How the MLB luxury tax works, and why there is no cap

A contract gets announced in December. The number in the headline has nine figures in it, the number is repeated everywhere for a fortnight, and it is not the number that matters. The one that matters is smaller, it is produced by a formula written into the collective bargaining agreement, and it is what the MLB luxury tax is actually charged on. Getting from the first number to the second is the whole of this article.

Baseball is the only one of the major North American leagues that never agreed to a salary cap. Not a hard ceiling like the NFL's, not a soft one buried under exceptions like the NBA's. There is no limit on what a club may spend, no compliance date in March, no sanction for carrying the largest payroll in the sport by a wide margin. There is a tax. It has a threshold, it has rates, and an owner who wants to can pay it every year until he dies.

Clubs nevertheless behave, most of the time, as though the threshold were a limit. That gap between what the rule permits and what thirty front offices actually do is the interesting part of the system, and the tax rates do not explain it. Three other things do: how the taxable payroll is calculated, what happens to draft picks and international signing money, and what the other half of the agreement, revenue sharing, does to the incentives underneath everything.

Baseball chose a tax instead of a cap, and it chose the hard way

Owners wanted a cap. They wanted one badly enough that the dispute over it cost the sport the 1994 World Series, which had survived two world wars and was cancelled by a labour stoppage. The union did not blink, and no agreement since has contained a payroll ceiling.

The union's objection was never that a cap is unfair in the abstract. It was structural. Under a cap, salaries stop being set by clubs bidding against each other and start being set by a formula, with the players' share fixed in advance and the only remaining argument being about how to divide a pot whose size is already decided. Baseball players have free agency, arbitration and no ceiling, and the union's position has been that trading the third of those for cost certainty is a permanent concession sold for a temporary peace.

What emerged instead was a payroll tax, and the reason the players could accept one is that a tax and a cap are not the same concession at all. A cap binds every club. A tax binds only the handful that go past the line, and it binds them with a bill rather than a prohibition. The compromise let owners tell each other they had put a brake on the top of the market while letting the union say it had conceded no ceiling. Both were telling the truth.

The contrast with the NFL's hard cap is the cleanest way to see what each side bought. American football has a genuine ceiling and, underneath it, contracts that are mostly not guaranteed, so clubs can walk away from almost any deal at a known price. Baseball has no ceiling and contracts that are guaranteed to the last dollar. A club that signs a bad deal in baseball owns it, in cash and against the tax, for every year it runs. Neither system is more generous than the other. They distribute risk to opposite parties.

What the MLB luxury tax is actually levied on

Start with the thing almost every report gets wrong. The competitive balance tax is not charged on what a club pays out in a season. Cash payroll and tax payroll are two different figures and, for clubs that structure contracts cleverly, they can differ by a very large amount in either direction.

Tax payroll is built from four components.

The average annual value of every contract belonging to a player on the 40-man roster. Not the salary that player will collect this season. The average value of his whole deal, computed once at signing and then carried unchanged for its life. This is the part that does the real work, and the next section is about nothing else.

The club's allocated share of player benefit costs. Pensions, health insurance, the medical and welfare programmes the agreement obliges the industry to fund. Every club carries a share of this and it runs into eight figures. It is not payroll in any ordinary sense, and it counts.

The club's contribution to the pre-arbitration bonus pool. The current agreement created a fund distributed to young players who are not yet eligible for salary arbitration, chosen by performance and awards, as a partial answer to the complaint that the best players in the sport are frequently among the cheapest. Every club pays into it, and that payment counts towards its tax payroll too.

Other salary obligations. Money still owed to players the club has released. Buyouts on options it declined. Cash sent to another club in a trade. Anything the club is contractually on the hook for that has not already been captured in the first component.

Two consequences of that list are worth sitting with, because they change how the threshold should be read.

The first is that a meaningful block of the threshold is spoken for before a club signs a single player. Benefits and the bonus pool are a fixed overhead that every one of the thirty carries identically. When a club is described as having a certain amount of space beneath the line, the honest version of that figure is the threshold minus that fixed block minus committed average annual values, and the fixed block is large enough that ignoring it produces a materially wrong answer.

The second is that a club cannot cut its way out. Baseball contracts are guaranteed, so releasing a player extinguishes nothing: the money is still owed, and it still counts against the tax at the same average annual value it always did. This is the sharpest practical difference between baseball and the leagues where contracts are only partly guaranteed. In the NFL, a release converts a future obligation into a smaller present one. In baseball, a release converts a player into an absence and changes the accounting not at all.

