Why do innovative businesses tend to “cluster” together in tech hubs like Silicon Valley? Why do these companies stay in these expensive cities when they could go to another city with significantly cheaper costs? Can these tech hubs be better governed? And how might other cities catch up? To explore these questions, I am delighted to speak with Enrico Moretti.

Enrico Moretti is the Michael Peevey and Donald Vial Professor of Economics at Berkeley. He is also an associate with NBER, a fellow for the Centre of Economic Policy Research and the Institute for the Study of Labor, and the editor-in-chief of the Journal of Economic Perspectives. His 2012 book, The New Geography of Jobs, brought to light much of the data showing that high-tech industries tend to cluster together in small areas, and he has repeatedly returned to this subject in his academic work since then, including in his new paper, “The Effect of High-Tech Clusters on the Productivity of Top Inventors.”

What follows is a lightly edited transcript of our conversation. 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: A lot of your research looks at how innovation-based businesses tend to cluster together in these tech-hub cities and metro areas — perhaps most famously in Silicon Valley. You also look at how the vitality of these hubs is important, not just to those regions, but more broadly to the national economy.

I want to start off by asking: How concerned should I be as someone who doesn’t live in San Francisco or the Bay Area that San Francisco seems to be a poorly managed city? You have tech companies leaving the city, they’ve just opened some sort of “Emerging Technology Licensing Office” — it seems to me that if San Francisco is an important tech-hub, it’s really important that it be a well-managed city as well.

Moretti: What we’ve seen over the past 30 years is an
increased concentration of good jobs in cities like San Francisco — and not
just there. Seattle, Boston, Austin, Raleigh-Durham, they’re also great
examples. San Francisco and Silicon Valley are by far the greatest
concentration of innovation-sector jobs. That’s in the tech-sector, narrowly
defined, but also much more broadly.

One main source of policy concern is that this area has been
quite hostile to new housing developments. So it’s an area where labor demand
has increased tremendously because of the vitality of the tech-sector, but
housing supply has not kept up. It doesn’t take a PhD in economics to figure
that the increasing cost of housing is going to have a big effect.

I think this is mostly due to policy decisions — not just on the part of the city of San Francisco and Silicon Valley, but also all the cities in the peninsula as well. They’re incredibly opposed to new housing. And that’s a major drag on the overall growth of the sector, but also on overall US economic growth.

So ideally, you’d
have these high-productivity clusters — because it’s good when other similar
businesses locate there as people move from business to business.

You have this
virtuous circle, and as you mention in your research, it’s not just people
working at those companies that get the benefits. They also create other jobs
in the service sector which are well-paying jobs. These clusters don’t just help
on the high end — it helps more broadly.

The housing is a big
issue because, one, people can’t move there. And then if they do move there, I
imagine that housing costs eat up a lot of their income gains.

That’s right. If you think about the boom towns of the past from the 50s and 60s — places like Detroit or Cleveland — those were the places that would attract talent and workers from all over the US. Those were places where it was easy for the average worker with an average family to move to and find affordable housing right away.

The boom towns of today are much more supply-constrained.
The San Franciscos, the Bostons, the Seattles, they have decided not to add
enough housing within their boundaries. This means that the average family
finds it much harder to find affordable housing if they want to move there.

It also means that a lot of the benefits of the incredible
economic dynamism in these areas end up being captured by incumbent landlords
who were lucky enough to have bought land before the current boom.

Is this an unsolvable
issue? Do you feel like you’ve identified a big problem that has big economic
effects, and the solution seems fairly obvious, but the politics are utterly
impossible?

Well, the first point to recognize is that the decision to
not add housing — although it has some geographical components, you know, you
can’t build on the ocean — is mostly a political decision. It is incumbent
voters, mostly homeowners, deciding that they would rather keep land very
scarce by constraining the number of houses that are allowed to be built.

There are solutions to it — there are-less than-optimal
solutions, and there are better solutions. One less-than-optimal solution is
probably what Austin is doing. Austin is implementing very few housing constraints.
It’s growing — I was just in Austin three weeks ago and I’ve never seen so many
cranes in my life. Austin has exactly the same labor-demand shock that San
Francisco has, and it has the same industries, really, and literally the same
employers. But housing in Austin is allowing for unchecked growth.

