Time goes back when clocks are adjusted for daylight saving time, which in the United States happens on the first Sunday of November at 2:00 a.m. local time. In 2025, that date is November 2. Clocks shift from 2:00 a.m. back to 1:00 a.m., creating a repeated hour. Separately, timekeeping laboratories occasionally insert a leap second, which adds a second rather than turning clocks backward. These are two distinct mechanisms with different causes and consequences.
The mechanism behind the November rollback
Daylight saving time is a scheduling convention, not a physical change in time. Governments define local time by statute, choosing an offset from Coordinated Universal Time (UTC). In the US, most states observe standard time as UTC−5 through UTC−10 depending on the time zone. During daylight saving, the offset shifts one hour forward, so clocks read UTC−4 through UTC−9.
The November transition reverses that statutory offset. At 2:00 a.m. local daylight time, the clock is reset to 1:00 a.m. standard time. The wall clock repeats the 1:00 hour, so any event scheduled between 1:00 and 2:00 a.m. occurs twice in civil time.
The underlying physics is unaffected. Earth's rotation, atomic oscillations, and the flow of entropy continue unchanged. What changes is the label society attaches to a moment.
The reason for the autumn change is practical. Shifting an hour of daylight from evening to morning reduces the mismatch between clock time and solar time during winter months. Sunrise arrives earlier on the clock, which matters for school commutes and outdoor work.
Plain language: the clock moves back, but nothing in the universe does.
Quantified context: dates, offsets, and duration
In the US, daylight saving time begins on the second Sunday of March and ends on the first Sunday of November. The fall transition date shifts year to year. November 1, 2026 and November 7, 2027 are the next scheduled rollback dates after 2025.
The change affects roughly 1 hour of clock offset. Travelers crossing time zones can experience larger adjustments. A flight from New York to Los Angeles in November crosses three time zones, so the local clock difference is 3 hours plus the 1-hour seasonal change.
Leap seconds are far smaller and far rarer. Since 1972, the International Earth Rotation and Reference Systems Service has inserted 27 leap seconds, most recently in 2016. Each adds one second to UTC to keep atomic time aligned with Earth's slightly irregular rotation.
The magnitude difference matters. A daylight saving shift is 3,600 seconds. A leap second is 1 second. They solve different problems: civil scheduling versus astronomical alignment.
| Adjustment | Typical size | Frequency | Cause | Who decides | |
|---|---|---|---|---|---|
| DST fall back | 1 hour | Annual | Statute and energy/social policy | National and local governments | |
| DST spring forward | 1 hour | Annual | Same | Same | |
| Leap second insertion | 1 second | Irregular, years apart | Earth rotation drift | IERS | |
| Time zone change | 1–12 hours | Per trip | Geographic boundaries | National governments | |
| UTC offset reform proposals | Hours or none | Proposed only | Policy debate | Legislatures |
Timeline of a fall-back night
The transition is not instantaneous in practice. Clocks, devices, and schedules update across a window of minutes to hours, and the order matters for anyone relying on timestamps.
| Local time (approx.) | Event | What changes |
|---|---|---|
| 1:00 a.m. DST | First pass through the hour | Clock reads 1:00 |
| 1:30 a.m. DST | Mid-hour | Systems may log duplicate timestamps |
| 2:00 a.m. DST | Statutory rollback point | Clock resets to 1:00 standard time |
| 1:00 a.m. standard | Second pass through the hour | Same labels repeat |
| 2:00 a.m. standard | Hour completes | Normal progression resumes |
| Morning hours | Devices sync | Network time servers update offsets |
| Following days | Schedules settle | Employers and transit adjust |
Networked devices usually update automatically through the Network Time Protocol, which carries time zone rules. Older appliances, wall clocks, and some industrial controllers require manual adjustment.
Edge cases that complicate the rollback
Not every US state observes daylight saving time. Arizona, except the Navajo Nation, and Hawaii stay on standard time year-round. Puerto Rico, Guam, American Samoa, and the US Virgin Islands also do not observe it. So "when does time go back" has no single national answer.
Federal law allows states to opt out of daylight saving, but current rules do not permit permanent daylight saving without congressional action. Several states have passed conditional laws that would take effect if neighbors or Congress acted.
The repeated hour creates genuine ambiguity in records. A timestamp of 1:30 a.m. on the fall-back date could refer to two distinct moments. Hospitals, financial systems, and logging infrastructure often store timestamps in UTC to avoid this problem.
Aviation and rail schedules publish in local time with explicit notes, but automated systems can still misread the duplicate hour. Software libraries that store local time without offset information are prone to errors on these dates.
Leap seconds create a different edge case. Because atomic clocks and Earth rotation drift apart, a leap second can in principle make a clock read 23:59:60. Most consumer devices hide this by smearing the extra second across a longer window. Financial exchanges and satellite navigation systems handle it with specialized protocols.
Misconceptions and a professional caveat
A common misconception is that daylight saving time saves substantial energy. Studies of national electricity consumption have found effects that vary by region and season, often small and sometimes in the opposite direction of the intended saving. The policy rationale has shifted toward traffic safety and evening recreation.
Another misconception is that clocks "go back" because of Earth's rotation or orbital mechanics. The seasonal change is purely a legal and administrative decision. Only leap seconds respond to astronomical drift.
The counterintuitive detail: the fall-back night is statistically associated with changes in sleep patterns, but the direction of health effects is not uniform. Some analyses report modest increases in certain cardiovascular events after the spring transition; the fall transition is often described as less disruptive. This is an active research area, and findings should not be treated as settled.
A professional caveat from timekeeping: UTC itself is defined by atomic clocks, and the relationship between UTC and Earth rotation is managed through leap seconds. There is an international effort to redefine the tolerance for that drift, which could change how leap seconds are handled in the future. The mechanism is stable today but not permanent.
Practical actions and red-flag warnings
The practical steps are simple but worth stating precisely.
Set manual clocks, including ovens, microwaves, and car dashboards, on the evening before the first Sunday of November. Verify that phones, computers, and smart home devices are set to automatic time zone updates. Check scheduled appointments and recurring meetings for the week after the transition. Update alerting and logging systems that store local timestamps without offsets. Confirm medication reminders and sleep schedules for shift workers and caregivers.
Red flags that deserve attention:
A critical system logs two different events with the same local timestamp on the fall-back date. A scheduled maintenance window silently runs twice because it was defined in local time. A trading or billing system produces duplicate or missing transactions around 1:00–2:00 a.m. Someone reports confusion about medication timing tied to a clock change. Industrial equipment with a manual clock shows an unexpected process start time.
For most households, none of these apply. For anyone running infrastructure or coordinating care, the duplicate hour is worth a specific check.
Limits of what science can settle
The physics of time is well established. The civil rules that govern clocks are not physics; they are policy. That means answers to "when does time go back" depend on jurisdiction and on legislative action that can change.
The health effects of the transition are studied but not fully resolved. Observational data can be confounded by season, latitude, and individual sleep behavior. Claims that the fall change causes or prevents specific outcomes should be read as tentative.
Leap second policy is under review internationally, so any statement about future leap seconds carries uncertainty. The mechanism for adjusting UTC exists, but its long-term use is a matter of coordination among timekeeping bodies.
What is certain is the distinction. Daylight saving is a scheduled, statutory shift of one hour, typically on the first Sunday of November in the US. Leap seconds are rare, one-second corrections managed by international timekeeping authorities. Neither changes the underlying flow of time; both change how we label it.
If you rely on precise timestamps, store them in UTC and convert for display. If you rely on a wall clock, check it the morning after the first Sunday in November.