• 40Players whose contracts count towards tax payroll
• 3Steps the rate climbs for consecutive years over the line
• 3Surcharge thresholds sitting above the base threshold
• 1Seasons under the line needed to reset the rate

Rule-defined counts rather than dollar figures. The threshold and the rates are published annually by the league.

Average annual value is the entire system in one phrase

Here is the rule that everything else hangs from. A player's charge against the tax is the total guaranteed value of his contract, discounted for anything deferred, divided by the number of guaranteed years. Not his salary this season. The average.

Say that a club signs a player to five years and $100 million, of which $40 million is a signing bonus handed over on the day of the announcement and the rest is salary of $5 million, $10 million, $15 million, $15 million and $15 million. In year one the club writes cheques for $45 million. In year five it writes cheques for $15 million. Against the tax it is charged $20 million every single year, because that is what $100 million divided by five is.

• Cash paid
• Tax charge

Year 145m20m

Year 210m20m

Year 315m20m

Year 415m20m

Year 515m20m

A constructed contract, not a real one. $100m over five years, of which $40m is a signing bonus paid on signature. The tax charge is identical every year whatever the cash does.

Show the numbers

Worked example: cash paid against tax charge, five-year deal

Item	Cash paid	Tax charge
Year 1	45m	20m
Year 2	10m	20m
Year 3	15m	20m
Year 4	15m	20m
Year 5	15m	20m

The reason for the rule is anti-circumvention, and it is obvious the moment you imagine the alternative. If the tax were charged on cash, every contract in the sport would be backloaded to the point of absurdity. A club sitting just under the line would sign a star to five years at $1 million, $1 million, $1 million, $1 million and $96 million, sail under the threshold for four seasons, and then trade him, restructure, or simply accept one enormous bill in a season it had already written off. Average annual value kills that idea in a single clause. Move the money wherever you like inside the deal. The charge does not move.

What the rule creates instead is a different lever, and it is the one that shapes the whole shape of baseball contracts: length. If the charge is total divided by years, then adding a year lowers the charge in every year. A $200 million commitment over eight years charges $25 million annually. The same $200 million over ten years charges $20 million. The club has not saved a dollar in cash, it has taken on two extra seasons of risk on a player who will be older in both of them, and it has bought itself $5 million of annual room under the threshold.

This is why baseball produces contract lengths that look insane from the outside and are perfectly rational from the inside. A club is not predicting that a player will still be good in year ten. It is buying tax relief in years one through eight and pricing the tail as the cost of that relief. Sometimes it is right. Sometimes the last three years are an anchor nobody can cut loose, because nothing can be cut loose.

Average annual value cuts the other way too. A club having a bad season cannot get relief by paying a player less, because his charge was fixed years ago. A player on the injured list for the entire year is charged in full. Money owed to somebody released in spring training is charged in full. There is no injured reserve mechanism, no stashing, no way to make a signed obligation quieter than it is.

The exceptions are handled by who holds the decision. A club option is not guaranteed money until it is exercised, so it sits outside the calculation while the buyout attached to it sits inside; exercise the option and the contract is recalculated with the new year included. A player option, being the player's to take, is generally treated as a guaranteed year from the start. Performance bonuses are outside the average annual value entirely, and are added to the club's payroll in the year they are actually earned, which is how a club can quietly finish a few million above a line it thought it was under.

Arbitration and pre-arbitration players make the arithmetic trivial and the strategy enormous. Those players sign one-year contracts, so the average annual value is simply the salary, and for pre-arbitration players the salary is near the league minimum. A club whose best five players are all in their first three seasons is carrying an outstanding roster for a fraction of a threshold. This is the real reason front offices are so attached to what wins above replacement actually measures: it gives them a common denominator, production per dollar of average annual value, and by that denominator a good pre-arbitration player is worth several times any free agent on the market.

