Does daylight saving time actually do anything?

Published on September 7, 2026

Tags:sciencetimekeepingsleeppolicy

Surreal painting of a melting wooden hourglass resting on a golden tidal beach at sunset, with a stag standing in the shallow reflective water

On 14 July 2026, the United States House of Representatives voted 308–117 to make daylight saving time permanent. The Senate has not acted; it is not law. But the direction of travel is clear, and it runs directly against the expert consensus. The American Academy of Sleep Medicine’s position, updated in 2024, is that the United States should stop changing its clocks — onto permanent standard time, “which aligns best with human circadian biology”.

Both sides of the DST argument deploy confident numbers. It saves energy. It wastes energy. It causes heart attacks. It prevents road deaths. This post is an attempt to do something simpler than either: say what the peer-reviewed studies actually measured, in which population, and what they did not.

The short answer: the energy question is settled and the answer is “basically nothing”. The health effects of the change are real, modest and brief. The benefits of the arrangement are also real. And that distinction — the change versus the arrangement — is the one almost every DST argument silently collapses.

Energy: three right answers, one conclusion

The original rationale for DST was fuel. It is the one claim we can close.

The United States government’s own answer: when Congress extended DST by four weeks in 2007, the Department of Energy measured the extension specifically — about 0.5 percent electricity savings per extended day, roughly 1.3 terawatt-hours, or 0.03 percent of annual US consumption. Savings came from evening lighting, partly clawed back by darker mornings and air conditioning.

The natural experiment: Indiana, until 2006, was a patchwork of counties that kept DST and counties that didn’t. Statewide adoption handed economists a clean before-and-after. The result, from billing data for roughly 224,000 households: DST increased residential electricity demand by 0.8 to 1.03 percent across the whole DST season, costing Indiana households about $9 million a year. The extra hour of hot evening landed in the air-conditioning window.

The whole literature: a meta-analysis of 162 estimates from 44 studies found a mean reported saving of 0.34 percent — but the best-practice estimates cluster at 0.01 percent, statistically indistinguishable from zero. Savings grow with latitude; nearer the equator, DST increases consumption. The authors’ conclusion is the quotable one: the effects are “too small to justify the biannual time-shifting”.

None of these studies contradicts another. Each measured something different — extension weeks nationally, residential demand across a season in one state, the entire literature. What they agree on is the size of the prize: a fraction of a percent, with a sign that flips with latitude. As energy policy, DST is over.

A horizontal chart on a percent axis showing the daylight saving energy effect as three entries: a saving of 0.5 percent for the US extension study, an increase of up to about 1 percent for the Indiana experiment, and a meta-analysis mean near 0.3 percent saving with its best-practice estimate at approximately zero.
Three studies, three different questions, one conclusion: the energy effect of DST is a fraction of a percent — and its sign flips with the study design.

One number deserves a burial: during the 2005 debate, politicians claimed each extra week of DST would save the equivalent of 100,000 barrels of oil a day. When the Department of Energy measured gasoline consumption and traffic volume, the effect was statistically insignificant. The claim was quoted into the record; the measurement came back empty.

Heart attacks: a real effect, almost always misquoted

Here is the number that circulates: heart attacks jump 24 percent after the spring clock change.

It is a real measured number — and quoting it that way inverts what the study found. The 24 percent comes from a Michigan registry of patients undergoing angioplasty, and it is a Monday-only spike. The same study found no difference in total weekly heart-attack counts, and its own conclusion is that the clock change “impacts the timing of presentations for AMI but does not influence the overall incidence”. Some heart attacks that would have happened later that week arrived on Monday.

A bar chart comparing a 24 percent Monday-only heart-attack spike with roughly 5 percent whole-population weekly increases, annotated to say the larger number reflects timing within the week rather than additional events.
The 24 percent is real and it is a Monday: plotted beside the whole-population weekly effects it comes from a single registry’s weekday timing, not a week of added risk.

The whole-population numbers are smaller and steadier. In the Swedish national registry — every recorded heart attack, 1987 to 2006 — incidence rose about 5 percent in the week after the spring shift (and only the first weekday was affected in autumn). A meta-analysis of seven studies and more than 115,000 subjects confirms the shape: a spring increase around 5 percent, no significant change in autumn.

Stroke tells the same story faster. Finnish hospitalization data across 2004–2013 shows ischemic stroke admissions up 8 percent — but only for the first two days after a transition, diluting to statistical noise across the week. The effect was strongest in the over-65s, women, and cancer patients.

These are registry studies: they count hospital presentations, not risk to any individual. The honest summary is that the spring change jolts vulnerable cardiovascular systems for a few days, and the week as a whole barely moves.

Roads: the harm is the change, the benefit is the arrangement

Every scary DST number measures the transition. Every benefit number measures where the hour of light sits. Roads show both at once.

