Give yourself a week for a two-hour job and the job takes a week. Everyone has watched that happen, and the line that explains it, “work expands so as to fill the time available for its completion”, is quoted constantly. The advice that follows from it, set a tighter deadline, is good advice. The explanation attached to it is probably the wrong one.
Parkinson’s Law says work expands to fill the time available. It has been tested directly only a handful of times, all between 1967 and 1975. The much better evidenced claim sits next to it: a specific, difficult goal beats a vague one, which is what a real deadline actually supplies.
Those two explanations point at different things to fix, which is why the distinction is worth an article rather than a footnote.
What does Parkinson’s Law say, and how is it taught?
As a rule of thumb it is stated exactly as above, and universities pass it on in that form. Bowdoin College’s learning strategies page lists it under time savers: “Set yourself a deadline or time limit for a task (Parkinson’s law = Work expands to fill the time available for its completion).”
Notice the assumption inside that instruction. It treats the law as a causal mechanism, a property of time that makes a task swell to fit whatever space it is given, and it treats a deadline as the counter-measure. If that is the mechanism, then shrinking the window is sufficient on its own, and you do not have to do anything else.
The law is attributed to the British historian C. Northcote Parkinson and entered circulation as commentary on administration rather than as a research result. A slogan is not evidence, and a slogan repeated for seventy years is still not evidence.
How much has it actually been tested?
Directly, three times in the era when psychologists were interested in it, and the citations are worth having in one place because almost nobody who quotes the law supplies them.
Bryan and Locke (1967), “Parkinson’s Law as a goal-setting phenomenon”, in Organizational Behavior and Human Performance, volume 2, pages 258 to 275. Aronson and Landy (1967), “Further steps beyond Parkinson’s Law: A replication and extension of the excess time effect”, in the Journal of Experimental Social Psychology, volume 3, pages 274 to 285. And Latham and Locke (1975), “Increasing productivity and decreasing time limits: A field replication of Parkinson’s law”, in the Journal of Applied Psychology, volume 60, pages 524 to 526.
Two things are worth taking from that list without opening the papers. The first is the vocabulary: the experimental literature calls this the excess time effect, and Parkinson’s Law is what the rest of us call it, which is why searching for evidence under the popular name finds so little. The second is the size: three papers across eight years, on a claim that is repeated daily in every language that has productivity blogs.
All three sit behind paywalls, and this article will not summarize results it has not read. That is a real limitation and it is also part of the story, because it is exactly why the maxim travels so much better than the research does. The slogan is free and the test is not.
What can be said without reading them is where they sit. Two of the three are from the research program that became goal-setting theory, and that program’s own results are open, extensive, and point somewhere specific.
What is the better-evidenced claim?
Locke and Latham summarized 35 years of that work in the American Psychologist in 2002, and their core result is stated plainly: “We found that specific, difficult goals consistently led to higher performance than urging people to do their best. The effect sizes in meta-analyses ranged from .42 to .80.” They give the reason in the next breath: “when people are asked to do their best, they do not do so. This is because do-your-best goals have no external referent and thus are defined idiosyncratically.”
Read that against the maxim and the picture rearranges. A deadline is not primarily a fence around the time. It is one of the few ways an ordinary task acquires a specific, external, non-idiosyncratic standard. “Finish the draft” is a do-your-best goal. “Finish the draft by 11” is a specific one. The evidence that the second kind of goal raises performance is well replicated and meta-analyzed; the evidence that time itself causes work to expand is three papers from the late 1960s and mid-1970s.
That is not a debunking. It is a relocation, and it changes what you should spend effort on.
What follows if the mechanism is the goal?
A deadline with no definition of done is not a goal. “I will finish this by 3” only functions if “finished” means something you could check from the outside. If it does not, you have put a time limit around a do-your-best goal and left out the part with the evidence behind it.
Shrinking the window is not free. The maxim implies that cutting the time costs nothing, because the extra was padding. The goal-setting literature is more careful: Locke and Latham note that performance leveled off or decreased when the limits of ability were reached, or when commitment to a highly difficult goal lapsed. A limit far below what the task needs does not compress the work, it breaks the commitment.
It says nothing about interruptions. A task that took four hours instead of two may have occupied four hours of elapsed time and ninety minutes of attention. Expansion is one explanation and fragmentation is another, with better measurement behind it. The cost of context switching covers what the observational research on interrupted work actually recorded.
How would you test it on yourself?
By measuring, and by being honest about how weak a personal test is.
Take a task you do repeatedly with a recognizable finish. Record how long it takes with no stated limit for a few weeks, then state a limit and a definition of done and record it again. Here is an invented example, with the figures made up to show the arithmetic rather than to suggest a result.
| Run | No stated limit (minutes) | With a stated limit (minutes) |
|---|---|---|
| 1 | 95 | 70 |
| 2 | 110 | 75 |
| 3 | 80 | 65 |
| 4 | 120 | 90 |
The unlimited runs total 95 + 110 + 80 + 120 = 405 minutes, so the mean is 405 divided by 4, which is 101.25. The limited runs total 70 + 75 + 65 + 90 = 300, so the mean is 75. The difference is 101.25 minus 75, which is 26.25 minutes per run, and 26.25 divided by 101.25 is about 26 percent.
Four runs against four proves nothing. The spread inside each column, 80 to 120 and 65 to 90, is wide enough that a different four weeks could reverse the sign. That is not a flaw in the method, it is the method being honest, and it is why the published work used crews and classrooms rather than individuals.
What a personal record can tell you is narrower and still useful: whether your own recurring work has a stable duration at all. Most people have never checked, and it is common for the same task to range over a factor of two, which makes any deadline you set a guess dressed as a decision.
What to do with it
Set the deadline, and spend the effort on the part with the evidence behind it. Make the goal specific, make “done” checkable from outside your own head, and put the limit somewhere you can commit to. Time blocking is one way to attach those limits to a calendar, with the same caveat that the technique’s popularity runs ahead of its evidence.
Then measure, because every deadline you set rests on an estimate of how long the work takes, and that estimate is usually a memory of the last time rather than a record of it. Timeness is for the record: make an activity, tap once to start and once to switch, and the day timeline and the week, month and year charts show the real durations. It has no countdown, no timer limit and nothing that enforces a deadline, because enforcing was never the part that worked. Give each kind of work its own activity and you can run the comparison above properly, with dozens of runs instead of four.
The maxim is memorable because it is a joke about bureaucracy. The useful version is duller: a specific goal beats a vague one, and you cannot set a specific goal about time without knowing how long things take. Keeping a work log is where that knowledge comes from, and a time audit is the fastest way to get the first month of it.
Sources
- Bowdoin College, Baldwin Center: Time Management Strategies
- Locke and Latham (2002), Building a Practically Useful Theory of Goal Setting and Task Motivation: A 35-Year Odyssey, American Psychologist 57(9), 705-717, hosted by Stanford University School of Medicine
- Bryan and Locke (1967), Parkinson's Law as a goal-setting phenomenon, Organizational Behavior and Human Performance 2, 258-275
- Aronson and Landy (1967), Further steps beyond Parkinson's Law: A replication and extension of the excess time effect, Journal of Experimental Social Psychology 3, 274-285
- Latham and Locke (1975), Increasing productivity and decreasing time limits: A field replication of Parkinson's law, Journal of Applied Psychology 60, 524-526