From a signature to a tax bill, step by step

1. 1Add up the guaranteed moneyEvery guaranteed dollar in the deal: signing bonus, guaranteed salaries, and any buyout attached to an option the club controls. Money that depends on performance is not in here.
2. 2Discount anything deferredPayments scheduled for after the contract ends are converted into what they are worth today, using the discount rate the agreement specifies. The further out a payment sits, the less of it survives.
3. 3Divide by the guaranteed yearsPresent value divided by the number of guaranteed seasons gives the average annual value. That figure, and not this season's salary, is the player's charge for every year of the deal.
4. 4Adjust for the calendarA player traded during the season splits his charge between the two clubs according to days on each roster. Cash attached to the trade moves part of the charge with it.
5. 5Add everybody else on the 40-manArbitration and pre-arbitration players are on one-year deals, so their charge is simply their salary. Players on the injured list count in full. So does money still owed to players who have been released.
6. 6Add the fixed blockThe club's allocated share of player benefit costs, and its contribution to the pre-arbitration bonus pool, go on top. Every club carries this before it signs anyone.
7. 7Compare against the thresholdThe total is the club's tax payroll. Only the amount above the threshold is taxed, at a rate determined by how many consecutive years the club has been over and how far past the line it has gone.

The same sequence runs for every player on the 40-man roster. The club's tax payroll is the sum of the results plus the fixed block.

The league runs that calculation after the season ends, on final figures, and bills in the winter. There is no in-season compliance test and no moment at which a club is out of order. A team can be well over the line in July, trade two contracts away in August, and finish under. The only date that counts is the end of the season.

Deferred money, and the discount that shrinks a headline

Now the part that turns a large contract into a smaller one on paper.

A club and a player can agree that some of the money will be paid long after the contract is over. Ten years after, twenty years after, in equal instalments stretching into a decade the player will spend retired. The club still owes every dollar, and the league does not permit the obligation to be a bare promise: deferred compensation carries funding requirements, so the money is set aside rather than left to the good intentions of a future owner.

For tax purposes, though, a dollar payable in twenty-five years is not a dollar. The agreement requires deferred payments to be discounted back to present value, using a rate it defines by reference to a published federal interest rate. The calculation is ordinary time-value-of-money arithmetic, and its effect on a heavily deferred contract is dramatic.

Take a constructed example. A club signs a player for ten years and $300 million. Two hundred million is paid during the contract in the normal way. The remaining $100 million is deferred, paid out in ten equal instalments of $10 million a year, beginning ten years after the deal ends. Discount that at a flat 5 per cent, chosen here purely so the numbers are followable, and the deferred hundred million is worth about $29 million today.

The tax therefore sees a contract worth roughly $229 million, not $300 million. Divided by ten guaranteed years, the charge is about $22.9 million a year rather than $30 million. The club has bought a headline it can put on a billboard and a tax number seven million a year lighter.

Headline value of the contract300m

Value the tax counts229m

Deferred money as it will be paid100m

Deferred money after discounting29m

A constructed ten-year, $300m contract. $100m of it is deferred into ten equal instalments beginning ten years after the deal ends, discounted here at a flat 5 per cent for legibility. The rate the agreement actually uses is tied to a published federal rate and moves with it.

Show the numbers

Worked example: what deferring a third of a contract does to the tax number

Item	Value
Headline value of the contract	300m
Value the tax counts	229m
Deferred money as it will be paid	100m
Deferred money after discounting	29m

Three things follow, and the third is the one almost nobody says out loud.

Deferral is not a trick played on the union, because the player agrees to it and is usually compensated for agreeing. He can take a larger nominal figure than the club would otherwise pay, he can spread income across tax years and jurisdictions, and if he believes he can invest less well than the club can, taking money later at a favourable nominal premium is a reasonable deal. Players with enough bargaining power to refuse deferrals outright generally get that too.

Deferral is also not free to the club. The money is real, the funding obligations are real, and a club that defers heavily across several contracts at once is committing future ownership to payments long after the players have gone. There is a version of this that ages badly.

And the value of the whole manoeuvre depends entirely on the interest rate environment, which is why deferrals come in and out of fashion for reasons nothing to do with baseball. When the discount rate is high, money twenty years out is worth very little today, and deferring is enormously effective. When rates sit near zero, the discount barely bites, and deferring $100 million saves the club almost nothing against the threshold. The single most aggressive contract structure in the sport is, underneath, a bet on the bond market.

The rate climbs for every consecutive year a club stays over

The tax is charged on the excess, not the whole payroll. A club that finishes ten million above the line pays a percentage of ten million. This gets misreported constantly, usually in a direction that makes the bill sound catastrophic, and it is worth being precise about because the marginal structure is what drives the behaviour.