The transition: across 732,835 fatal crashes in the United States, 1996 to 2017, the spring clock change increased fatal-accident risk by about 6 percent. The risk was worse in the morning, and worse the farther west you live within your time zone, where the sun rises latest on the clock. The autumn change showed no effect. (A published commentary in the same journal disputes how much of the spring effect is DST-specific rather than seasonal; the attribution is contested, the observation is not.)

The arrangement: shifting an hour of clock-shifted evening daylight into the commuting hours is protective. A counterfactual estimate — not an observation, a calculation — put year-round DST at about 901 fewer fatal crashes over 1987–1991, 727 of them pedestrians, had it been in force.

These do not contradict each other. One is the cost of moving the clocks. The other is the value of where they sit. Any argument that quotes one without the other is arguing in bad faith or hasn’t noticed the difference.

A two-panel chart, the left showing a 6 percent rise in fatal crash risk after the spring clock change as an observed bar, the right showing an estimated 901 fewer fatal crashes under permanent daylight saving time rendered in outline to mark it as a calculation.
The harm is measured, the benefit is computed: the spring-change crash risk is an observation, the 901 lives a counterfactual estimate — different epistemic classes the debate routinely conflates.

What sleep loses

The mechanism behind all of the transition effects is measurable, and it is sleep.

On the Monday after the spring change, US workers sleep an average of 40 minutes less than on other Mondays — this from time-use data, not self-report. US mine workers that same Monday record more injuries, and more severe ones. After the autumn change, when the hour is returned, neither effect appears.

The deeper finding is chronic, not acute. In a survey of 55,000 people plus 50 monitored subjects, sleep timing followed the seasonal progression of dawn under standard time — but not under DST. The authors’ conclusion, in their words: “the human circadian system does not adjust to DST”. The body keeps tracking the sun while the schedule tracks the clock, and the two drift apart for as long as the policy lasts. This is the evidence base for the sleep-medicine position, and it does not evaporate after the transition week.

Crime: one crime, honestly

DST’s effect on crime is real, specific, and routinely overgeneralized. In US crime-reporting data covering 2005–2008, daily robbery fell about 7 percent in the weeks after DST began — down 27 percent in the sunset hours specifically, which is the point: robbery is a darkness crime, and evening light removes its cover. For aggravated assault there were no consistent effects; results for other crimes were too fragile to lean on.

The defensible sentence is “evening light reduces robbery”. “DST reduces crime” is a different sentence, and the study does not support it.

So which clock?

Here is the curious shape of the current argument. Almost everyone — the House, the sleep doctors, most of the public — agrees the twice-yearly change should stop. The disagreement is which clock to stop on.

The House voted for permanent DST: lighter evenings, darker mornings, forever. The American Academy of Sleep Medicine recommends permanent standard time, because permanent DST means permanent morning darkness and a circadian misalignment the transition studies only sample. Russia chose standard time when it quit changing its clocks in 2014; Turkey chose permanent DST in 2016. The evidence, honestly read, underdetermines the choice — but to the extent it underwrites anything, the sleep-physiology literature underwrites standard time.

Most of the world has simply left the table. A majority of countries never observe DST, and in the last decade Azerbaijan, Brazil, Iran, Jordan, Namibia, Russia, Samoa, Syria, Turkey and Uruguay have all ended seasonal changes. Mexico abolished summer time in 2022, except in its border cities.

A world map with a minority of countries shaded as observing daylight saving time and most countries unshaded, showing the practice concentrated in Europe, North America and parts of the southern cone.
The table most of the world already left: Europe keeps changing its clocks; a majority of countries never do or have stopped — and the quitters of the last decade span every continent.

Two folk stories to drop on the way out. Farmers did not want DST — they were historically among its most organized opponents, working an extra hour of morning darkness to hit market times. And Benjamin Franklin did not invent it: his 1784 essay was a satirical joke about candles, with the modern clock-shifting policy credited to George Vernon Hudson.

(The tale that candy makers lobbied for the fall extension to cover Halloween surfaces in the literature only as “reportedly”; the primary evidence could not be traced, so treat it as unverified.)

The part that never changed

DST moves an hour of daylight on the clock — it creates none of it. Everything above is an argument about where a fixed quantity of light should sit, and about whose schedule the clock should serve. It is a wall-clock policy in the fullest sense: a time zone is a rulebook, and DST is its most argued rule, rewritten by the same kind of political decision that drew the zones in the first place.

Elapsed time sits outside the argument entirely. Whether your legislature lands on standard time, DST, or keeps flipping the clocks twice a year, the interval between two moments on one clock is untouched by any of it. When you press start on TiCaNo Stopwatch, nothing about the debate reaches the measurement.

Which is, perhaps, the one thing everyone in it can be grateful for.

Sources

← Back to all posts