In Depth
Basketball · The NBA’s aprons · 1984 – today

The ladder

The NBA has a soft cap. It did. Every tool a team can use to add a player, the exact line at which each one stops existing — and the point, reached in 2023, where the list runs out and a soft cap becomes the hardest ceiling in American sport.

The NBA is usually described as having a soft cap — a limit teams are allowed to exceed. That was always a slightly odd way to put it, because the league never set a spending limit at all. What it set was a limit on the tools. Below the cap a team may sign anyone it can afford. Above it, a team may add salary only through a list of named exceptions written into the collective bargaining agreement: Bird rights to keep its own players, a mid-level exception to sign somebody else’s, a bi-annual exception, salary matching in trades, the minimum. That is the whole system, and everything strange about the modern NBA follows from it. Hardness has never been a matter of how much a team may spend. It is a matter of how many doors are still open.

Which makes the system countable, and the count is the surprise. Two of those tools require cap room and die at the cap. At the luxury-tax line — the line every fan knows, the one that generates the headlines and the nine-figure bills — exactly none of them die. Crossing it has never cost a team anything but money. Seven die at the first apron: the full mid-level, the bi-annual, acquiring a player by sign-and-trade, signing anyone off the buyout market above a certain size, taking back more salary in a trade than you send out, and two more besides. Five die at the second: any mid-level at all, combining two salaries to match a third, sending cash, and both halves of the sign-and-trade. Above the second apron, at $221.686M in 2026-27, four things are left — minimum contracts, your own Bird rights, the draft, and a trade that sends out at least as much as it brings back. No exception exists above that line, which is the technical definition of a hard cap, arrived at without the word ever being redefined.

The rule binds in two directions at once, and this is the part that catches people. Where a team sits decides which tools it may reach for. But reaching for a tool fixes a ceiling it may not cross for the rest of the league year, whatever its payroll was at the time. Sign a player with the full mid-level in July and the team is hard-capped at the first apron until the following June — so a trade that was legal that morning is illegal that afternoon. It also means a tool can be simultaneously available and unusable: a team $9M below the first apron may use the $15.044M mid-level in principle, and cannot in practice, because the signing itself would breach the ceiling the signing creates. For twenty years crossing a line cost money. Since 2023 it costs the ability to improve, and the difference shows up as champions being dismantled — Denver losing its bench in the first week of the new agreement, Boston trading two starters in 2025 to duck the second apron by about $4.5M, and trading the 2024 Finals MVP a year after that.

This guide works the whole system. The ladder below is the signature figure: a money axis with the five lines on it — floor, cap, tax, first apron, second apron — and beside it every roster-building tool as a rung, each going dark at the line where it dies. Move the payroll marker up the axis, or drop a real 2026-27 team onto it, and the rungs go out beneath you, with a readout naming what was just lost and what survives. The empty band at the tax line is not an oversight in the drawing; it is the argument. Beyond it the guide catalogues every exception at its 2026-27 price, runs a summer of transactions through the hard-cap rule to show the ceiling clamping down, computes the tax bill under all three of the rate schedules the league has used since 2002, and lays out thirty franchises against the lines for every season the luxury tax has existed.

Season
$148.465M
the salary floor
$164.961M
the cap
$200.428M
the tax line
$209.015M
the first apron
$221.686M
the second apron
$214.01M

The slider steps in thousands, so it can never land exactly on a line — the buttons put the marker one dollar over each one, which is where the rules change.

  • Survives everything

    • Minimum contracts
    • Bird rights
    • The draft
    • An even trade

  • Dies at the second apron — $221.686M

    • The taxpayer mid-level$6.064M· using it hard-caps you at second apron
    • Aggregating salaries in a trade· using it hard-caps you at second apron
    • Cash in a trade
    • Matching with a signed-and-traded player
    • A sign-and-trade trade exception

  • Dies at the first apron — $209.015M

    • The non-taxpayer mid-level$15.044Mgone
    • The bi-annual exception$5.477Mgone
    • Acquiring by sign-and-tradegone
    • The mid-level in a trade or claimgone
    • The buyout marketgone
    • Taking back more than you sendgone
    • Last year’s trade exceptiongone

  • Dies at the tax line — $200.428M

    Nothing. Crossing the luxury-tax line costs a team money and not one tool. This empty band is the argument.

