by Jay Pscot
On the shortest day of this year the sun will come up over Rockwood, Tennessee at 7:45 in the morning and over Crab Orchard at 6:46. The two towns sit about eleven miles apart on the same interstate. The sun reaches them less than a minute apart. Their clocks disagree by an hour, because the boundary between Eastern and Central time runs between them.
Nobody there finds this remarkable. You learn to say which time you mean.
On July 14 the House of Representatives passed the Sunshine Protection Act by a vote of 308 to 117. It went to the Senate Commerce Committee the next morning, where it sits now. For the first time since 1974 the country is close to ending the twice-yearly clock change, and a supermajority of the House has said so on the record.
The bill would end it by putting the whole country on permanent daylight saving time. The American Academy of Sleep Medicine, which has spent years arguing the clock change should stop, condemned the bill inside a week. Its president said the House had “ignored the overwhelming evidence supporting the health and safety benefits of permanent standard time and instead chose to force dark, dangerous mornings on millions of Americans.”
That is the whole deadlock in one week of news. Everybody wants the change to stop. Nobody can agree where to stop it, and the disagreement has now killed the reform more than once. The point of contention? Sixty minutes in which direction.
The bill as written would also leave the country more fractured than it found it, for reasons that have nothing to do with sunrise.
The Senate should send the bill back for thirty minutes’ worth of change, and it does not even have to write one — the bill it wants has been sitting in a House committee since February. But more on that later.
I live in Tennessee, about a mile from that boundary, on the Central side. I keep my clocks on Eastern.
Tennessee runs more than four hundred miles end to end and the time zone line cuts it in two. The boundary follows the eastern edges of six counties. Sixty-five of the state’s ninety-five counties are Central; the remaining thirty are Eastern. Tennessee legally divides itself into three distinct geographical regions, East, Middle and West — the three stars on the state flag. Three of the counties on the Central side are part of East Tennessee by any measure except the time zone.
Locally the two are called fast time and slow time. It is a disambiguator rather than a joke, used by schools, businesses and county offices, because a meeting set for two o’clock without a qualifier is a meeting somebody is going to miss.
I reside in slow time and I live on fast time. The county courthouse is in Central. So is my dentist. Everything else I do — my work, what I watch, what I listen to, where I go, who I meet — runs on Eastern, so that is what my clocks say. I set them by hand.
Devices that update themselves are the difficult ones; a cell phone or Assistant device will look up where it is, decide it knows better, and quietly put me back on Central. Things that stay wrong on purpose take maintenance. My phone lets me select the time zone. The Assistant does not. My computers do. Some of the kitchen appliances do not. And so it goes.
Then there is calling Nashville. State government sits in Central time, which is also, legally, my time. An office that opens at 8:30 opens at 9:30 by my clock, and one that closes at 4:30 is open until 5:30 my time. I am in the same time zone as my state capital and I still do the arithmetic every time, because the statute and my life disagree about what time it is and I sided with my life.
I have lived in Oregon, Washington, Texas, Tennessee, Georgia, Florida and Ohio — Pacific, Central and Eastern, three of the four continental zones. Three of those seven states are themselves split by a time zone line. That is not unusual. Depending on how you count Arizona, somewhere between thirteen and fifteen states are divided, and the fact that reasonable sources disagree about the number is itself the point. Even the count is a convention.
What living on a line teaches you is that the national argument has two questions welded together. Whether the clock should shift twice a year is one of them. Where you sit inside your zone is the other, and it is the bigger of the two. People who live on a boundary stopped treating clock time as a fact a long time ago. Those who live near the east or west edge of the same time zone have very different perspectives on the clock relative to the sun.
The medical case against permanent daylight saving time is not soft and it should not be dismissed lightly.
The evidence that the transition itself does damage is about as good as population-level evidence gets. Janszky and Ljung, writing in the New England Journal of Medicine in 2008, found heart attacks in Sweden rising about five percent in the week after the spring shift. Fritz and colleagues, in Current Biology in 2020, went through 732,835 fatal American motor vehicle accidents from 1996 through 2017 and found the spring transition raising fatal crash risk about six percent.
