The Spacetime of the US Economy

Visualizing 75 years of value creation

Fun (false) fact: Frank Lloyd Wright inadvertently invented the bubble graph in 1912 when he made these stained-glass windows for the Coonley Playhouse in Illinois

You've probably noticed that almost nobody who thinks, writes, or talks about technology maintains that we're living in UNexceptional times. I doubt you've heard a keynote speaker or TED talker say “AI is not that big a deal. The technologies developed before any of us were born were much more important.” The big exception here is economist Robert Gordon, who made that argument in his meticulous 2016 book The Rise and Fall of American Growth.

Bob makes a strong argument,¹ but he's pretty much the only one making it these days. Most of the other economists who study technology think that AI is pretty dang powerful and important. Some, like my MIT colleague (and recent Nobel prize winner) Daron Acemoglu , think that its potential to do harm to our societies, economies, and livelihoods is so great that it needs to be more closely governed by the state.

Others, like my other MIT colleague David Autor, are more optimistic. Autor believes that AI could do a huge amount of good and perhaps even shore up the American middle class. My frequent coauthor (and cofounder) Erik Brynjolfsson thinks that we’re not being ambitious enough with our goals for AI. Me? I think AI’s crystal clear, unfolding-before-our-eyes benefits massively outweigh its nebulous, could-happen labor market downsides, and that putting governments in charge of rolling out technologies juuuuust fast enough to protect workers — and no faster! — is about as bad an idea as economy-wide rent control.²

The other members of technology's chattering classes are like the economists; they vary in their optimism, but not really in their belief about modern technology's importance. And I’m with them! I believe that we are living in NOT-unexceptional times — otherwise known as exceptional times — when it comes to the puissance and salience of new technologies.

This Graph Proves Nothing, But I Still Like It

But I keep feeling a need to prove it. I keep on trying to find ways to test this idea that we haven't been here before - that we're living in demonstrably exceptional times. My tests often take the form of data visualizations. For example, long-time readers of this newsletter might remember the graph below, which I drew and posted a while back (green bubbles represent companies in digital high-tech industries; blue bubbles represent all other companies):

I think this graph supports the we've-never-been-here-before argument. It shows that in recent years there's been an unprecedented and large-scale youth movement in America's most valuable companies. This youth movement corresponds with the rise of the modern digital bundle: the Internet, smartphones, Cloud Computing, and AI. Here’s another graph from the same post making clear that this youth movement is a sharp reversal of the previous trend toward corporate fogey-dom:

Now those two trends (the arrival of the Digital Age and the corporate youth movement) could be unrelated. As we learned in our stats classes,³ correlation doesn’t imply causation. And of course this graph doesn’t prove that we humans have never been here before. It only goes back about a hundred years. It doesn’t cover the gradual (second) Agricultural Revolution, the sudden Industrial Revolution, or any periods before them. It doesn’t even cover the intense electrification of US manufacturing during the 20th century’s first couple decades (because we don’t have good share price data going back that far.) So I haven’t come anywhere close to proving the “never before” argument.

But those graphs are still interesting, aren’t they? For most of the 20th century America's big companies were a pretty stable group, and most of them were founded in the 19th century. But a new, much younger group has appeared in the 21st century and appears to be leaving behind the behemoths of the previous era. As I wrote

Between 1926 and 1986 the trend was for the top 50 [most valuable US companies] to age by an average of 0.75 years every year. Between 2006 and 2024 they got younger on average by 1.25 years each year. A whole lot of very young companies entered the top 50; in 2024 the group was as young as it had been 75 years earlier.

I've written here and here about whether we should expect those behemoths to bounce back. The single-sentence summary of all that writing is: I don't think we should expect them to bounce back, and would very much like to be wrong about that.

Same With These

As much as I like the graph above, it fails to capture something important. It fails to capture that not only is there a youth movement going on among America's most valuable companies, there’s also a dense clustering — a pronounced geographic concentration. Most of its big, young bubbles represent companies from a very small patch of West Coast real estate. I talked about America’s 21st-century West Coast gold rush here, and visualized it with the below snapshots of where the country’s most valuable companies were located at the turn of the 21st century, and then again at the start of last year (in the below, the green-ness of a bubble represents how young it is; old companies are grey)⁴:

I found this comparison striking, but it still wasn’t everything I was looking for. I was looking for a visualization that would combine both the geography and the age of America's most valuable companies, and how this spacetime landscape has changed over time. Tall order, right? Here's what I've come up with.

Maybe These Graphs Will Do the Trick

In all the graphs below, the horizontal axis shows each company's longitudinal distance from Palo Alto, California.⁵ The West Coast is on the left-hand side of this graph, and the East Coast on the right, just as it would be on a standard map. This visualization, then, is like a map of the US that doesn't give any information about whether the company is headquartered in the northern latitudes (near International Falls, Minnesota, say) or the southern ones (down by McAllen, Texas). I don't think we lose that much by ignoring the North-South axis. Trust me, the real action is East-West.