  • Requires cap room — $164.961M

    • Cap roomgone
    • The room exception$9.366Mgone

A team over the second apron at the end of the regular season has its first-round pick seven drafts away frozen: it cannot be traded.

$214.01M in 2026-27 is over the first apron. $7.67M below the second apron. 9 of 18 tools remain; 9 are gone — cap room, the room exception, the non-taxpayer mid-level, the bi-annual exception, acquiring by sign-and-trade, the mid-level in a trade or claim, the buyout market, taking back more than you send, last year’s trade exception.

solid rung — still available at this payrolldashed and struck through — gonedashed line — the salary floorthe outlined band is where the marker sits

Every line, in words

In 2026-27 a team’s payroll can cross five lines, and they do different kinds of work.

  • the salary floor — $148.465M. The minimum a team must spend. A club below it at the end of the season pays the shortfall out to its own players, which is why nobody stays below it.
  • the cap — $164.961M. The line most people think of as the limit. It is not one: it is the point past which a team may no longer sign whoever it likes, and must use a named exception instead. Most teams are above it, most of the time.
  • the tax line — $200.428M. The line that costs money. Above it a team pays a bill on a bracket schedule, rising steeply, doubling for repeat offenders — and loses nothing at all. Every tool it had below the line, it still has.
  • the first apron — $209.015M. The first line that takes things away. Seven tools stop here, including the full mid-level exception and the ability to take back more salary in a trade than you send out.
  • the second apron — $221.686M. The ceiling. Five more tools stop here, including every mid-level and the ability to combine two salaries into one trade. Nothing above this line can be bought — only drafted, re-signed, or found at the minimum.

Every tool, in words

The whole toolbox, in the order the ladder groups it. Each entry is worked out at length in the toolbox section; this is the list itself, so it can be read without moving anything.

Cap room
— Sign any free agent for anything up to the space between your payroll and the cap. Dies at the cap.
The room exception
— A consolation exception for a team that has just spent all of its cap room. Dies at the cap, worth $9.37M.
The non-taxpayer mid-level
— Sign a free agent for up to $15.044M a year, for up to four years, without cap room. Dies at the first apron, worth $15.04M.
The bi-annual exception
— A second, smaller signing exception — usable only every other year. Dies at the first apron, worth $5.48M.
Acquiring by sign-and-trade
— Take in a free agent who signs with his old team and is traded to you in the same act. Dies at the first apron.
The mid-level in a trade or claim
— Absorb an incoming salary through the mid-level rather than by matching salaries. Dies at the first apron.
The buyout market
— Sign a player waived mid-season whose contract was larger than the full mid-level. Dies at the first apron.
Taking back more than you send
— Come out of a trade with more incoming salary than outgoing. Dies at the first apron.
Last year’s trade exception
— Use a traded player exception created in a previous league year. Dies at the first apron.
The taxpayer mid-level
— Sign a free agent for up to $6.064M a year, for up to two years. Dies at the second apron, worth $6.06M.
Aggregating salaries in a trade
— Combine two or more contracts to match one larger one. Dies at the second apron.
Cash in a trade
— Sweeten a deal with money — up to $8.495M across the league year. Dies at the second apron.
Matching with a signed-and-traded player
— Use an outgoing sign-and-trade player’s new salary to bring salary back. Dies at the second apron.
A sign-and-trade trade exception
— Spend the trade exception a sign-and-trade would otherwise generate. Dies at the second apron.
Minimum contracts
— Sign any player, at any time, for the league minimum. Never dies.
Bird rights
— Re-sign your own free agent for any amount up to the maximum, cap or no cap. Never dies.
The draft
— Sign your own draft picks on the rookie scale, whatever your payroll. Never dies.
An even trade
— Swap one contract for another of comparable size, sending out at least as much. Never dies.

Common questions

What is the second apron in the NBA?