Read the second study closely, though, because it says something the debate has not absorbed. The increase was concentrated in the morning, and it was larger for people living further west inside a time zone. The fall-back transition in autumn produced no such effect at all. Janszky and Ljung found the same asymmetry from the other side: the autumn shift, which returns an hour of sleep, is followed by a small drop in heart attacks.
Position within the zone shows up in the mortality data as a risk factor in its own right.
Solar time moves four minutes for every degree of longitude. Eastern time is built around the meridian at 75 degrees west, which runs down through eastern Pennsylvania. At Lubec, Maine, the easternmost town in the country, the sun peaks thirty-two minutes before the clock says noon. At Marquette, Michigan, on the far side of the same zone, it peaks fifty minutes after. That is a spread of about eighty-two minutes inside a single time zone — wider than the entire hour the country is fighting over.
Put it in sunrises. On the shortest day of the year, under permanent standard time — the option the sleep researchers are defending — the sun comes up in Marquette at 8:30 in the morning. Under permanent daylight saving time, the option they are calling dangerous, it comes up in Lubec at 8:02. The result?
The healthy policy produces a darker morning in Michigan than the dangerous one produces in Maine.
None of which makes permanent standard time the wrong choice. It is the best of the two by the measure its defenders care most about: the worst misalignment it inflicts anywhere in the Eastern zone is about fifty minutes, against a hundred and ten for permanent daylight time.
And the misalignment is not a theoretical quantity. Fritz’s crash data found the spring transition hitting hardest in the western part of each zone. A 2023 analysis of American county data reported higher suicide rates in western partitions. A broader literature, usually called the western edge effect, reports shorter sleep and higher rates of obesity and cardiovascular disease toward a zone’s western boundary.
That literature is not settled, and it would be dishonest to present it as though it were. An earlier finding linking longitudinal position to cancer incidence was reexamined in 2023 against more recent registry data and did not survive: the authors concluded that relative position has no significant effect on cancer incidence, and that their results challenge prior research and reveal numerous inconsistencies. The traffic and sleep findings stand on firmer ground than the disease findings.
The direction is consistent enough to take seriously, though, and it points where the sleep researchers have been pointing all along. Closer to the sun is better.
There is also a country that ran the experiment. Russia abolished its seasonal change in 2011 and stayed on permanent summer time — the same policy H.R. 139 proposes. Parents, teachers and doctors spent three years objecting to the dark mornings, and in October 2014 the Duma put the country on permanent standard time, where it has stayed.
That is the case for permanent standard time, and it is a serious one. It is also a case that describes a meridian, and nobody lives on a meridian. Optimal is what you get by averaging a country down to a line and then reporting the line’s experience as everyone’s. The eighty-two minutes do not go away when Congress votes. No statute moves anybody’s longitude. Whatever we choose, most of the country will be somewhere between twenty and eighty minutes out of step with the sun, as it is today, as it has been since 1918.
The coverage of H.R. 139 has described it as making daylight saving time permanent and letting states opt out. That is close enough for a headline and wrong in the way that matters.
The engrossed text does two things. It repeals Section 3 of the Uniform Time Act outright. Then it goes into the Calder Act of 1918 and performs a find-and-replace on the zone definitions: strike “5 hours,” insert “4 hours.” Strike “6 hours,” insert “5 hours.” Down the whole list. Eastern time stops being UTC−5 and becomes UTC−4, and the seasonal shift is gone because the summer setting is now the only setting.
The opt-out is the part worth reading twice. The new subsection lets a state keep the old offset only if it had already exempted itself from daylight saving before the Act takes effect. And the Act repeals the provision that lets states exempt themselves. The door closes behind it.
So there is a window, and it is a race. A state that wants permanent standard time has to pass a law before enactment. A legislature out of session, or slow, or not watching a bill in a Senate committee, does not get a second chance. It gets decided for them.
