By James
Pethokoukis and Mark Muro

When we think of technology, especially software, Silicon Valley immediately comes to mind. But during the pandemic, remote work has gotten a lot of attention. So is this the end of Silicon Valley, or will the Bay Area’s dominance continue for the foreseeable future? In this episode, I spoke with Mark Muro, who has used job postings to study that very question.

Mark is a senior fellow at the Brookings Institution, a think tank in Washington, DC, and the policy director of Brookings Metro. He and Yang You recently authored “Superstars, rising stars, and the rest: Pandemic trends and shifts in the geography of tech” for Brookings.

What follows is a lightly edited transcript of our conversation, including brief portions that were cut from the original podcast. You can download the episode here, and don’t forget to subscribe to my podcast on iTunes or Stitcher. Tell your friends, leave a review.

Pethokoukis: Why has the American tech sector remained so concentrated in so few places, particularly San Francisco and the Bay Area? Is that because of broad economic forces? Is that because other things happening in those regions? What do we know for sure about that?

Muro: The first descriptive fact is all of these places
had really high BA attainment in 1970, so they are high human capital places.
And then they had early-mover or first-mover advantage. And then my part of the
thesis or the theory here is that there are particular things about digital
that angle it towards winner-take-most dynamics and superstar dynamics.

And nobody’s doing anything wrong. These are how the technology works. And these are issues around network effects. That’s a huge one. They’re centers of big platforms, the way these industries have joined, so they’re command-control centers. And then the most obvious one that is talked about most is the huge power of conglomeration effects in human capital around skills acquisition. And being big and being identified center early has made a huge difference for those places. There are a limited number of what we call superstar hubs. And I would say, they’re not going anywhere.

Just one
point of clarification: What do you mean by conglomeration? What is
conglomeration economics?

These are the effects of large concentrations of people,
or other effects, where there’s a significant benefit to having large returns.
So you get significant concentration here, and there are benefits of that
concentration, most obviously, acquisition of highly skilled workers.

This might be
obvious, but I’m going to ask it anyway: What is special about the digital
economy that seems to lend itself to this concentration supported by these
conglomeration effects?

Well, this is not primarily an infrastructure economy;
it’s hugely a human capital economy. So I think that has allowed an accentuated
version of the conglomeration effect. There are other aspects to affect the
availability of particular inputs and so on, but it’s particularly pronounced
because it’s about the acquisition and concentration of people, and that can be
also marshaled quickly. And I think that adds to the speed that you can see in
the accumulation of these economies.

It’s about
human capital, so it’s not because they’re located by rivers or the ocean or by
coal mines or something. Those aren’t the key factors making these places
long-lasting hubs. It’s who’s there.

Absolutely. And you might think that would be more
ephemeral than a river mouth or something, but it turns out to be quite
durable. And because of these huge returns to the same inputs, you can get this
self-reinforcing dynamic that we think has been really powerful in this economy
and extremely central to the special sauce of US achievement in this field.

As far as
cities being superstar cities for this kind of talent—they’re not over. We’re
not going to talk about scenario where they quit being important, where the Bay
Area quits being important for tech. Maybe I’m assuming some of the other
cities—New York, Boston, Austin—are these the cities we’re talking about? And
they will be durable as far as being centers of this talent.

Yeah. I don’t see them going anywhere. And in fact, we’ve
seen them increasing their share. Through the entire last decade, I think lots
of observers have been waiting for some decentralization and it has not
occurred. And I think that is because of the special superpowers that we’re
talking about here. And it’s especially important for early-stage activity. All
of the things that we’ve been talking about really piles on for innovative,
creative work in these fields. We should expect durability. We maybe shouldn’t
expect permanency though.

But there are
other cities which are rising, doing better. What are those cities and what do
we know about why they seem to be doing better as far as attracting this kind of
talent? It’s a human capital-centered phenomenon.

