Is the speed of skyscraper construction declining?

Jose Luis Ricon writes:

…taking not averages but just the fastest building built every year (again taking out Pyongyang and Broad Group), regardless of country. This seems to indicate that with the exception of The Belcher’s tower(s) in HK and 60 Wall St, in general construction speed has remained stagnant for almost a century, and has actually declined since its peak in the Great Depression.

The early 1930s look pretty amazing.  The data on meters built per year, and the fastest built buildings by that standard are interesting too:

There is much, much more in this post, which I consider to be one of the best things written this year.  Here is Artir’s conclusion:

…the Empire State Building was an impressive achievement compared to the present, but it’s not true that the past has been better than the present; rather, the skyscrapers built during the Great Depression were.

Highly recommended, via Patrick Collison.  And here is Artir on Twitter.


I agree with the author that it is complexity.

Parts count is a good measure of complexity of design. I bet parts per square meter climbed substantially for buildings. They sure did for the airplanes mentioned.

I think you are right - people add more features over time. It is sort of like computers - clock speed goes up but the number of features go up so overall speed remains the same.

Maybe another factor is the quality of people managing the project - perhaps the Empire State Building was able to hire the most capable people as a very prestigious project, now the profession of building is a lot less prestigious and there are many more being built, so the great people are more spread out.

I have to thank you for the efforts you've put in writing this website.
I'm hoping to see the same high-grade blog posts from you
in the futute as well. In fact, your creative writing abilities has encouraged me to get my own blog now ;)

Also safety is a factor; way back when they used guys who did steel girder work without a safety harness. These days safety harnesses are mandatory (sadly, even for dare devils like that Christian guy that walked across Niagara Falls).

I'm backing the safety factor as well. My totally unresearched guess is that this is why the North Korean buildings went up so quickly yet more recently.

All of these explanations have some truth to them, no doubt.

But still, it seems deeply problematic that after more than 80 years, during which technology has advanced greatly, we are building things slower. And this is just the construction phase. I suspect it looks even worse if you add planning, permitting, etc.

Not being able to build infrastructure quickly and efficiently imposes huge costs, both visible and invisible.

LOL, including North Korean buildings in the dataset!

I'm not sure what conclusions we can draw from time-to-build data, because it's a highly endogenous variable, in some ways dependent on the builder's choices and in some ways dependent on factors outside of their control.

I lived several blocks from a Taco Bell, that closed and was quickly demolished several years ago. And then the lot sat there for several months with nothing happening, and then some construction activity started happening.

I was eager to see what they would build there (the construction signs did not say); the neighborhood is mixed but basically middle class with a little bit of gentrification happening ... maybe they'd build a bistro or a coffeehouse on the site.

But after some initial excavation and foundation laying, construction went very slowly, sometimes halting entirely. IIRC it took well over a year, maybe almost two years, until they finished the new building and had the grand opening.

It was ... another Taco Bell.

We know they could've built the Taco Bell faster than that. Why didn't they? Who knows, but whatever those reasons were, I think they create a huge identification problem for someone trying to analyze time-to-build statistics. What do those numbers even mean? Regulatory issues, management issues, ownership issues, macroeconomic issues, economic choices, subjective feelings of urgency (the race to the moon) ... they all affect the speed of construction. If one building was built more quickly than the other, can we get anything useful from that information?

Which might be what Artir is trying to say in his article. Patrick Collison's question is not well formed and seems to be based on not just faulty data but a faulty sense of what the data can tell us.

There has been quite a bit of construction of various kinds in my area over the last 15 years; highways, schools, fairly large public buildings, low rise commercial space.

I’ve often wondered at the economics that (I presume) drive the appearantly slow pace of construction. One would think that at least for the commercial space, there would be a significant value in rapid completion to leasability once construction started.

A new water main to augment the one that William Mulholland built in 1915 to bring Owens Valley River water to Los Angeles has been going in across the San Fernando Valley. The block near my house was under construction for about 8 years, but is now finally finished. For much of that time, the construction site mostly appeared to be used as a parking lot for million dollar pieces of equipment.

Mulholland built his original water main in something like 15 months, using pick axes and mules.

Voters voted to hike taxes back in those days.

If taxpayer voted for bigger bond issues (requiring higher taxes to repaying the bonds), construction would go much faster.

When Obama was elected, the GOP actively worked to kill jobs by cutting spending on infrastructure. Capital does not vanish, just go down in price, so the now idled capital was parked near your home waiting for money funding paying workers to use the capital.

From the 20s to the 70s, paying workers to work was a popular public policy. Voters voted to hike taxes to pay workers to work, almost all the workers in the private sector.