Now consider what the winners of that race look like on a map. Arizona is currently out of step with its neighbors for about eight months of the year and back in step for the other four. Under this bill, every border between a state that opted out and a state that did not is an hour wide in January and in July and every month in between.
I live on a line like that, and I want to be careful about how I say this, because I am not describing a hardship. It works. People manage. But it is a tax, it is paid almost entirely by the people nearest the line, and none of them voted to levy it. It costs a phone call to Nashville a little arithmetic. It costs school districts and hospitals and dispatchers considerably more.
The bill takes a country with somewhere around fourteen split states and offers to create a fresh set of permanent internal boundaries, in exchange for ending a clock change that a thirty-minute amendment would end as completely. The patchwork is the thing everyone says they want to fix. This bill does not fix it. It makes more of it, and it makes what it makes permanent.
What am I proposing?
Advance the clocks thirty minutes from standard time and stop. Eastern becomes UTC−4:30, Central −5:30, and so on west. The one hour intervals between American zones are untouched: the country’s internal map keeps exactly the shape it has now. The seasonal change is gone, because there is no longer anything to change to.
This is not a thought experiment, and here is the part that ought to be the story. Somebody already wrote the bill.
H.R. 7378, the Daylight Act of 2026, was introduced by Representative Greg Steube of Florida on February 4 of this year and referred to the Committee on Energy and Commerce. It is two pages. It amends the same sentence of the Calder Act that H.R. 139 amends, and differs from it by a half hour: strike “4 hours,” insert “3.5 hours”; strike “5 hours,” insert “4.5 hours”; down the list. It repeals seasonal daylight saving outright. It takes effect ninety days after enactment.
It also does something H.R. 139 does not. It contains no opt-out, because it does not need one — there is no state that has to choose between two permanent standards, and so there are no new borders. The patchwork that H.R. 139 would make permanent never gets created.
A thirty-minute offset is not better than permanent standard time. By the same measure the last section used — the worst solar misalignment anywhere in the Eastern zone — permanent standard time costs about fifty minutes and permanent daylight time about a hundred and ten. A thirty-minute advance costs eighty. It lands between them, which is what splitting a difference does, and permanent standard time still beats it. The half hour is not the healthy choice.
What it is, is the only proposal on the table under which neither side has to be photographed losing. The doctors give up thirty minutes from the alignment they want. The evening-light constituency gives up thirty minutes from theirs. Both get the thing they have each said, repeatedly and in public, that they want most: an end to the twice-yearly change.
The obvious objection is one I have already stated. If the country tolerates eighty-two minutes of solar drift inside a single zone without complaint, why is thirty minutes of policy worth an act of Congress?
Because the thirty minutes is not the prize. The spring transition raises fatal crashes about six percent; the autumn transition raises nothing. Heart attacks climb after the spring shift and dip after the fall one. The measured harm attaches to the change, not to either destination. The country has spent a decade arguing about the destination while the thing doing the damage goes on happening twice a year, on schedule, because the argument cannot be settled.
Thirty minutes is not what we gain. It is what the settlement costs, and it is priced so that both sides can pay it.
Late in the 1990s the company I worked for started getting complaints from its electronic trading partners. Our data was crashing their systems. It was, they were clear about this, our fault.
It was not our data. We ordered further ahead than anyone else in that network, so we were the first to send purchase orders dated 9/9/99 — and a good deal of the software receiving them treated “all nines” as a sentinel value meaning stop, or end of file, or nothing here. The assumption had been sitting in those parsers for years. Nothing had ever tested it, because until we came along nobody had sent a date from far enough out. I spent a while diagnosing other people’s defects and then explaining them, across a company boundary, to people who had every reason not to want to hear it.
I have been through most of the rest of them too.
The Y2K hysteria, which was a mix of real, severe technical risks and exaggerated media alarmism that drove public panic. The cause was the use of two-digit years and the in-built failure on January 1 of 2000. One solution was the windowing patch — 00-19 read as 2000s and 20-99 as 1900s — which did not fix the problem so much as move it, and which came due again around 2020. That’s when New York City’s fourteen thousand parking meters stopped taking credit cards on New Year’s Day, unpatched Splunk installations lost the ability to parse two-digit years on January 1 and Unix timestamps that September, and a wrestling video game shipped in 2019 would not run in 2020.