The second apron is a payroll level — $221,686,000 in 2026-27, $17.5M above the luxury-tax line when it was introduced and indexed to the cap since — above which a team loses the last of its roster-building tools. It cannot use any mid-level exception, cannot combine two or more salaries in a trade to match a larger one, cannot send cash in a trade, and cannot use either half of a sign-and-trade. Its first-round pick seven drafts away is frozen at the end of any season it finishes above the line, and drops to the end of the first round if it finishes above the line in two of the following four. Because no exception exists above it, the second apron is a genuine hard cap: a team there can want a player, be willing to pay him, and have no legal mechanism to add him.

What is the difference between the luxury tax and the aprons?

The tax charges you; the aprons forbid you. Crossing the luxury-tax line has never removed a single tool from a team — it generates a bill, calculated on a bracket schedule, and that is the entire penalty. The Golden State Warriors paid $170,331,194 in 2021-22, the largest bill in league history, and kept every exception they had. The aprons are different in kind rather than in degree: they are the lines at which the CBA starts taking things away, and above the second one there is nothing left to take. This is why a team can be comfortably able to afford a signing and still be unable to make it.

How does an NBA team get hard-capped?

By using a tool, not by spending money. A team is hard-capped at the first apron for the remainder of the league year if it acquires a player by sign-and-trade, uses any part of the bi-annual exception, uses more than the taxpayer share of the mid-level exception to sign a player, or uses any part of the mid-level to take a player in a trade or off waivers. It is hard-capped at the second apron if it uses any part of the mid-level to sign a free agent, combines salaries in a trade, or sends cash. The cap applies from the moment the transaction completes through the following 30 June, and it applies regardless of what the payroll was when the tool was used — which is why teams sometimes decline a signing they can plainly afford.

When did the NBA aprons start?

Earlier than most people think. The 2011 CBA created a single line above the tax, set at a flat $4M above it, with its restrictions in force from 2013-14: a team more than $4M over the tax line could no longer take a player in a sign-and-trade or use the full mid-level. The 2017 CBA moved that line to $6M above the tax and indexed it to the cap. It was not called the "first" apron until the 2023 CBA added a second one, $17.5M above the tax, and widened what both of them prohibit. Those restrictions were phased in across 2023-24 and fully in force from 2024-25. Calling the 2015 line "the first apron" is a small anachronism this guide avoids: until 2023 there was only the apron.

How is the NBA luxury tax calculated?

Marginally, in brackets, and the schedule has been rewritten three times. From 2002-03 to 2012-13 it was a flat dollar of tax for every dollar of team salary over the line. From 2013-14 the 2011 CBA replaced that with $5M brackets at rising rates — $1.50, $1.75, $2.50, $3.25 per dollar, then $3.75 rising by $0.50 in each further bracket — plus a repeater surcharge of a dollar in every bracket. From 2025-26 the 2023 CBA’s schedule applies: $1.00, $1.25, $3.50 and $4.75, with repeater rates of $3.00, $3.25, $5.50 and $6.75, and bracket width now growing with the cap rather than sitting at a flat $5M. The first two brackets got cheaper and everything above them much more expensive, which is a deliberate shape: a small overage is survivable and a large one is not.

What is the repeater tax?

A surcharge for teams that stay over the line. It first applied in 2014-15, to teams that had paid tax in each of the three previous seasons; from 2015-16 the test settled into its modern form, three of the previous four. A repeater pays a higher rate in every bracket — under the current schedule, $3.00 per dollar where a first-time payer pays $1.00, rising to $6.75 where the first-timer pays $4.75. Nothing about repeater status is stored in this guide’s dataset: it is computed from each franchise’s own record of taxpaying seasons by the same rule the CBA states, which is the only way to be sure the two agree.

Were there seasons when no NBA luxury tax was paid?

Two, and they are the reason this guide’s record has declared holes in it. The 1999 CBA wrote a luxury tax into the agreement but made it conditional: it applied only if league-wide salaries and benefits exceeded a set share of basketball-related income. The tax was bargained to begin in 2001-02, the threshold was not reached, and no team paid a cent however large its payroll. The same happened in 2004-05. The first season the tax was actually levied was 2002-03, when $173.3M was collected. The 2005 CBA removed the condition, and it has been assessed every season since.

Can a team over the second apron still improve its roster?