Anyone can build anywhere — it’s largely unconstrained growth. I think this is positive for housing prices because housing prices are lower than in San Francisco, although they’re growing as well. But by allowing growth everywhere in the region, they’re also adding some negative side-effects such as traffic congestion. Like, you don’t want to be on one of those freeways at peak hours — it’s really jammed.

A better way would be to concentrate development to allow
much more housing development but to focus it near the urban core, and increase
the provision of public transit. Then you can get additional housing supply
increases, which help keep cost of living in check, but you add less to
congestion and traffic.

So finding the right balance between an extreme case like
the Bay Area where nothing gets built and another extreme where everything gets
built but without a lot of planning — it’s not rocket science. It’s not like
some secret, it’s political will.

Are there cities that
have been able to strike this balance?

I would place Seattle in the middle, between San Francisco
and Austin. Seattle has been much more pro-growth than San Francisco. It has
traditionally had a much less political housing entitlement process — so if a
developer wants to build it’s much less political. It’s more shielded from the
endless series of appeals we see here in the Bay Area.

At the same time, it has been successful at concentrating
much of the growth in parts of the city that are near downtown that were
totally underutilized 20 years ago while successfully limiting sprawl on the
outskirts. That’s typically what generates most of the traffic and
environmental costs. I don’t think that Seattle is by any means perfect — cost
of living has increased significantly in Seattle, but by less that it would have
increased had it not been for a pro-growth housing policy in the right places.

How do those sorts of
policies happen? Do you just happen to get the right people in office — maybe
people who live there have a positive attitude towards growth and development? Do
you just have to get lucky, or is this a solvable political problem?

I think it’s a solvable political problem, and I think that
even in the Bay Area you’re starting to see some progress. For example, the
city, which was incredibly restrictive as of late, has started to become a
little bit more open.

I think the progressives in the city are beginning to
realize that more housing, yes, might mean more profit for the developers, but
it also means cheaper rent for the renters. So it’s quite a progressive policy.
The worst offenders are now, not the city of San Francisco or Oakland, but
these smaller communities in Silicon Valley and the East Bay.

They remain incredibly restrictive, and you see crazy situations where there could be a semi-empty parking lot next to the train station, and the citizens will be up in arms fighting any building that’s taller than two stories. Even if this would be next to the train station with an empty parking lot.

So I think it takes some time for these battles to be won,
but in the most progressive parts of the Bay Area — meaning San Francisco and
Oakland — this notion is beginning to percolate.

So if we did policy
better, would these hubs have struck a better balance? Would they just have
much bigger populations because they’d be more affordable, and more Americans
would move there because they would see that they could afford it, and they’d
have good jobs at a tech company or maybe even a service job?

They would be bigger, and there would be two main effects. First of all, as you point out, they would be more affordable. But second — this is the point of a recent paper that I wrote with my colleague Chang-Tai Hsieh — there would also be important aggregate benefits in terms of economic growth for the state, and for the broader region in the United States.

The idea here is that by allowing more growth in the most
productive cities of the US, a larger share of the US workforce would have
access to these higher productivity jobs that are now essentially constrained
by the housing supply. So by making these areas larger, local employers would
be able to hire more, expand more, and more US workers would access these
higher-productivity, higher-paying jobs that are somewhat limited by housing
supply.

In that paper we find that there would be measurable, economically large effects, not just for the region, but also for the US as a whole.

How significant would
the aggregate impact be on the broader economy?

We find that the broader economy would experience faster
economic growth, and that the gain in the average earnings for the average
worker in the US would measure in thousands of dollars. Depending on exactly
the assumption you make in the model, it could be $3-5,000 a year for the
average worker. That’s if the most productive US cities relax their housing
restrictions to the level of the median US city.

It seems that suddenly
we’re talking a lot about wealth inequality in the news. Isn’t this housing
issue a big part of the wealth inequality problem? And when you’re recommending
these looser housing policies, that’s also an anti-inequality policy, right?

