Save Daylight Savings Time – by Nate Silver

Last week, President-elect Trump pledged to “eliminate” Daylight Savings Time, which he called “inconvenient, and very costly to our Nation”. The idea may have been inspired by DOGE, the Department of Government Efficiency, an agency set to be run by Elon Musk and Vivek Ramaswamy, after Musk and Ramaswamy pitched a similar plan earlier this month.

I suppose I’m not anti-DOGE, though it’s hard to say because its structure and its mandate are unclear. I’d certainly like for it not to cost $2.5 billion to build a single mile of subway track in New York, for instance. But without having any real teeth to implement its policies, DOGE may focus more on gimmicks — and will run the risk of violating the precept of Chesterton’s Fence by changing things that are perfectly fine or where the current rules are in place for a reason.

Daylight Savings Time is a perfect example of this. Eliminating it would deprive the average American of 40 minutes of waking daylight in the summer months. This is a terrible idea.

Just to be explicit about what we’re talking about — because I know it can be confusing exactly which is “DST” and which is “standard time” — Daylight Savings Time pushes the clock forward by an hour in the summer months, along with the spring and early fall. (This year, DST was in place from March 10 to November 3.) For instance, with DST — the status quo — the sun rises in New York at 5:24 a.m. and sets at 8:31 p.m. on the longest day of the year in June. If we reverted to year-round standard time as Trump proposes, those times would shift forward an hour, meaning sunrise at 4:24 a.m. and sunset at 7:31 p.m. No more of those glorious summer evenings spent at Citi Field or Yankee Stadium where it’s still light out until the sixth inning.

True, the twice-annual time changes — you know the drill: spring forward, fall back — are annoying and associated with some tangible consequences, such as an increase in traffic accidents because of sleep deprivation. But to be honest, those consequences are pretty minor. One estimate is that they only amount to about 30 extra traffic deaths per year in a country of 335 million people. And we’re talking about just one hour, something most people can take in stride. Let’s say you’re living in New York, and a friend who popped into town cancels plans to hang out because they complain of jetlag. If they’re arriving from Paris or Hong Kong — sure, total pass. West Coast? Fine. But what if they’ve flown in from Chicago, just one time zone away? You’d think that person was an incredible loser or was just making an excuse because they clearly didn’t want to hang. If we change a program like Daylight Savings Time that’s been in place for almost 60 years, we need a better reason than that.

The Trump/Musk/Ramaswamy proposal should not be confused for its counterpart: year-round Daylight Savings Time, which has also been proposed recently by Florida Sen. Marco Rubio and Massachusetts Sen. Ed Markey. Some Republicans are apparently hoping Trump is confused. Sen. Katie Britt of Alabama, usually deeply loyal to Trump, sent a tweet on Monday that appeared to nudge him in the direction of year-round DST:

This is much closer to being a good idea. Year-round DST would at least increase the amount of time that the average American spends in daylight in the winter but at the cost of ignoring some strong revealed preferences for people not wanting to wake up in darkness for required activities like work and school. On the shortest day of the year in New York in December, the sun rises at 7:20 a.m. under standard time but sets at 4:28 p.m. With year-round DST, those times would shift forward to 8:20 a.m. and 5:28 p.m. Even as a relative Night Owl, I’m not sure I’d want such late sunrises.

Spend any time looking at the data, though, and you’ll see that eliminating summer DST is an awful plan — New York is hardly atypical in having summer sunrises before 5 a.m. under standard time.

Let’s start by asking a basic question: when are Americans awake? There’s actually data on this in the form of the American Time Use Survey (ATUS), a Bureau of Labor Statistics project that calls a sample of Americans and asks them to log how they spent their previous day. Here’s what the data said in 2023:

The average American claims to be sleeping from 10:06 p.m. to 6:42 a.m. I say “claims to be” because that’s a lot of sleep — more than eight-and-a-half hours. Data from sleep-tracking software suggests Americans get less sleep than that: an average of 7 hours and 19 minutes. It may be that people count all hours lying in bed or attempting to sleep as “sleep” — lots of people are in the habit of fiddling with their phones until they pass out, or wake up for stretches in the middle of the night. There may also be some element of social desirability bias in the ATUS: you won’t want to admit that actually, you were binge-watching NHL goalie fight videos until 3 in the morning. But anyway, we’ll run with the ATUS numbers.