In the 60s, the school's I attended were overcrowded, so the elections meant debate over voting for bond issues that always came with tax hikes to pay to build new bigger schools. Or new roads, water and sewer, for new homes and businesses. A big issue was where the private construction companies getting paid were. Marion was small, so often businesses in Indianapolis, Muncie got the construction contracts. By the 70s, that argument ended support for hiking taxes to fund bonds, based on no local benefit for tax hikes. The result was an end to growth, loss of jobs, and Marion's population growth reversed into population loss.

Note, Trump promised the Midwest lots of construction jobs funded by cutting taxes. Workers know government creates jobs to grow the economy, but workers have been taught that someone else will pay for the jobs. Unfortunately, economies are zero sum: workers always pay workers to work. Tanstaafl so Trump and the GOP has failed to pay workers to get to work building needed infrastructure.

Instead, Trump and the GOP have rewarded not paying workers by cutting taxes on the money not paid to workers. And as Milton Friedman argued circa 1970, lower business profit taxes will kill jobs and lower worker pay, and reduce the building of capital, so returns on scarce capital will go up generating high economic profits.

By delaying building of water lines, the existing capital with access to water will be able to charge higher rents, transferring money from workers to owners of old capital.

Ditto here in what is routinely-described as the fastest-growing city in the country. A Wendy's at one of the busiest, most high-profile intersections in the city went out. Then a fence went around it, five or six years ago seems like, announcing the coming of a Starbucks. The Wendy's was very slowly stripped and gutted (and graffiti-ed). Ah, they want to avoid code hassle by sticking to the footprint, doing a remodel, I supposed. As time went on, more and more of the Wendy's was removed. Finally, as of a month ago, a whole new Starbucks building is in place.

That things seem to happen even more slowly here on high-dollar land is not surprising, but that it would be a Starbucks that took years to get going seemed odd.

Buildings? Where we're going, we don't need "buildings."

Next you're going to tell me that they passed more amendments then than they do today.

Consider the SF Bay Bridge. Built in the 1930s in 3 years. Half replaced 50 years later in 25 years. It was damaged in the 1989 earthquake, and the replacement opened in 2014, with significant problems. Construction has resisted evolution like no other institution save...government.

I think it's government that has caused construction to slow down. More permitting, more environmental impact statements, more rules and regulations, more union work rules, more of everything.

'I think it's government that has caused construction to slow down.'

For some definition of government - because oddly, this vast edifice was constructed very quickly too.

'General George C. Marshall, the Army’s chief of staff, turned to Brigadier General Brehon B. Somervell, head of the Army’s Construction Division, for a solution.

Somervell’s proposal was audacious: a headquarters big enough for 40,000 people, with 4 million square feet of office space. A building this large could not fit in Washington, so Somervell chose a site across the Potomac River in Virginia, just east of Arlington National Cemetery.


The House of Representatives passed the necessary legislation for the project on July 28, 1941; the Senate on August 14. By that time, however, controversy had arisen over the scale of the building, as well as its location so close to the hallowed ground of Arlington National Cemetery.

Moved by the protests, Roosevelt declared that the project should be moved to a site three-quarters of a mile south of Arlington Farm, adjacent to Washington-Hoover Airport. He also directed Somervell to reduce the size of the building to no more than 2.25 million square feet.


Construction on the Pentagon began without fanfare on September 11, 1941. By early December 1941, 3,000 workers were on the site during the day, but construction was still behind schedule. Their supervisor was Corps of Engineers Colonel Leslie R. Groves, who would later be chosen to head the Manhattan Project and build the atomic bomb.

On December 7, the Japanese attacked Pearl Harbor, and the acceleration of the U.S. move toward war gave Somervell free reign to expand his project. The already-tight construction schedule was moved up, and by March 1942, more than 10,000 people were working on the site. At one particularly intense stage, 15,000 people were working three shifts, 24 hours a day, with floodlights illuminating the site at night.

The Pentagon’s first employees moved in on April 30, 1942; the building officially opened on January 14, 1943. Its massive bulk—6.24 million gross square feet—held 410,000 cubic yards of concrete, made from some 700,000 tons of sand dredged from the Potomac River.

From the $35 million Somervell had originally proposed, costs had ballooned to $75 million, though some claimed it was even higher.'

Admittedly, as one can see, efficiency was ignored, in typical bloated government style, using a national security justification for all the undoubted waste - like running shifts so as do construction over an entire 24 hour period.