I would say it’s an anti-inequality policy in terms of
income even more so than wealth, in that it would allow more people access to
well-paying jobs that are now harder to get because of housing cost. So it
would increase the number of US workers who can afford to have access to good
careers and good employers. There is also an effect on wealth, but I think most
of the research on wealth focuses on other forms of capital, not just housing.

So it would be great
if more people could move to those cities. And maybe there’ll be housing reform
that allows that to happen. But if that doesn’t happen, then what?

Do we eventually see
cities become more expensive, and do both potential workers and entrepreneurs
just end up going to other places? Do they end up starting their businesses in
Columbus, Ohio or other smaller cities? Could you end up creating smaller tech–hubs,
but more distributed around the country where the wealth gets spread around?

Up to this point, we don’t see that happening on a large
scale.

Let me be more precise: In some of my work, I look at the
productivity differences that workers in the innovation sector exhibit in
different cities. What you see is that these star cities, yes, cost way more.
That’s to a firm — in terms of labor costs, in terms of real estate costs. But
the productivity advantages are still larger than the costs.

So it’s still a good deal for an employer, especially those
that hire a lot of engineers, a lot of PhD’s in the sciences, and focus on new
technologies and products. What you see is some evidence of outsourcing or
opening of new offices in other cities for parts of these firms that aren’t
crucial. It’s not R&D, or the core functions of the innovation firm.

So, a lot of Bay Area firms have offices in Utah, or in Austin
as I mentioned, in Colorado, and increasingly in the South — Texas first, but
also Tennessee and other states. But the types of offices that they open there
at this point are not the core jobs that make those firms successful — not the
engineers or scientists. They are auxiliary positions — sometimes it’s HR,
sometimes it’s customer service.

Don’t get me wrong, those are all great jobs, so it’s great
for the recipient cities to have them. But at this point, the productivity
advantages of being in these innovation centers still outweigh the costs.

Do you see that
changing? Are we anywhere near a tipping point where it’s just not worth it,
and people decide that they’re going to move their companies to Austin, or Salt
Lake City?

No, I don’t think we’re close to that tipping point yet.
Part of the reason is that you still see an increase in the agglomeration of
certain types of jobs in the Bay Area, or in Seattle, or in Boston. The
productivity advantages are still higher.

We keep talking about the Bay Area, but the story is much
broader — you know, Boston is the same. If you look at what happened in
Cambridge, the amount of life science and biotech-related jobs agglomerated in
Cambridge is staggering. Cambridge and Boston as a whole is one of the most
expensive areas to do business in the US, just like San Francisco — it’s lower,
but not all that different. And yet, companies keep agglomerating there despite
having the option of moving to cheaper locations.

So I think that the fact that companies and VC investment
keeps agglomerating in a limited number of star cities, but that some types of
offices get opened in other cities, are both perfectly consistent with each
other.

In fact we see the same thing with finance. Finance started
this process 60 years ago when a lot of the banks, investment companies, and
later hedge funds started delocalizing some of the positions to cheaper places.
First, it was New Jersey, and then it was Arizona, and then it was Bangalore.
But the core jobs in finance remained in headquarters in either New York or
southern Connecticut in spite of this decades-long process.

So I don’t think this is all that different. There will be outsourcing of some positions, but the core will remain in these innovation hubs.

Public officials
around the country would love to have a tech–hub in their state, which is one
reason we saw all of this bidding for Amazon to relocate its employees. It
seems like that’s a very difficult thing to do — trying to purposefully say,
“I’m going to create an innovative technology hub in my state or in my city,”
from the top down. How successful have governments been in doing that in the
United States?

I am skeptical that it’s something you can engineer, top
down. If you look at the history of the innovation hubs in the US, whether
they’re the big ones — the Silicon Valleys, the Seattles, the Bostons — or the
small ones, it’s hard to find examples where an innovation hub was explicitly
created by a deliberate policy on the part of the county or state saying,
“We’re going to create the next Silicon Valley, there.”

In my reading of the history of the innovation hubs, there
is no example like that. The typical story through which innovation hubs come
about is much more organic. Typically, it’s the success of one local company in
one new technology sector that then becomes the seed around which the cluster
agglomerates.