For this story, I’ll look at a sample of 27 American cities. These include the 11 largest metro areas, plus other medium-big metros I’ve chosen for geographic coverage, plus some smaller cities like Bangor (Maine), Boise (Idaho), Traverse City (Michigan) and Pensacola (Florida) that I’ve selected because they represent edge cases of various kinds. East Lansing, Michigan, for instance — where I grew up — has very long summer evenings because Michigan is on the western frontier of the Eastern Time Zone and should really be in Central Time. This is even more pronounced in Traverse City, north and west of East Lansing.

This map, adapted from a Reddit post, shows the location of 24 of our cities in the continental U.S.; not pictured are Anchorage, Honolulu, and San Juan, Puerto Rico, but I’ll include them in the data tables.

This map reflects the current situation under DST, showing the offset relative to Solar Noon — basically the difference between observed time and each location’s “natural” time zone. Locations shaded in red have late sunrises and sunsets relative to solar time, and locations in green have early ones; yellow is “just right.” Arizona is particularly green, for instance, because outside of Navajo Nation, it doesn’t use DST; nor do Hawaii and Puerto Rico. Overall, however, there’s a pretty good balance of green and red under the status quo with DST. Eliminating DST would undermine this and make the entire map greener.

Around the world, cities and countries are offset from their natural time zones for various economic and political reasons. For example, all of China is in a single time zone, even though the country spans about as much territory from west to east as the Continental U.S. But in general, around the world, the bias is unidirectional: states and countries adopt a more easterly time zone (making them redder on the map above) than they naturally would have, giving them later sunrises and sunsets than solar time would dictate. And then on top of that, North America and Europe almost universally use Daylight Savings Time, which sets the clock back further during the summer months.

Boise, Idaho is a good example of this. It’s on Mountain Time, even though it arguably “should” be on Pacific Time — it’s actually west of Las Vegas, which is on Pac Time. But suggest that it should move to Pacific Time, and Boiseans will get real mad at you. Let’s look at its options:

Under the status quo (#4), the sun rises just after 6 a.m. on the longest day of the year in Boise and doesn’t set until 9:30. If Boise was “astronomically correct” and was on Pacific Time and observed standard time year-round (option #1), sunrise would come at 4 a.m. at the summer solstice but sunset at 7:30 p.m.

But Boiseans have more common sense than Trump, Musk and Ramaswamy and recognize that having the sun rise at 4 a.m. is a waste because almost everyone is still asleep then. They are willing to shift essentially two time zones over from solar time to avoid this: a city that “should” be in Pacific Time instead operates on what’s essentially Central Standard Time in the summer. (In the winter, there’s no DST so Boise is only one hour offset from its natural time zone, operating on Mountain Standard Time.)

Looking at its menu of options, you can see why Boise chose this course. Under year-round Central Standard Time (#5) — that is, no DST — Boise would get the sunrises it likes in the summer, but sunrise at the winter solstice wouldn’t come until 9:19 a.m. after most people have already been awake for a couple hours. Year-round Pacific Standard Time (#1) is likewise a non-starter with those 4 a.m. sunrises. Options #2, #3 and #4 are all basically reasonable. All of Idaho could go Mountain Time (the state is currently split between two time zones) but not observe DST (option #3); this is what Arizona does, as I mentioned.