These days, we would probably just rent some already available property - or buy some undeveloped land. As happened with the FDIC and the GMU Foundation, showing just how efficient the free market can be when it involves taking advantage of taxpayers - 'The George Mason University Foundation later sold approximately half of the land on the western side of the parcel to the Federal Government for nearly five times the amount it paid for the entire property. On it the Federal Deposit Insurance Corporation built its L. William Seidman Center.' Trump may be a real estate magnate, but he clearly has little of the skill of Northern Virginia real estate players like Hazel when it comes to making money.

heat and noise insulation, fire control, no ethernet cables during the 1930s.

I'd be more interested to look at the efficiency of capital use instead of absolute construction time.

The Trump Building? Do you mean 40 Wall Tower? Or some other Trump building? Buildings can change names every few years, which is why architects and historians usually use the original name.

Wikipedia has a detailed description of many clever techniques used to make the construction of the Empire State Building so efficient:

But it doesn't offer an overarching explanation for its incredible speed of construction. My guess is that NYC skyscraper construction was one of the world's most exciting industries before the 1929 stock market crash and attracted much talent. Builders evolved ever more efficient methods during the 1920s working on ever taller buildings.

But few tall buildings were constructed for about 15+ years after the Empire State was finished in 1931, and a lot of these meta-skills of how to organize such a vast project had decayed by the time construction got going again at the end of the 1940s.

I'd like to see a movie set in a alternative timeline where the stock market didn't crash and there was no WWII and the glories of 1920s Manhattan just kept going, getting bigger and more self-confident. The closest thing to this is "Sky Captain and the World of Tomorrow," although I'd like to see a movie portraying a more peaceful world.

"The Empire State Building only started becoming profitable in the 1950s, when it was finally able to break even for the first time."

Perhaps paying for 2 twelve-hour shifts or building specific infrastructure to deliver construction materials affected the profitability
of the project.

Perhaps what's not existing today are investors that go after faster and expensive construction instead of profits.

Until Khrushchev gets The Bomb first in 1960...

Seriously, it would be nice to see a non-dystopian urban future for a change. At least one that doesn't make you want to vomit like Star Trek.

Right. I love 1930s architecture and golf course design, but very little of it exists in the US outside of Tulsa, Oklahoma, which had an oil boom in the 1930s. Eritrea has some spectacular Italian moderne 1930s buildings that didn't get blown up in WWII.

You can see the kind of buildings that Americans would have built in 1935 in sets for Fred Astaire movies, but that's about it.

But post-1945 modernism seems to be motivated largely by self-loathing. This is not to say we didn't deserve it, but it took a long, long time for buildings to express some pride again. I can recall going in Trump Tower around 1984 and saying to myself, "The long postwar malaise is OVER."

I don't think it was self loathing that drove the post war pause in building styles - it was that WW2 was the triumph of the state controlled economy. This meant efficiency rather than ornament, because everyone understood there was a higher purpose - defeating the enemy. That solidarity around a common higher goal felt so natural to that generation, just read books set in that period.

Apples, oranges, tangerines, apricots, pears, tomatoes. I guess until I read the article, I shouldn't presume that the author is comparing disparate processes. But isn't he? What does it tell us? Nothing - we already know not only are the manufacturing techniques, equipment, and materials DIFFERENT today, but so is the surrounding environment.
Oh, also, shouldn't the metric be space rather than height per unit time? "occupiable" space? (there's probably a term for that, but it doesn't come to me) And then we get into mean or median, etc. Although it does get me to thinking about what the Empire State Building's lifetime is. When will it cease to be a safe structure?

Construction of the Empire State Building seems to be particularly sturdy, so barring terrorist attack or war, it's lifetime would probably be a few centuries. Earlier New York skyscrapers like the Woolworth Building 1912 and Flatiron Building 1902 are still standing.

In 2007 the American Institute of Architects published a survey of the public's favorite American buildings. The Empire State Building was #1 (I'd be partial to the Chrysler Building at #9). The only No-Longer-Existent buildings on the list were WTC, Penn Station, and old Yankee Stadium, along with parts of the Astrodome. (I miss old Comiskey Field in Chicago.)

Here's a list of 10 no longer existent buildings of note:

IIRC, a twin engine B-25 bomber flew into the Empire State Building during WW2 during heavy fog, with only minor damage to the building.

One reason that the top two buildings on the list were built in New York just a year or two apart is that they were part of the Race to the Sky in the late 1920s along with the Chrysler Building to be the tallest building in the world. 40 Wall Street (now the Trump Building), the Chrysler Building, and the Empire State Building were each in turn the tallest building in the world. Speed of construction was important for the buildings to lay claim to that title because an even taller one might come along soon and if they dallied in building, they might never even briefly hold the coveted title.