Once that seed is in place, you start seeing this self-reinforcing mechanism of increased concentration. But the initial seed in the US has never been a deliberate policy on the part of local government. It’s really hard to know which company is the next Microsoft or which company is the next Amazon, exactly. It’s hard for venture capitalists, let alone mayors or governors.

Given all of the
economists’ warnings when Amazon said that they were going to put 50,000 jobs
in one city, and given how hard it is to engineer a tech-hub, was it crazy for
all of these cities and governors to try to pitch Amazon? Where they tried to
immediately create a tech–hub from 50,000 Amazon jobs out of this
once-in-a-generation opportunity — was it rolling the dice by attempting to
grab those jobs?

The second headquarters of Amazon is very interesting,
because if you’re thinking about a big shock that could generate a hub where
there was none, I mean, that’s as big as it can get. 50,000 tech jobs all
concentrated is huge. The typical new opening is much smaller by an order of
magnitude.

So I’m not a great fan of offering $6 billion, which was one
of the bids made by one of the 20 finalists, for attracting a new company. The
second headquarters of Amazon, especially when it was won, was going to have a
lot of clustering benefits for the community where it went. And I think the
city will get significant benefits.

The benefits are twofold: The first one is that it’s going
to be a magnet for other companies that will agglomerate around it. The second
benefit is all of the additional jobs that are not in tech, but they are going
to be in the local service sector that will exist because you have so many
well-paying tech jobs in that city.

That said, it doesn’t justify billions of dollars in bids. I
think that DC, or Virginia I guess, was quite savvy — they were one of the
lowest bidders, and they realized early on that Amazon was most likely not
going to choose just based on the bids. They were going to choose based on the
economic fundamentals of an area.

So they didn’t have to bid as much because it’s an area with
a highly-skilled labor force, they have incredibly good fundamentals, and they
didn’t need to throw away a lot of taxpayer money with their bids. In fact,
they won the second headquarters with one of the lowest bids among 20 finalists
— a very good decision on their part.

It always seemed
unlikely to me that Amazon was going to take 50,000 jobs and pick Cleveland, or
Indianapolis, or a Midwestern city. I always felt they’d end up on the coasts.

You hear this talk about the broader Rust Belt area, these “left behind” areas — what advice would you give policymakers in those areas? For instance, there’s a proposal to send federal agencies to various cities in those regions to try to help them. Does that sound like a good idea to you, or are there any kinds of economic development pathways they could take if they’re unable to create a tech-hub?

First of all, let me say that I agree with your earlier
statement — it seemed unlikely that Amazon was going to go to Detroit or
Cleveland even when there was a lot of talk about it. As I said then, even
before the names of the finalists were released, the most likely outcome was
Amazon going to an area with a very highly-skilled labor force — a very deep
talent pool.

If you look at the history of US cities, the most important
predictor for attracting tech and innovation sector businesses is the share of
workers with a college degree in the labor force. It’s a share that varies
enormously across cities — some cities have 50 percent of workers with a
college degree, with DC being an example. And other cities have 10 percent of
workers with a college degree. These are enormous differences, and I think
that’s the best predictor of the attractiveness of a city for these sorts of
companies.

Just to put this difference in perspective: The US regional differences
in college degrees are much larger than the difference of the US as a whole and
developing economies like Peru and Bolivia. So within the US, there are some
economies that are incredibly well-educated and incredibly attractive to these
companies, and there are economies that are much less attractive. To me, it was
clear well before they announced the names that the former group had the chance
of attracting, and the latter didn’t.

But the question is what to do about the latter group. One
thing that states have been doing — not with great success — is engaging in
bidding wars, where they’re offering big subsidies to pretty much any new plant
or office opening. It’s an incredibly bipartisan policy — blue state and red
state governors love them equally. The biggest provider of the subsidies are
taxes, which, in spite of the free-market rhetoric, are really providing a lot
of subsidies.

Overall, US states are spending an excess of $80 billion
annually in subsidies offered to firms to come to their jurisdiction. Now, this
is larger than most welfare programs we have in the US, in aggregate. You’ve
got to wonder whether this makes sense, because essentially you have states and
counties that are bidding against each other for an investment that was going
to go somewhere within the US. To some extent — probably not fully, but still —
it’s a zero-sum game across communities. We’re shifting where the investment
goes, but the main winners are the firms and the main losers are local
taxpayers.