The rate itself depends on repetition. A club over the threshold for the first time pays the lowest rate. Over for a second consecutive year, it pays a higher one. From a third consecutive year onwards, a higher one still, and it stops climbing there. The steps are large rather than cosmetic, and by the third year the marginal cost of payroll is high enough to change what a general manager is willing to add in July.

Then comes the rule that shapes half the transactions you see in a given winter. The counter resets after a single season under the line. Not two seasons, not a probationary period. Finish one year underneath and the next time the club crosses it is a first-time payer again, at the lowest rate.

That single clause explains a great deal of otherwise baffling behaviour. A club with an expensive roster and a squad it does not believe in will deliberately duck under for one year, shedding salary at the trade deadline, declining to replace departures, and writing off a season it was not going to win anyway. It is not rebuilding. It is resetting the clock, so that the following winter it can spend hard as a first-time payer instead of a third-time one. The saving is not the tax bill on the year it skipped. The saving is the lower rate on every dollar it intends to spend for the next three years.

This is also why the tax is a poor tool for restraining the very richest clubs and a decent one for restraining everybody else. An owner prepared to sit in the top rate indefinitely has pulled the escalator's teeth simply by refusing to step off it. An owner not prepared to do that has to plan his crossings, which means planning his competitive windows around a tax calendar rather than around his roster.

Surcharge tiers stack on top, and one of them costs a draft pick

Above the base threshold the agreement sets further thresholds, three of them under the current deal, each a fixed distance above the last. Cross one and a surcharge is added to the rate charged on the money above that tier.

The two dimensions compound. The consecutive-year rate and the surcharge tiers are separate escalators applied to the same excess, so a club in its third straight year over, sitting deep in the upper tiers, faces a marginal rate on its last slice of payroll that is a serious multiple of what a first-time payer a few million over the base line faces. At the top end, adding a player costs meaningfully more than the player costs.

The surcharge tiers also carry the first non-financial penalty in the system. Go far enough past the line, at the tier the agreement designates, and the club's highest selection in the following year's draft is moved back ten places. There is a protection for clubs whose pick sits near the very top of the order, on the sensible ground that a rule intended to discipline heavy spenders should not fall hardest on a club that has just finished last; in that case a later selection moves instead.

Ten places does not sound like much. It is much.

The penalties that actually deter clubs are not the money

For a club at the top of the sport's revenue table, a tax bill is a cost, and costs are what large businesses are for. Set against a deep run in October, the extra broadcast inventory, the ticket revenue, the sponsorship renewals and the effect on the value of the franchise itself, several million dollars of tax on the last few players is not obviously a bad trade. The expansion of the postseason field has, if anything, made it a better one, because it raised the value of the marginal win for every club hovering around the last qualifying place.

What such an owner cannot buy at any price is amateur talent, and that is where the real deterrent sits.

Draft picks are rationed by rule and only weakly tradeable. Losing ten places in the first round is not losing a lottery ticket; it is losing the difference between the group of players a club's scouts have ranked in the top fifteen and the group they have ranked below it, in a draft where the gap between those two groups is the entire point of employing scouts. It costs money in a second way as well, because each pick carries an assigned bonus value and a club's total draft spending pool is the sum of the values of the picks it holds. Move a pick back and the pool shrinks, which constrains what the club can do with every other selection it owns.

International amateur signings work the same way and matter more than people outside the sport realise. Players from outside the draft's territory, largely teenagers from Latin America, are signed out of a bonus pool whose size is fixed by rule, with only limited scope to trade for additional pool space. The best sixteen-year-old shortstop available in a given signing period goes to one of the clubs that has kept enough pool money to pay him, and no amount of ownership wealth creates pool money the rules did not allocate. When a penalty strips a club of international pool space, it is removing access to a market rather than charging admission to it.

That is the design insight underneath the whole penalty structure. Money penalties discipline clubs that are short of money. Scarcity penalties discipline everybody, and they discipline the richest clubs hardest, because a scarce asset is the one thing a rich club cannot simply purchase more of.

Where the MLB luxury tax money goes

The proceeds do not sit in a commissioner's discretionary fund. The agreement directs them, and the direction tells you what the tax is for.