By the way, the mighty helmet on top of the Chrysler Building was secretly assembled inside the building and then raised into place to take the title from 40 Wall Street on October 23, 1929, the ultimate day of the Roaring 20s. Unfortunately, the next day the stock market crashed.

2017 set the record for the number of new skyscrapers at 144 (including 15 super tall buildings reaching at least 980 feet). And they are being built in more places (69 in 2017), from NYC to Jakarta to Shenzhen to Pyongyang (North Korea). An in more small towns, in China but also in Africa (including one in Nairobi, Kanya). Africa is likely to get many more skyscrapers, including a 300 meter hotel in Kenya that will be completed in 2019. And 2018 is expected to beat last year's record, with 160 completed this year. Here's my guess is to why it's taking longer to build them: they are taller and are being built in new places without experienced labor for such projects. All those skyscrapers were built faster during the Great Depression because of the experienced labor: build one, then start on another next door. Of course, all these new skyscrapers being built in far flung places might amount to little more than vanity projects.

Skyscrapers are about urbanization. My observation is that many of these places are building skyscrapers in order to promote urbanization, not in response to it. Then there's my frequent observation about looking to rising asset prices to achieve prosperity as opposed to investment in productive capital, in this case rising real estate prices. All these vanity projects are evidence of too much capital chasing higher returns through inflation in asset prices. It's a leap of faith; and it will not end well. I suppose that if prices do collapse, the leap of faith may end up in an altogether different type of leap.

"The early 1930s look pretty amazing."

When construction projects are shutting down freeing up workers and equipment, an lenders are getting scared and pulling back money, builders and developers will rush to pay workers to get buildings built.

The Empire State Building cost $40 million but the 11 building Rockefeller Center build throughout the 30s cost only $100 million. The Empire State Building was unprofitable for years while the Rockefeller Center was profitable soon after completion because it was fully occupied, which the Empire State Building still wasn't in 1939, or even 1949.

Rockefeller Center added 5 million square feet of rental space while 1+ million square feet of the Empire State Building was largely unoccupied.

The 30s and 40s destroyed the workforce able to build skyscrapers, and the owners of those built in the 20s got government to block construction of any competing skyscrapers.

Except for the government WTC development, building landmark pskyscrapers ceased for the most part until the 21st century. Of the top ten NYC skyscrapers, two were completed by 1930, the rest were built in the 21st century. But their height was capped to prevent them from being taller than WTC twin towers, or its equal height replacement.

Am I the only one that reads this graph different? It looks to me like there was a major ramp up in the late 20s/early 30s and a few spike years but mostly the numbers are pretty constant.

Also, the speed chart isn't adjusted for height. Would be interesting to see a floors per day rating n

Projects can be completed quickly or slowly depending on a ton of factors. If speed of construction were the only thing that mattered, skyscrapers would be built more quickly than they are -- but it is not the only thing that matters, and to focus on it as Patrick Collison wants to is a mistake.

The interesting question is not how long did it take to build the Empire State Building vs the Sears Tower, rather the interesting (but exceedingly difficult) question is what does the tradeoff curve look like, between speed on the one hand vs all of the other desiderata (cost being a prominent one) on the other hand. Has that speed-vs-cost curve shifted over the years? Or have other factors caused builders to choose a different optimum along that curve?

But the actual speed of completion gives us only limited information. Some projects are finished quickly, typically because there's urgency behind them (examples in the comments and the article include the Empire State Building, the Pentagon, and the Apollo program). Some are finished slowly (examples include the Taco Bell that I mentioned, or the Bay Area's slow process of replacing it's broken freeways after the Loma Prieta earthquake compared to Southern California's faster process of repairing I-5 after the Northridge quake).

If Taco Bell had truly needed to, it could've replaced that restaurant in months, perhaps weeks. The fact that that Taco Bell took well over a year to complete whereas another one in another city was built in say four months tells us almost nothing. Ditto trying to compare the construction time of the Empire State Bldg vs Sears Tower.

And some projects take exceptionally long amounts of time, or are never finished at all. It took China a thousand years to build its Grand Canal. And the Corinth Canal took at least 1,8000 years. But again, those data points tell us almost nothing. If say Augustus Caesar had decided that the Corinth Canal was an important priority, it would've been built in probably a few years. When we look at the actual data, we might observe a construction time of a few years, or of 1,800 years, just as Artir observed a range of construction times. But I don't think we learn anything useful from those comparisons. It's like comparing NBA teams by comparing the average height of their players: a variable of some interest but only limited usefulness.

Comments for this post are closed