The vertical axis shows company age, with older companies down at the bottom and younger ones up top. And what exactly are we graphing on this weird spacetime coordinate system? Not the most valuable companies at any point in time, but instead the 50 US public companies that experienced the largest rises in their valuations over a quarter-century period. In other words, we're showing both the age and east-west location of the 50 companies that did the most of a major thing that public companies in a market-oriented economy are supposed to do: increase their valuations over a long time horizon.⁶ The time horizon we picked is 25 years.

The first quarter century we’ll look at is 1950-1974, those go-go postwar years when the US economy was generally going great guns. Let’s see who created a lot of value back in that time of executive dining rooms, bar cars, skinny ties, and oil shocks. (In the visualizations below, we're back to using green bubbles to represent high-tech companies.)

The top third of this graph is unpopulated, which makes sense: the y-axis goes as high as 2020, yet the end date of the data is 1974. Companies are pretty evenly distributed across the earlier years, with plenty of representation from both the 19th and 20th centuries.

The x-axis reveals a pattern that we'll see consistently: that the Rocky Mountains, Great Plains, and deserts of the Southwest are not fertile ground for high-growth companies. There’s a bit of action in California and then not much as we move east until we get to Texas, where oil companies benefited greatly from the embargoes and other upsets in the global hydrocarbon markets of the early 1970s.

Continuing eastward, we see a diversified clutch of Midwestern companies, including GM and Ford (Michigan), Eli Lilly (Indiana), and Sears, Roebuck (Illinois). And then we get to the industrial and financial powerhouse of the Northeast.

This visualization makes clear that the information age was already gearing up by 1974, and that investors were already excited about it. IBM is the largest bubble on this graph, bigger even than monopolist AT&T. And although IBM was a near-monopolist in computing, a few other green bubbles like HP and Texas Instruments were also big enough to be inclucded here. You can see the digital age coming in this graph, and it looks like it's coming from all across the country.

Now let's look at the next quarter century:

I hereby name this period the “confetti economy.” It was colorful and all over the place. Old companies and young companies were creating a lot of value. So were companies in the East, West, and Midwest (the Mountain Time Zone remains unrepresented). And fast-growing tech companies were found throughout the country.

But an interesting phenomenon took shape in the upper left corner of the graph. A large cluster of green bubbles formed inside the “15 by 75 box:” founded since 1975, and within 15 east-west miles of Palo Alto. Microsoft, Oracle, Sun, Yahoo, and Cisco are all within this box.

Over the next quarter century, that cluster became the center of mass for value creation in the US:

There are still some smallish flakes of confetti floating around the rest of the economy, but the main party is clearly happening with the 15 x 75 box. That box is so densely populated it's hard to see all of its contents, even after making the bubbles partially see-through. Uber, Meta, Amazon, Netflix, NVIDIA, Apple and Microsoft are all there, and they have plenty of company. It looks like the Mountain Time Zone finally incubated a green bubble during this time, but no: Palantir’s headquarters is now in Denver, but the company was born in Palo Alto and spent more than fifteen years there before moving east shortly before its 2020 IPO.

I find the concentration of value creation shown in this visualization stunning. Fully 70% of the graph’s total market cap growth is contained within the 15 x 75 box, as is about 62% of the net market cap growth in all US companies between 2000 and 2024. As I look at the three graphs above, it's hard for me to escape the conclusion that we went from having a diversified economy to having a diversified economy with a concentrated tech sector to, in the first quarter of the 21st century, having a young, West Coast tech sector with a national economy attached to it . That’s an exaggeration, of course, but it does capture how fast, broad, and deep the shifts in value creation have been in the second machine age.

Will those shifts continue? Will the green bubbles remain so concentrated, or will they spread out? When we draw the same graph for the next quarter century, will there be homegrown green bubbles throughout the country? And no matter where they are, will they continue to be so much bigger than all of America's blue bubbles? Right now, my answers are “concentrated,” “no,” and “yes.” But like I said, I want to be wrong here. I say to America's builders, entrepreneurs, and innovators, please prove me wrong.

1. I've had the pleasure of debating the issue with him a couple. I like to think I held my own, but probably didn’t. Bob is good.

2. This is the worst insult I could think of to fling at economists.

3. And keep forgetting.

4. Sorry for the palette flip-flopping, but these visualizations were made before I settled on a consistent color scheme.

5. Palo Alto is the mythical center of Silicon Valley because it’s the home of the garage that birthed Hewlett-Packard.

6. A horizon long enough to make us buy-and-hold investors happy.