Yes, but only in four ways, and none of them involves buying. It can re-sign its own free agents using Bird rights, which the aprons never touch. It can sign anybody at the league minimum, which is why the most expensive rosters in the league fill their last places every February with veterans on minimum deals. It can draft well and sign its picks on the rookie scale. And it can make a trade that sends out at least as much salary as it takes back, one contract at a time, without aggregating and without cash. What it cannot do is turn three good players into one great one, which is how almost every star has ever changed teams.

Going deeper: Larry Coon's Larry Coon’s NBA Salary Cap FAQ has been the public reference for this rulebook for over twenty years, question by question, and remains the place to go when a specific provision needs its exact wording; Hoops Rumors’ CBA glossary keeps a working entry for every mechanism here — the aprons, the hard cap, each mid-level — updated as the figures change, and publishes the annual lists of which teams are hard-capped and at which line; the NBA and the National Basketball Players Association's The 2023 Collective Bargaining Agreement is the source everything above is downstream of; Article VII on the salary cap and its exceptions is where the tool list this guide counts actually lives; Mark Deeks's A Complete History of NBA Luxury Tax Payments is the compilation the record layer leans on for the seasons before the cap sites had trackers, and the source of the largest single tax bill in league history.

The tax is assessed on team salary and the aprons on apron team salary, which is a different computation — it counts unlikely bonuses, raises contracts with none or one year of service to the two-year minimum, and treats qualifying offers and required tenders as salary, among other adjustments. This guide keeps the two figures separate and takes published apron salaries where they exist rather than deriving its own.

The simulator is a teaching instrument, not a compliance tool. It models the rules the guide teaches — whether a tool may be used, whether more salary may come back than goes out, whether salaries may be aggregated, and what ceiling a tool fixes — and deliberately omits the CBA’s full salary-matching bands, trade-exception expiry, roster-count rules and the many timing provisions around the moratorium. A move it permits is not thereby legal.

Rule statements are made under the 2023 CBA unless an earlier agreement is named. The apron restrictions were phased in across 2023-24 and fully in force from 2024-25. Band names are era-aware throughout: there was no "first" apron before there was a second one to distinguish it from, and the 2011 and 2017 agreements’ single line is called simply the apron.

Cap and tax levels are the league’s announced figures where it published them and SalarySwish’s cap-history table cross-checked against Hoops Rumors elsewhere. The apron figures for 2013-14 to 2016-17 are the 2011 CBA’s flat $4M-above-the-tax rule applied, which is how they were derived at the time; a test pins them to it. Where two sources differ by rounding, the league’s own figure is used.

The record covers 2001-02 through 2025-26 — every season in which a luxury tax existed, plus the two in which one was written into the agreement and never levied. It records whether a team was a taxpayer, not whether it was over or under the salary cap: team-by-team cap positions before the apron era come from reconstructions rather than league publications, and this guide does not print a digit it cannot check. Repeater status is not stored either; it is derived from each franchise’s own taxed seasons by the rule that applied in the season being asked about, and the derivation is checked against five seasons of independently published repeater lists.

Two declared gaps in the apron layer. Apron standings are recorded for 2023-24 and 2024-25, the seasons with certified end-of-season tables. For 2025-26 the last published apron tracker is a 6 March 2026 snapshot, taken with New York $369,736 below the second apron and Minnesota $221,323 below the first and a month of the season still to play, so that season’s taxpayers are shown without apron standings rather than have a March position presented as a result. 2026-27 has no standings at all — the league year opened on 1 July 2026 and no tax has been assessed.

The grid was built from per-season taxpayer lists and then cross-checked against two independently published aggregates that were not used to build it: the number of taxpayers in each season, and the number of taxpaying seasons for each franchise through 2017-18. All 23 season counts reconcile and 29 of 30 franchise counts do. The exception is Portland, and the check found the error in the aggregate rather than in the grid — see the record section for the three constraints that establish the missing 2010-11 payment. Both numbers are kept, and a test asserts that this is the only place they differ.

Episodes describe transactions that happened and the reasons the people involved gave for them. Where a move would plausibly have been made anyway, the entry says so rather than claiming the rule as the sole cause. No team has yet actually lost a first-round pick to the second apron’s freeze — the earliest frozen pick falls in 2031 — and the guide does not describe one as though it had.