Basketball's spending rules are usually described as a cap, but the exceptions to it do most of the work. They determine which teams can add players and which cannot.
A pure limit would freeze the league
If teams could only sign players with room under a single number, almost every competitive roster would be locked. Good teams sit at or above the limit by definition, because paying good players is what puts them there.
A frozen roster is bad for the sport. Injuries would go uncovered, teams could not replace a departing starter, and continuity would be impossible to maintain.
So the system pairs the limit with named routes that allow spending above it, each with its own size and conditions.
Keeping your own players is treated differently
The most important exception lets a team re-sign a player it already has, even if doing so takes the payroll past the limit.
The reasoning is continuity. Without it, a team that developed a player would lose him the moment he became good enough to command real money, and drafting well would be punished.
This is why rosters change less than the raw numbers suggest. Incumbency is a genuine advantage in negotiations, and players know it.
Mid-sized exceptions decide the middle class
Teams above the limit are given a fixed allowance to sign an outside player. That allowance sets the market price for the entire tier of useful non-star players.
Because most contenders have the same allowance available, they compete for the same players at roughly the same number, and the deciding factor becomes role and location rather than money.
Trades have to balance
A team above the limit cannot simply absorb salary in a trade. Incoming and outgoing money must roughly match, which is why deals accumulate players who exist mainly to make the arithmetic work.
This matching requirement explains a great deal of trade-deadline behaviour, including contracts that appear to have no basketball purpose.
It also explains why some obvious swaps never happen. Two teams can agree on player value and still be unable to construct a legal deal.
The result is a hierarchy of flexibility
Teams end up sorted less by wealth than by which tools remain available to them, and those tools tighten as payroll climbs.
A team well under the limit can sign anyone. A team just above it has its allowance. A team far above it has almost nothing but its own players and the draft.
Roster building is therefore a question of staying in a band where the useful exceptions still apply, which is a different skill from simply identifying good players.