We see a fairly large number of places of all different
sizes seeing some upticks during the pandemic. But you’re exactly right. There’s
a list of fairly sizeable, often Sunbelt metropolitan areas ranging from the
Atlantas and Dallases of the world to a Denver, for instance, that are actually
quite sizable so they can replicate some of the virtues of size that the
superstars enjoy. They are almost all anchored by a substantial research
university, having pretty decent climate seems to be showing up as something
here, and then having pretty rich pools of technical talent.

And so not for all activity, but it’s beginning to look
like those are very attractive places for a lot of activity in tech. And we’re
beginning to see either moves into them by firms or just organic growth in
these places. And so I think there are a number of things going on, but we’re
beginning to develop a next tier that has been quite formidable and challenging
for superstar status.

What
advantages does that next tier have? Is it an affordability issue where those
places are cheaper to live? Is there something else that makes them
additionally attractive if companies are looking to locate, if people are
looking to relocate, other than some of these other factors like universities?

You can look at that group and they do seem to have some
feel for entrepreneurialism. They seem to be fairly free-wheeling. Clearly, at
least for now, they’re significantly cheaper. I think that’s part of it. And
those things seem to be counting. And then they do have warm weather, which I
think is part of it.

Did the
pandemic change anything?

It did. I think it’s a two-track reality. There’s no doubt
that those big superstar centers on the coast remain entrenched, remain dynamic
in many cases, and actually slightly increased their share if you look at the
eight biggest ones. So that remains. But we do see a very vibrant performance
by these rising stars, these sizeable, often Sunbelt, bigger places.

And then I think quite interestingly, dozens of medium-size
and smaller places, often quality of life meccas or college towns, actually
started growing in a new way. And we see growth at least for a year and with
some further signals through the pandemic. We count something like 60 places
that grew faster during the pandemic than before it. Now, I think this is
evidence of some of the remote work effect. These might be the Zoom towns that
people are wanting to talk about that are supported or gain from foot-loose
tech workers arriving with their laptops but working for elsewhere.

And Mark, is
it also places where you can also buy a house? So you’re not stuck in an
apartment if you have a family, you can actually have a house where you can
have a home office. Is that part of it?

Yeah, I think very much. And I think these are not huge
numbers in these places, so that we may be seeing a temporary effect, but there
seems to be something real there. And I think you put your finger on aspects of
it. It’s a mix of quality of life, ability to work in this industry, and then
maybe cost effects for sure.

And we’ve
talked about durability, and maybe this is temporary. So what is the durability
of that phenomenon? If we don’t see big deadly waves of COVID, do things begin
to drift back to the way they were?

I think that the open questions are around remote work for
one, and then whether there’s a substantial change in the siting posture of big
companies. But the remote work question is interesting. To the extent that we
drift into a kind of hybrid—and that seems to be the leading candidate right
now: more hybrid work, more people doing some work remotely but some in the
office—that will limit the extent of this remote work, Zoom town piece of this.

But we don’t know what’s going to happen. I think there are signals that we’re going to pull back a bit, but I think most people agree that things will not be the same and there’ll be substantial amounts of remote work. Nick Bloom in Stanford is talking about two to three days a week in an office. That’s substantial.

I’ve seen his research. That’s a pretty massive change from people maybe working out of the office one day a month or something to doing it two or three days a week. That is a huge change.

Right. So I think this is unfinished. The effects I’ve
been discussing may be a temporary disruption, but I also think we’re not going
back to where we were before. So I think it’s going to be very interesting to
see what happens. Meanwhile, there are other effects. I think we are maybe
talking too much about Zoom towns and foot-loose workers and not talking enough
about the siting predilections of these companies. And I think there are
reasons that they are moving into the interior as well: maybe some regulatory,
some about wanting a different sect—the coastal mindset and group think—but
also seeking just an insatiable need for talent and the recognition that people
aren’t necessarily going to move to Seattle, move to the Bay Area.