Ad hoc leap seconds, when time-keeping technology has to realign itself with Earth’s rotational reality.
Early remediation on the 2038 rollover, which is still coming.
In 2016 I spent a stretch not writing code at all but briefing executives and warning a thousand employees not to set their iPhones back to 1970, because doing so bricked them permanently.
That last one is worth separating out, because it was a communications problem rather than an engineering one, and most of what a thirty-minute transition would actually require looks more like that than like a code change.
The pattern across all of them is the same. The work was real and the schedule was ugly and the deadline held anyway. The public’s experience of every one of them was mostly the news coverage rather than the event.
There is a precedent closer than any of these, and it is not an analogy. In 2005 the Energy Policy Act moved the daylight saving dates, effective 2007. That is exactly the same class of change as a fractional offset: a legislated redefinition of local time, propagated through the IANA time zone database to every device on earth. The industry called it Y2K7, worried loudly, patched everything, and shipped it. The new dates took effect on March 11, 2007. You have been living under them for nineteen years and you did not notice the transition.
The objection specific to a half hour is that American software assumes whole-hour offsets. Some of it does. But the assumption is untested rather than load-bearing, which is precisely the pattern of the 9/9/99 problem — a thing nobody had exercised because no orders had yet been sent from far enough out.
We also have a live test of the actual question. Venezuela ran a thirty-minute offset for nine years, from 2007 to 2016. It was abandoned. Why is sometimes misunderstood. One research briefing asserts the fractional zone had wreaked havoc on international trade. It had not. Maduro moved the clocks back to a whole hour to get more evening daylight during a drought that had crippled the Guri reservoir, which supplies most of the country’s electricity. A nation ran successfully on fractional time for nearly a decade and gave it up over a hydroelectric emergency, not over a database or international trade.
Fractional offsets are not exotic in any case. India and Sri Lanka run on UTC+5:30, Iran on +3:30, Afghanistan on +4:30, Myanmar on +6:30, Nepal on +5:45, Newfoundland on −3:30, central Australia on +9:30, the Chatham Islands on +12:45. The global network carries all of them today without editorial comment. The whole hour is a convention somebody picked. The planet does not have an opinion about it.
And one place has been running the experiment itself. Lord Howe Island, in the Tasman Sea, keeps standard time at UTC+10:30 — half an hour off the Australian mainland, because that is what its longitude wants. Then, twice a year, it changes its clocks by thirty minutes: forward to UTC+11 on the first Sunday in October, back on the first Sunday in April. It is the only place that shifts by a half hour rather than a whole one, and it has done so since 1985, for exactly the reasons this argument turns on. The half-hour offset matches its position relative to the sun. The half-hour shift puts it level with Sydney through the summer, when it has business with Sydney.
Four hundred and forty-five people live on Lord Howe Island, and that matters: it is an existence proof for the mechanism, not for the scale. But the mechanism is the part that was supposed to be impossible. Every operating system carrying the Australia/Lord_Howe zone already computes a fractional transition correctly, twice a year, forty years running. The code path is not hypothetical. It is in production and it runs on a schedule.
The hard part is timestamps stored without a zone — local wall-clock time in a column with no offset, which is a great deal of enterprise legacy data. What happens if the 30 minute shift occurs? A company may already assume such dates and times are in US/Mountain time. In that case, nothing changes, Mountain time is still Mountain time. Maybe they adjust displays so users in Fresno or Buffalo see times in their own context. That doesn’t change as the hours difference is still the same. What does change is if the Toronto office wants the same treatment. Toronto and US/Eastern are no longer the same. Or if the application stored future events as UTC — a 9:00 AM Eastern job persisted as 14:00 UTC carries the old offset baked into it, and fires at 9:30 local the day the offset changes. The intent was nine o’clock local. What got saved was arithmetic that stopped being true. That means coding work, and finding every such row is most of it.