Again, that’s one of
the reasons that many economists were against bidding for Amazon.

But I can tell you
that whenever I write about your research, I’ll always get an email or a tweet
that says, “Is this the only answer you have for unemployed coal miners? Tell
them all to move to San Francisco? That’s it — you want everybody to move to
these high-productivity cities, abandon their families, and make the leap?”

And not everyone is
going to move, so what answer do you have for those who aren’t going to?

I think that one of the best investments that state and
local government can make today is public education. I think there’s a wealth
of evidence that that’s one of the best industrial policies that we can create.

It can be at the lower level of schooling, the high level, it could be community colleges. I think community colleges are vastly underrated as an important source of skill upgrading for the local workforce. I completely agree that you cannot go to a declining community in the Rust Belt and say, “Hey the only solution is to go to the Bay, Boston, or Seattle.”

I also agree that by investing in public education, you
cannot expect that community to turn around in the next year. These are
long-term economic strategies that take time. But I think the overwhelming
evidence in economic research is that one dollar spent in schooling for the
local labor force and students has an aggregate return much higher than one
dollar for that community in the long run. And that’s one of the best
investments that local communities can make.

The return is much higher than investing in roads, bridges,
or rail. It’s higher than the return that can be made in other forms of public
spending. This goes back to what I was saying before: Over the past 30 years,
the best predictor of economic growth for US communities is the share of
workers who are well-educated.

Does that guarantee the mayor that if they invest in the
local community college, or their high school, or lower-level schooling, that
they’re going to get the next Silicon Valley in their jurisdiction? No,
absolutely not — we need to be honest about that. It’s a tall order. It does
guarantee, though, that the investment will eventually turn into better-paying
jobs in the long run.

The other point I want to make is that it’s not as though
every community needs to copy the Silicon Valley blueprint. The Silicon Valley
industry mix is unique, and it’s not necessarily the best recipe for every
single community. There are communities that might have a local industrial
employer, and maybe the solution is to work with the community college to make
sure that whatever skill the local employer uses can be provided by the
college.

That could be an industry quite far removed from what we see
in Silicon Valley — it might nothing to do with the internet, biotech, data
mining, or advanced material. It could be working within the existing strength
of the community.

So you’ve studied
that idea of dispersing federal agencies around the countries — such as sending
the Department of Agriculture to Kansas. Another idea is to have more
place-based visas, where you encourage immigration to declining areas. Do
either of those ideas sound good to you?

Yes, I think that both seem plausible. Jonathan Gruber has a very interesting new book, where he lists a dozens of potential policies of that kind. I think they’re all very plausible scenarios.

There are some federal agencies that you can disperse without much effect on their productivity, so I’m in favor of dispersing bureaucracy [laughs]. I would be much more reluctant to disperse certain jobs that require proximity with an innovation cluster.

Just for an example: I wouldn’t put an IA job in the middle of nowhere, because it’s important that those scientists are a part of the intellectual community in their industry. It would harm their productivity in their industry if we outsource them.

I know that Jon was on your show recently — he’s collected a number of serious proposals along those lines that seem to me economically valid. Are they going to completely reverse the decline of the Rust Belt or other regions? No, but they could help on the margins.

And immigration?
Bringing people in with place-based policies?

Sure. One thing, though, is that immigrants tend to go where
good jobs are. The thing that is most striking is that when you look at the
high-skilled immigrants, they are even more concentrated in star cities than
natives.

Silicon Valley here is dominated by Chinese, Indian,
Italian, and French engineers who are here for good jobs. I guess there are
policies that are not unlike the ones in northern Europe where, for example,
refugee status is tied to a given community. And yes, I think this could be
very helpful at least in the short run to revitalize some of these communities.

Enrico, thanks for
coming on the podcast — it’s been great.

Thank you for having me, I enjoyed it.

The post Tech hubs and the labor market: A long-read Q&A with Enrico Moretti appeared first on American Enterprise Institute – AEI.