They’re also moving in. I think there’s in the pipeline
already substantial, significant adjustment of the current footprint. Because
you’ve got a pretty long list of either headquarters moves or new satellite
office moves. I think there’s going to be a lot of desire to access diverse
talents in the south or in Atlanta and communities like that. And so I think we
have a lot of things, several cross currents at once. So that may be the more
durable and then a somewhat moderation of the foot-loose story.

Is there a
difference between the kinds of workers who are not showing up at the office or
maybe moving elsewhere? Because when we talk about these conglomeration effects
and how they’ve played out in places like the Bay Area, you’ve seen people who
may have been engineers at a tech company who leave or start their own company.
Are those the people moving to these areas or is it more like the marketing
people? And does that make a difference—who’s going into the office, who’s at
headquarters—or does that not matter?

I think the fundamental purpose of the core superstar hubs
to date has been to congregate technical engineering design talent to the
highest-end creative activity. Those workers have been able to exert a lot of
worker power here and they may, on the one hand, be most likely to demand to be
somewhere else, but they’re also going to be very important to the firms.

And so this is where I think the firms may wind up
creating decentralized networks within themselves in a lot of satellite
locations so they can retain those workers and square the circle with remote
work tied to our Texas office or our Nashville office. But I haven’t seen
research on the difference of different behavior of different types of workers,
but I think there could be something there.

Has anything
during the pandemic changed the ability of places which don’t have very many of
these companies, which don’t have much of their talent? Maybe they’re called
these “left-behind” areas in the Midwest, and there’s all this interest in creating
little innovation tech hubs in these areas. Is that happening at all? And most
of what we’ve talked is very bottom up: what companies want, what workers want.
Are they just going to stay left behind? I can see Salt Lake City or Miami, but
for some cities, I’m not sure if there’s an obvious fix.

I think there are limits to this. I think a popular response
has been to encourage foot-loose work by handing out incentives and so on. But
I think these are inherently small bore and are not going to be transformative
to communities. But the remote work vision is very attractive and I think has
been embraced by some of these communities, but I don’t think that it could be
transformative at scale. So I think that is a challenge.

I think that we are working out some ways to help somewhat
larger places that have a toehold in this sector to gain traction. And I think
that we can talk about some of the interplay of federal support with, in my
view, what needs to be bottom-up strategy and so on. But I think that is, for a
bit bigger places, something that is possible.

Do those need
to be places that have a toehold, that have something going for them? Maybe
there is a university there or there’s already some existing tech company.
Maybe the weather’s not so bad, I don’t know. But they can’t just be a blank
page.

Yeah. And I think within reason, we know some of the
things that matter. An R1 university with a substantial computer science
presence, is one thing that matters. A pool of technical and engineering workers
is a core: some presence already so that there is an emerging differentiated
technical presence. Those things are really important. And I think to the
extent we agree that a degree of decentralization of this industry and the
emergence of more centers is a good thing, then we should be focusing on a
particular group of these places that do have some of these starting
attributes.

I have two
questions. Why is decentralization a good thing? And what can government do for
these cities which maybe have a bit of a toehold? So, why do we care? And then
assuming we do care for a good reason, what should government do?

Right. So I think we care from a national competitiveness
and welfare perspective. We now have too much tech in too few places with
incredible accumulating side effects, externalities of costs, real estate
costs, transportation costs, and some would argue problems of group think in
the industry, narrowness of the industry.

And meanwhile, though this is our most dominant industry
globally, we actually have relatively few truly significant hubs. I think we
need more of them to compete, even with China’s depth. China has created dozens
of centers, scores of them really. So I think we need more of these in a broader
array of places. I think also there’s a political economy dimension: I think
we’re losing a sense of consensus about innovation and about even tech because
so much of the country feels it is extremely remote, feels it’s a coastal elite
activity. And so for all these reasons, I think we do need more.

Let me just
jump on that. Some might say that despite issues of affordability and maybe
strains on infrastructure and maybe group think, clearly these companies still
find a lot of value being where they’re at. So the concentrated nature is passing
the market test. It must still seem to work, and just because China has a lot
of tech centers doesn’t mean we should necessarily do that.