However, by implementing DST for part of the year, Boise essentially gets to have its cake and eat it, too. It’s still a bit unusual for tolerating 8:18 a.m. sunrises. But given its northerly location, it has to make some trade-offs: year-round Pacific Time with DST would mean 4 p.m. sunsets in December. Taking the Arizona route isn’t terrible, but you can see why Idaho chose the path it did (#4). By putting up with (the horror!) two one-hour time shifts each year — one of which actually lets you sleep in — it can manage these other trade-offs well; it’s probably the least-worst option.

To suggest something heretical, in northerly locations, you could even argue for having a clock shift of more than an hour. For example, if Boise joined Pacific Time but moved its clocks forward by an hour-and-a-half in the summer — option #6 in the table — it could have summer daylight until 8:30 p.m. but winter sunrises at a more reasonable 7:48 a.m.

Granted, that isn’t going to happen, and a time change of more than an hour would be overkill in more southerly locations, which don’t face the same scarcity of winter daylight. Let me go ahead and show you the current sunrise and sunset times in all 27 cities as currently observed on the longest and shortest days of the year. I’ll also show you something else: the average time at which workers in the metro area associated with each city begin their morning commutes as according to the Census Bureau:

Commuters do show some preference to shift their schedules based on how their time zone lines up relative to solar time. Commutes begin notably earlier in the three cities (Phoenix, Honolulu and San Juan) that don’t use DST. They don’t begin until 8:29 on average, conversely, in Traverse City, Michigan, which is very late-shifted because it “should” be in the Central Time Zone. However, commutes start at 7:45 a.m. in Bangor, Maine, so far east that it “should” really be in the Atlantic Time Zone like Nova Scotia. Having spent a fair amount of time “Down East,” you can see this in action: Maine gets up super early, but it can be hard to find a restaurant open past 8 p.m. for dinner. In other cities, though — most notably, Las Vegas, where the average commute doesn’t begin until almost 9 a.m. — lifestyle factors or the nature of the dominant industry in the city dominates other considerations.

Overall, though, most places demonstrate similar revealed preferences. They select their time zones and manage their use of Daylight Savings Time such that the sun comes up at ~7:30 a.m. even on the shortest day of the year, and then begin their commutes about 30 or 45 minutes later — just enough time to have breakfast and send their kids to school. But then they opt to avoid wasted, too-early daylight in the summer by adopting DST — and in some cases also by shifting over to a later time zone than solar time would recommend.

As currently observed, the 27 cities have an average sunrise at 5:39 a.m. and sunset at 8:37 p.m. on the longest day of the year. Take away DST — affecting all cities but Phoenix, Honolulu and San Juan — and the average sunrise would come at 4:46 a.m. and the average sunset at 7:43 p.m. Per the ATUS, only 14 percent of Americans are awake at 4:46 while 93 percent are still awake at 7:43 p.m. So you’d take away daylight at a time when nearly everyone is awake while adding it when nearly everyone is asleep.

People would probably adapt to some extent by adopting earlier schedules; some states might even switch time zones. But without wanting to make this too “political,” this is a highly unconservative proposal from Trump and Musk. It’s the technocratic state imposing preferences on people for the sake of alleged efficiency or convenience: literally what Communist China does.

Share

The Markey/Rubio proposal for year-round DST would also cause some problems. The U.S. actually adopted year-round DST on an experimental basis in 1974 — but it was extremely unpopular and the trial was ended early. Under current clocks — states may opt out of DST but cannot opt into year-round DST, so everyone is on standard time in the winter — the sun rises at 7:37 a.m. on average in the 27 cities on the shortest day of the year, and the sun sets at 4:45 p.m. With year-round DST, you’d get 5:45 p.m. sunsets but 8:37 a.m. sunrises.

I’ll go into more detail on how winter would be affected by year-round DST in a moment. But let’s return to the summer since that’s the higher-stakes question. You’re just taking away sunlight from people with no real benefit other than avoiding the twice-yearly time change.