A first slice is applied to player benefit costs, which is to say straight back into the industry's obligations to the players themselves. Of what remains, a defined share goes into individual retirement accounts for players, and a defined share is distributed among clubs that finished under the threshold, subject to eligibility conditions attached to that distribution. The precise proportions are set out in the agreement and have moved between negotiating rounds.

The important thing about the destination is its scale. Even in a year when several clubs pay, the total collected is small relative to what the sport turns over and tiny relative to what moves through revenue sharing. Nobody is made competitive by their share of the tax proceeds. The tax is a deterrent priced in money, not a redistribution engine, and treating it as the mechanism by which baseball equalises its clubs gets the system backwards. That job belongs to the other half.

Revenue sharing is the other half of the machine

Baseball's revenue divides into two kinds. Central revenue, which is the national broadcast deals, the league's digital and licensing businesses and anything else sold collectively, is shared equally among all thirty clubs before anybody does anything with it. Local revenue, which is regional broadcast rights, gate receipts, local sponsorship and ballpark income, is where clubs differ from each other enormously, because a club in a large market with its own regional network operates on a different scale from a club in a small one.

The revenue sharing plan attacks the second kind. Each club contributes a defined percentage of its net local revenue, a proportion just under half under the current agreement, into a pool that is then divided equally among all thirty. The arithmetic of that is simple and severe: a club with local revenue far above the average pays in much more than it takes out, and a club far below the average takes out much more than it pays in. Clubs in the largest markets are barred from receiving altogether, whatever their own accounts happen to show, on the reasoning that a large-market club with poor revenue has a management problem rather than a structural one.

Recipients are not free to bank the transfer. The agreement obliges a club receiving revenue sharing to use those receipts to improve its on-field performance, and the union has, more than once, formally complained that particular clubs were not doing so. That obligation is easier to write than to enforce, because almost any expenditure can be characterised as improving performance eventually, and the argument over it is one of the most persistent in the sport's labour relations.

The two systems are designed to work as a pair, and they compress payroll from both directions at once. Revenue sharing raises the floor by funding clubs that could not otherwise compete for free agents. The tax lowers the ceiling by making the top of the market progressively more expensive. What sits between them is a band, and the sport's payroll distribution clusters inside that band far more tightly than the raw difference in club revenues would predict. That mechanism is worth setting beside the way other leagues cap their spending, because baseball reaches a similar compression through economic pressure where others reach it by prohibition.

The criticism of the pair is that they can reward doing nothing. A club that receives revenue sharing, keeps payroll low, loses cheerfully and collects high draft picks can be profitable while being bad, and the incentive to be bad is strongest exactly where the competitive need is greatest. The current agreement made two attempts at the problem. The pre-arbitration bonus pool pays young players more, which reduces the profit available in a roster made entirely of them. And a lottery was introduced for the top of the draft order, with limits on how many years running a club may collect a lottery pick, so that finishing last stopped being a reliable purchase order for the first selection. Whether either has changed behaviour is an argument for a longer piece than this one.

The qualifying offer, and how a draft pick becomes a price tag

The last piece connects the tax to the free agent market directly.

When a player is about to become a free agent, his club may make him a qualifying offer: a one-year contract at a value calculated by averaging the salaries of the highest-paid group of players in the game, currently the top 125. The price is therefore substantial, it is identical for every player in a given year, and it moves with the market rather than with the individual.

Eligibility is narrow. The player must have spent the entire season with the club making the offer, which is why players traded in July cannot be given one by their new team, and he must never have received a qualifying offer before in his career. He then has a defined window in which to accept or decline. Accept and he is under contract for a year at that price. Decline and he reaches the open market with a string attached.

The string is draft compensation, and its size depends on circumstances that have nothing to do with the player himself. If he signs elsewhere, his former club receives a draft pick, and the position of that pick depends on whether the former club pays the competitive balance tax, whether it receives revenue sharing, and how large a contract the player ultimately signs. The signing club, meanwhile, forfeits picks, and forfeits more of them, and deeper ones, if it is a taxpayer than if it is not. A taxpaying signing club also gives up international bonus pool money.

Read that structure carefully and the design is plain. The system is arranged so that a club already over the tax line pays a heavier non-financial price for signing a free agent than a club underneath it. Tax status is not just a bill. It is a modifier applied to the club's cost of doing business in every other market it operates in.