Absolutely. And nobody should be talking about erasing the
hubs that we have built up, which are demonstrably have been extremely
successful over the last 25 years.

So to what
extent will this decentralization make companies and tech, and therefore our
economy, more productive? Is that the reason we should do this? And how much is
if that it was more spread out and there were more Americans and more
legislators who felt invested, it would be easier to spend more money on
R&D because they would feel that the whole country was benefiting? So is it
both, or is one more the issue?

I think it is both, but I think we do have fundamental
innovation challenges. I think the talent-matching process seems to be breaking
down. We can’t rely on mobility and migration from the interior to these
coastal, expensive metropolises. So just from a matter of accessing the ideas
necessary to drive the innovation system, I think we need a more decentralized
system. But I’m the last one to think we should erase or reduce the centrality
of these hubs. We can have both.

And I think that we have seen things become so much more
concentrated that I think we’re hitting the wall on actual productivity and
innovation questions. Remember one thing, I’m not just talking about the
presence of these hubs in relation to the rest of the sector, we have shown
that their share of the overall national industry has increased and it’s still
increasing. So I think that we need to be thinking about tuning this platform
to make sure that it can retain its effectiveness.

When thinking
about government policy, is this more a federal policy, is this state and
local? What do you want done and who’s going to do it?

Because we’re talking about a continent-scale system,
there’s clearly got to be a national purview. If we want something different to
happen, we’ll have to intervene in that system. With that said, I think the
sweet spot is that of federal resources flowing into promising places with a
bottom-up, public-private suggestion of strategy and technology strengths based
on locally accessed strength.

Nobody’s talking about a top-down effort, but we are
talking about national resources flowing into bottom-up problem solving. And I
think that’s the story of a Madison, a St. Louis, an Austin, and especially
somewhere like the Research Triangle. So there’s a role for independent,
strong, well-informed regional stakeholders and initiatives close to their
strengths and the industry. But federal support is needed to really alter the
trajectory somewhat.

Are you familiar with Jonathan Gruber’s plan in his book Jump-Starting America? It would spend a lot of money, and he had all these metrics, different places would be competing for federal funds. Is that what you’re talking about, or do you have a different approach?

Yeah, very much so. And we worked with Jonathan and Simon [Johnson]
also on a proposal for regional tech hubs, which is essentially an R&D and
innovation surge at quite a large dollar amount into a relatively few places.
And with some targeting of particular strategic industries of national and
strategic importance, I think this is the broad architecture that can work in
the United States because it’s both centralized in terms of resources but
decentralized in terms of strategic growth, strategic development and, I would
say, marshaling public-private interactions—and all competitive by the way.

We began
talking about how the digital economy was different, it lent itself to this
concentration. But do we need better of the old economy? Do we need better
infrastructure? Maybe so people can agglomerate with each other? Do we need high-speed
rail? Do we need faster broadband? Do we need any of that stuff so people can
bounce ideas off each other and have meetings and everything’s not Zoom? And is
that important here or is that just an old way of thinking?

Well, and absolutely at the top of the list are steps to
make sure that we are building sufficient housing. Because that becomes the
pinch point in every one of these trajectories, including the new hubs that we
hope will emerge or that are emerging, which are already going down the same
trajectory often of seeing tremendous housing price inflation that becomes then
a limit on their own effectiveness, their own innovative capabilities. So all
of this is extremely important and very much part of the system that’s needed.

Mark, thanks
for coming on the podcast.

Great. Thanks, Jim.

James Pethokoukis is the Dewitt Wallace Fellow at the American Enterprise Institute, where he writes and edits the AEIdeas blog and hosts a weekly podcast, “Political Economy with James Pethokoukis.” Mark Muro is a senior fellow at the Brookings Institution and policy director of Brookings Metro. He and Yang You recently authored “Superstars, rising stars, and the rest: Pandemic trends and shifts in the geography of tech.”

The post Is the Tech Sector Too Concentrated? My Long-Read Q&A with Mark Muro appeared first on American Enterprise Institute – AEI.