Newer programs rely on time zone tables, so the issues are already mitigated. It is the sea of older code, with what are now understood to be bad habits, that are the problem. This was the Y2K problem. Not the then-current programs so much as old COBOL and other legacy code. Correcting it means a large, risky migration in production. It is real work and some of it will go badly. Most of it will not.
It is work those shops need to do regardless. A system that stored local time without an offset has been broken since it was written; it has not yet been asked a question that exposes it. That is the untested assumption again, in a different column. Current conventions call for storing future times, not with the actual offset value, but rather the tzdb label. Historical dates use the actual offset, as that represents the actual time at the time.
And it is worth saying plainly what the machinery on the other side of the ledger looks like, because the debate keeps describing it as fragile. Time zone tables update as a matter of routine. NTP corrects drift continuously and invisibly. GPS disciplines itself. The apparatus for pushing a redefinition of local time to billions of endpoints is mature, ubiquitous, and running right now for other reasons. The device on my desk is so aggressive about it that it overrules me, from orbit, about what time zone I am standing in.
That is not a system too brittle to absorb thirty minutes.
There is a cleaner solution than the one I am proposing, and it is worth describing because it establishes the scale of what is actually being asked.
Redefine the reference. Pick a date, move Coordinated Universal Time thirty minutes, and every interval on earth is preserved exactly. New York and London stay five hours apart. Nothing about international scheduling changes at all, because nothing about the relationships changes — only the label everybody is using, changed everywhere at once.
We have absorbed far worse. In September 1752 Britain and its colonies moved from the Julian calendar to the Gregorian and deleted eleven days; September 2 was followed by September 14. The riots that story usually comes with appear to be largely a myth, traceable to a satirical magazine and a Hogarth painting. What actually happened is that people went to bed, got up eleven days later, and got on with it.
That option is not available, and I want to be honest about why rather than pretend it is a technical problem. The United States is powerful enough to wield a great deal of influence and not nearly powerful enough to reset the world’s clock by itself. And it should not try. A good part of the world would follow us, another part would want the thirty minutes in the other direction, and the rest would want no change or some other change, and all would be right to want what they want.
So the fantasy stays a fantasy, and what is left is the small version: one country, half an hour, everybody else untouched.
The rest of the world will no doubt react. I won’t presume to predict what that looks like.
The Sunshine Protection Act will probably die in committee. That is what happened in 2022, when a version of it cleared the Senate by unanimous consent in March and expired without ever reaching a House vote, and the reason it will happen again is not that anybody wants the clock change. Nobody wants the clock change. It will die because the only replacement on offer is the one the medical establishment has committed itself to opposing, and because a bill that hands one camp a total victory gives the other camp nothing to do but block it.
Meanwhile the spring transition arrives in March, on schedule, with its six percent.
And nobody has to draft anything. The Committee on Energy and Commerce has had the Daylight Act since February. It had the Sunshine Protection Act too, and in the spring it moved that one toward the floor, where it passed in July with three hundred and eight votes. The other bill is still sitting there. Same committee, same sentence of the same 1918 statute, a different number in the blank. Only one of them got a vote.
The half hour does not require anyone to concede the science, and it does not require anyone to concede the evenings. It asks both sides for thirty minutes, which is less than the accident of geography already charges most Americans, and which nobody would be able to feel a week after it took effect.
On both sides of the line where I live, they have been keeping two clocks for over seventy years. My 11:15 AM dentist appointment is at 12:15 PM my time. I have to leave my home by 11:55 AM to get there on time at 11:15 AM. My sun came up at around seven this morning. My neighbor’s showed itself around six. Same sun, same moment, different clock. In terms of life, neither has ever been the hard part. The line has to be somewhere; however, there is one thing we all agree on — we would love to stop resetting the clock twice a year.
© 2026 Jay Pscot
All rights reserved.
editor@pscot.com
www.pscot.com