And the players it lands on hardest are neither the elite nor the marginal. An elite free agent is signed regardless of compensation, because no club walks away from a great player over a second-round pick. A marginal one is never given a qualifying offer in the first place, since the club would risk him accepting it. It is the good-but-not-overwhelming player, priced somewhere near the qualifying offer value, whose market goes quietly thin because every interested club is weighing him against picks it would rather keep.

Why a tax nobody is obliged to pay behaves like a cap

Put the pieces together and the soft cap emerges without anybody legislating one.

The threshold is a point at which the marginal cost of payroll jumps. Below it, a dollar of payroll costs a dollar. Above it, the same dollar costs a dollar plus tax, at a rate that depends on decisions the club made in previous years. Cross far enough and the dollar costs a dollar plus tax plus surcharge plus, at a certain point, ten places in the draft. Sign a qualifying offer free agent while over the line and the picks it costs are worse than the picks it would have cost underneath. Every one of those outcomes is legal, permitted and purchasable, and every one of them makes the next dollar more expensive than the last.

A rational front office facing a step change in marginal cost at a known, published, precisely located point will do what a rational anything does at a step change. It will bunch against it. Payrolls cluster just beneath the threshold not because a rule forbids crossing, but because the last few million before the line are the cheapest few million available and the first few million past it are the most expensive money in the sport.

That is what a soft cap is. It does not need to exist in the text.

Is it collusion, or is it just arithmetic?

The union's suspicion is neither paranoid nor new. The agreement contains an explicit prohibition on clubs acting in concert with respect to free agency, and that clause is in there because in the 1980s arbitrators found that owners had done precisely that, and the owners paid to settle it. Baseball's labour history includes a proven conspiracy against the free agent market, which is why every subsequent slow winter gets examined for a second one.

The argument for collusion runs like this. Thirty independent businesses, competing for the same players with wildly different resources, ought to produce a wide spread of payrolls and unpredictable bidding. Instead a suspicious number of them stop at the same number, publicly describe that number as a limit it is not, and go quiet in the same weeks. When the outcome looks coordinated, coordination is the obvious explanation.

The argument against is simpler and, on the evidence available in public, sufficient. Thirty firms facing an identical, published, precisely located jump in marginal cost will converge on the same behaviour without exchanging a word, because they are all solving the same optimisation problem with the same inputs. Identical incentives produce identical conduct. That is not a conspiracy, it is a mechanism, and the fact that it is indistinguishable from a conspiracy at a distance is the genuinely uncomfortable part.

Both arguments can be right at once, which is the usual state of affairs in labour disputes. A structure that makes coordinated restraint the individually rational choice does not require anyone to coordinate, and it also provides excellent cover to anyone who does. The union's real complaint is less about proving that a meeting took place and more about the design itself: the threshold was sold as a brake on the top of the market and behaves as a ceiling on the middle of it.

How a front office plans around a line it is allowed to cross

The practical version of all this is that a general manager is not managing a budget. He is managing a schedule of future average annual values against a threshold that rises modestly each year, at a rate determined by his own recent history, with penalties that fall on assets he cannot buy.

If you want to judge a club's position rather than react to a payroll table, five things tell you most of it.

Committed average annual value two and three years out, not this season's payroll. This season is already decided; every contract on it was signed months or years ago. The interesting question is how much of the club's room in the winter after next is already spoken for, and by whom.

Where the club sits in the consecutive-year sequence. A club in its first year over faces a very different marginal cost from one in its third, and the same signing means different things to each. If a club is contemplating a reset year, it shows up in trade deadline behaviour long before anybody announces it.

How much of the threshold the fixed block eats. Benefits and the bonus pool are neither optional nor small. Any analysis that treats the entire threshold as available for players overstates every club's space by the same substantial amount, and overstates it most misleadingly for the clubs closest to the line.

How many of the club's best players are still pre-arbitration, and when that ends. A roster carried by cheap young talent has a dated advantage, exactly as a rookie-contract quarterback does in football. The date is knowable years ahead. The clubs worth watching are the ones spending aggressively just before it arrives, because they can read the calendar too.

Whether the club receives revenue sharing. It changes the compensation picks it collects, the picks it forfeits, the scrutiny its payroll decisions attract, and the political weather around every winter it spends quietly.

None of those five is a secret. All of them are more informative than the number in the December headline, which describes a contract's nominal value in a system that does not charge on nominal values, does not care when the cash is paid, and settles up quietly in the winter on a figure hardly anybody bothers to calculate.