A lot of sunlight, it turns out. I calculated, using the ATUS data on sleeping and waking times and sunrise and sunset tables in each city, how much daylight each American would receive under standard time and DST throughout the year. Here’s the data for Chicago, for instance. There’s also one other twist: I’ve shifted each city’s waking and sleeping times over from the national averages based on its average commute times. For instance, Miami gets off to a late start, with commute times beginning about 15 minutes later than the national average; I assume this time shift pertains to waking and sleeping hours, too.

In Chicago, eliminating DST would cost the average resident more than 45 minutes of waking daylight during late spring and early summer. Year-round DST would also give people more daylight in the winter — in fact, this is true for all 27 cities for all 366 days of the year. However, the differences are much less: the average Chicagoan would gain about 15 minutes of waking daylight in the winter under year-round DST. Losing that quarter-hour is a much more reasonable price for not wanting to commute or send your kids to school in the dark. (I’ll ignore the fact that school should start later: that would be a better DOGE initiative.)

Here’s this data for every city at the summer solstice, which was on June 20 this year. In every city except Anchorage — Alaska has more sun than it knows what to do with in the summer — residents get at least 30 more minutes of waking daylight with DST when without it. The biggest differences are in Bangor and Boston since they’re so far east — Maine, as I mentioned, should probably be in the Atlantic Time Zone. There, residents would lose an average of 49 and 48 minutes of waking daylight, respectively, under Trump’s plan. The less obvious case is Las Vegas — 47 minutes of lost daylight — because the city is so late-rising.

This is a very large amount of daylight to lose. In addition to considering people’s revealed preferences for daylight, traffic accidents occur much more frequently at night and are more fatal. Since one of the ostensible purposes of avoiding the spring and fall time changes is to prevent accidents associated with groggy drivers — but the actual number of deaths associated with “spring forward, fall back” is very small — it’s hard to imagine it isn’t outweighed by an order of magnitude or two by depriving Americans of ~40 minutes of waking daylight during the summer. Crime also rises at night, until it declines sharply at around 1 a.m. because so many people are in bed. (Certain crimes are hard to commit against people when they’re asleep and safe at home.)

Overall, this is just not a close call. Eliminating DST might even be a political risk for Trump and Musk. Big, top-down-imposed changes that nobody is asking for check all the boxes for an unpopular policy. It would be a highly noticeable change that would affect Americans of all social classes. If people are used to it being bright out at 7 p.m. in the summer — whether for their kid’s softball game or long evening jogs on the beach or al fresco dining in the West Village — they’re going to be very annoyed if that gets taken away.

Here’s the data for the winter solstice:

Year-round DST would produce more waking daylight in the winter too, although the differences are much less — about 15 minutes per person rather than 40 minutes in the average city. Here, however, the benefits of earlier sunrises are more obvious for both biological and sociological reasons. Waking up in darkness messes with people’s circadian rhythms. There’s a lot to be said for at least giving people the option to avoid that since work and school begin toward the start of the day and are the “mandatory” part of a person’s schedule that they can’t avoid.

Let’s say we place a premium on AM daylight; I think that’s pretty reasonable. I ran the numbers and calculated what the optimal start and end dates would be for DST to maximize waking daylight on average in the country given various “morning multipliers” — that is, how much more we value morning sun than evening sun.

• With a morning multiplier of 1.5 — that is, we weigh AM daylight 50 percent more than PM daylight — DST would begin on February 22 and end on November 16, actually longer than under the status quo.

• With a morning multiplier of 2x, DST would start on March 26 and end on September 30.

• Even if we value morning daylight at 3x evening daylight, we’d still have a 3+ month period of DST from April 27 through August 5. Only a morning multiplier of 4x or higher would call for a year-round standard time as Trump proposes.

So I say keep the status quo — or make only minor tweaks, like letting states opt into year-round DST, which isn’t currently allowed. The twice-annual time changes aren’t such a big burden. Year-round DST is debatable, but year-round standard time is terrible. DOGE can shove it where the sun don’t shine.