You go on a one-week vacation with the big suitcase and you fill it. The next year you go for the same week with the small one, you fill that one too, and you had everything you needed.
Work does something very similar. A task with a one-day deadline gets done in a day; with a two-week deadline it gets done in two weeks. Nobody was lying in the sun. Reviews and details that kept piling up filled the time.
In 1955 the historian Cyril Northcote Parkinson published a half-joking, half-serious essay in The Economist about how the British bureaucracy kept growing, adding employees even as the empire it ran was shrinking. That's where Parkinson's law comes from: work expands so as to fill the time available for its completion.
What I like most about this law is that nobody decides to stretch the work. It just happens.
The meeting booked for an hour lasts an hour, even though the topic could have been settled in fifteen minutes. The feature estimated at one sprint takes up the whole sprint, with a couple of refactors nobody asked for and two more passes on the design. The document due Friday gets finished Friday at six. And the project with no date never gets finished.
Inside that expansion, pretty reasonable things happen: what was already fine gets polished, what didn't need discussing gets discussed. Parkinson had another law for this, the law of triviality, with a great example. A committee approves a nuclear reactor with barely any discussion and then spends ages debating what material to use for the roof of the bike shed. Nobody understands much about the reactor; everyone has an opinion on the roof. Teams do the same: we gravitate to what we understand.
The obvious temptation is to tighten deadlines, and it's also the approach that falls apart fastest. You cut the time, the team rushes, something ships, and next time you cut it again. It works a couple of times, until the shortcuts show up, then the tech debt and the burned-out people. If on top of that you turn the deadline into the metric that matters, Goodhart's law does the rest: the date is met and what gets delivered stops being useful.
For me the deadline is useful for something else, which is forcing the conversation about scope. With little time someone has to decide what's in and what's out, and that decision is what actually organizes the work. What shrinks the work is deciding what we're not going to do. The deadline only forces you to decide it.
That's why I'd rather bound the problem than the time. Make it clear what has to happen for something to be done and what metric it has to move. With that, the work has an edge and the expansion has something to bump into.
Not all extra time is waste, either. Sometimes that second pass on the design was necessary, and sometimes the conversation that went off on a tangent led to the best idea of the month. The law describes a tendency, and the hard part is noticing when the extra time adds something and when you're just stuffing another pair of sneakers into the suitcase just in case.