three results that I will list in the order in which the Pittsburgh Steelers violated them.
First: outside options and participation constraints are endogenous. What a buyer is willing to pay depends on what happens to the object if they don’t acquire it, which depends on which other buyer would acquire it instead. There is no exogenous reservation price; reservation prices are equilibrium objects, derived from the entire game.
Second: the seller can extract surplus from buyers who do not acquire the object, by leveraging the externality. In the nuclear-weapon framing, you can collect protection money from the United States and the European Union by credibly threatening to sell to a less savory state. In the NFL framing, you can extract trade compensation from a team that wants a player by credibly threatening to draft him yourself, even if you don’t really want him.
Third, and most striking: when externalities are large relative to use-value, the seller is sometimes better off not selling at all, while still collecting payments. The right policy for nuclear weapons is sometimes to keep them and accept tribute. The right policy for a draft pick is sometimes to sit on it. (No, this is not a post about Iran and Trump, but you can read it how you like.)
Each of these results has a folk-wisdom translation that every general manager in the NFL knows by heart, even those who could not pronounce “Stacchetti” without injury. The single most durable rule of NFL roster management is “(almost) never trade within division.” The reason is precisely the negative externality: a draft pick traded down to a divisional rival lands inside an opponent you will play twice a year, against whom every additional win for them is a probable loss for you. The trade compensation you receive is a payment denominated in the same accounting unit as the harm the trade inflicts, and the harm generally dominates. There is no behavioral story here. There is no overconfidence, no anchoring, no failure of expected-value reasoning. There is just an externality term that surplus-value calculations cannot see, and that NFL front offices treat with the gravity of canon law because the externality bites them every September.
Pittsburgh and Philadelphia are not divisional rivals. They are, however, the two NFL franchises located within the borders of Pennsylvania, which makes them rivals in the regional sense, the fan-dollar sense, the media-narrative sense, and — this is the honest one — the schadenfreude sense. The cross-team negative externality term that determines the Eagles’ real value of Lemon includes the on-field cost of the next Pittsburgh-Philadelphia matchup whenever the schedule produces one, the regional bragging-rights cost of being the lesser PA team, and the very large and very real Eagles-fan utility surplus generated by acquiring a player Pittsburgh wanted in front of Pittsburgh on national television. Roseman captured all three at the cost of moving up one slot, which on any standard chart is a price near zero.
Which brings me back to the first-year law student. The auction-theoretic name for what Pittsburgh did on the phone with Lemon is information leakage. Calling a prospect before you are on the clock — broadcast, on the league’s own television network, to every other team in the league including the team currently picking — reveals your reservation price to all other bidders simultaneously. In an auction without externalities, this is merely a tactical error; in an auction with externalities, it is a structural blunder. Once Pittsburgh’s preference for Lemon was common knowledge, every team that valued denying Pittsburgh that asset — for any reason, divisional or not — had a clean opportunity to either preempt by trading up, or extract trade compensation from Pittsburgh by threatening to take Lemon themselves. Philadelphia chose preemption. The price was one slot.
The reason the NFL has a specific rule against non-clock teams calling prospects is not that the league is squeamish about awkward television. It is that the league has, somewhere in its institutional memory, the recognition that the draft is an auction with allocative externalities, and that an auction with allocative externalities does not function if reservation prices are public. The rule exists because the structural blunder is bad enough that the league wants to prevent it from happening even when individual general managers fail to. The Steelers’ mistake was not tactical. It was a category error. They treated the draft as a parallel set of single-agent decision problems — Massey-Thaler’s frame, exactly — when it was a strategic-form game with payoff interdependence and information leakage. The first-year law student knows not to reveal the reservation. The auction-theory student knows the reservation only matters because there are other bidders whose payoff depends on whether you get the asset.
One more observation, because the post would not be complete without it. The cost side of “surplus value” — what teams pay drafted rookies — is set by a wage scale that emerged from the 2011 collective bargaining agreement. Slot values are essentially fixed. This is sometimes described as a market price for rookie labor. It is not. It is a settlement payment in a long-running antitrust dispute about whether the draft is itself a legal restraint of trade. Players accept the scale because the alternative is no guaranteed early-career money for late-round picks; owners accept the scale because the alternative is the antitrust litigation the Players Association periodically rattles. Both sides are settling a dispute about whether the institution should exist, denominated in dollars that flow only because the institution does exist. The “compensation” term in any surplus-value calculation is therefore endogenous to the legal-political bargain that produced the draft. Treating it as a market price is the move that quietly extracts all the political economy from the analysis.
Which brings me, finally, to the chart. The Jimmy Johnson draft value chart, which assigns each pick a numerical value that lets teams compute trade equivalences in real time, has been demonstrated to be wrong in roughly the way Massey and Thaler describe. It overweights early picks, underweights late ones, and bears no particular relationship to either surplus value or the externality-adjusted values that Jehiel and Moldovanu would tell you actually matter. This is, in the analytics commentariat, treated as a problem with the chart. It is not a problem with the chart. It is what the chart is for.
The chart’s job is not to price picks accurately. Its job is to make every trade defensible in real time at a speed that allows three days of televised content to function. A chart that produced visibly correct prices would invite every team to relitigate every trade, and a chart whose authority depended on demonstrable accuracy would lose authority the first time it was clearly beaten. The Jimmy Johnson chart works because nobody believes it is accurate and everybody behaves as though they do — the Schelling-point function of “the chart we all use” matters more than the price-discovery function it nominally serves. This is the same reason a constitution persists despite being known to be incomplete: its function is coordination, not accuracy. The losers’ curse, in the relevant sense, is being right loudly enough to break the device that lets everyone agree.
So: did the Eagles overpay to move up one slot for Makai Lemon? On the Massey-Thaler ledger, somewhat. On the Brill-Wyner ledger, possibly not, depending on what the Eagles thought their wide receiver room needed. On the Jehiel-Moldovanu-Stacchetti ledger, almost certainly not, because the externality term they captured by acquiring Lemon in front of Pittsburgh on national television is not visible on either of the other ledgers and is, in this case, plausibly larger than the trade cost. There is a settlement, an institution, a chart whose function is to make the settlement liquid, and a moment of theater that transferred a player from one organization to another at a cost both organizations agreed to pretend was the right cost. The Steelers were on the phone. The Eagles were on the chart. The chart was on television. Everyone was bound by the same rules — which is the only thing that made any of it worth watching, except for the part where one team broke one of them.
I lived in Pittsburgh once, Maggie and I got married at the Mattress Factory, and we still kind of consider it one of our “true homes.” Even after this weekend, I am still a Steelers fan. I will spend the rest of this week irritated about Howie Roseman, irritated about Omar Khan, and disproportionately interested in the NFL’s procedural review, and I will watch Round 7 of next year’s draft with the same intensity, because the institution depends on me doing all of it. (Ed. Oh, I do love it when you get grandiloquent…)
With that, I leave you with this.