THE CASE for GROWTH CENTERS - How to spread tech innovation across America - ROBERT D. ATKINSON, MARK MURO, and JACOB WHITON - Governor Tom Wolf
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
THE CASE for GROWTH CENTERS
How to spread tech innovation across America
ROBERT D. ATKINSON, MARK MURO, and JACOB WHITON
December 2019
Table of contents
Executive summary 3
1. Introduction 10
2. The entrenched geography of America’s innovation
16
industries
3. The costs of hyperconcentration 25
4. Why markets alone won’t solve the problem 32
5. Why place-based intervention is essential—and how it can
42
succeed
6. Countering regional divergence through growth centers 47
7. Spreading tech hubs across America: A proposal 52
8. Candidates to be America’s next top innovation hubs 62
9. Addressing objections and concerns 67
10. Conclusion 71
References 72
Endnotes 75
Appendix A 85
Appendix B 88
Appendix C 90
Acknowledgements 92
Executive summary
I
t has become clear that while the future are they likely to. Instead, these deeply seated
of America’s economy lies in its high-tech dynamics appear ready to exacerbate the current
innovation sector, that same sector has divides.
widened the nation’s regional divides—a fact
that became starkly apparent with the 2016 This is why the nation needs a major push
presidential election. WRFRXQWHUWKHVHG\QDPLFV6SHFLoFDOO\WKH
nation needs—as one initiative among others—a
Dependent on intense agglomerations of highly massive federal effort to transform a short list
skilled workers and based on winner-take-most of “heartland” metro areas with compelling
network economies, the innovation sector has strengths into self-sustaining “growth centers”
JHQHUDWHGVLJQLoFDQWWHFKQRORJ\JDLQVDQG WKDWZLOOEHQHoWHQWLUHUHJLRQV
wealth but has also helped spawn a growing
gap between the nation’s dynamic “superstar” The present paper, therefore, proposes that
metropolitan areas and most everywhere else. Congress assemble and award to a select set of
metropolitan areas a major package of federal
Neither market forces nor bottom-up economic innovation inputs and supports that would help
development efforts have closed this gap, nor those areas accelerate transformative innovation-
THE CASE FOR GROWTH CENTERS 3
sector scale-up. Along these lines, we envision enhancing nature of the regional economics
Congress establishing a rigorous competitive market.
process by which the most promising eight to 10
potential growth centers (all not geographically For much of the 20th century, market forces had
located near existing successful tech hubs) would tended to reduce wage, investment, and business-
UHFHLYHVXEVWDQWLDOoQDQFLDODQGUHJXODWRU\ formation disparities between more- and less-
support for 10 years to get “over the hump” developed regions. By narrowing the divides,
and become self-sustaining new innovation the economy ensured a welcome “convergence”
centers. Such an initiative would not only bring among communities and regions.
VLJQLoFDQWHFRQRPLFRSSRUWXQLW\WRPRUHSDUWV
RIWKHQDWLRQEXWDOVRVLJQLoFDQWO\ERRVW86DQG However, in the 1980s, that trend began to break
innovation-based competitiveness, including in down as digital technologies and innovation
the competition with China. moved to the center of economic activity. Intense
new demands for talent and insights increased
What follows is a discussion that situates the the value of “agglomeration” economies,
nation’s divergence problem, and highlights a set unleashing self-reinforcing dynamics that
RIUHOHYDQWoQGLQJVDQGUHFRPPHQGDWLRQV LQFUHDVLQJO\EHQHoWHGELJFRDVWDOFRUHUHJLRQV
often to the detriment of cities and metro areas
THE PROBLEM in other parts of the nation.
Rather than growing together, the nation’s Amid these conditions, the convergence trend
regions, metropolitan areas, and towns have been gave way to “divergence,” as a top tier of big,
growing apart. That has been a shock, including tech- and innovation-heavy metro areas such
for an economic and policy mainstream that as Boston, San Francisco-San Jose, and Seattle
has long trusted the self-regulating, welfare- began to consistently outperform less-tech-
Indexed average annual wages Indexed employment level
1969 - 2017 (1969 = 100) 1969 - 2017 (1969 = 100)
180 280
170
250
160
220
150
140 190
130
160
120
130
110
100 100
Bottom third of metro areas Median metro areas Top 2% of metro areas
Source: Brookings analysis of BEA data
THE CASE FOR GROWTH CENTERS 4
based places on measures of innovation-driven recommendations in the process of laying out
prosperity. what a federal innovation-based growth centers
program might look like. These takeaways include
The result is a crisis of regional imbalance. the following:
Among the superstar metro areas, the winner-
take-most dynamics of the innovation economy 1. Regional divergence has reached extreme
have led to dominance but also livability and levels in the U.S. innovation sector. The
competitiveness crises: spiraling real estate costs, innovation sector—comprised of 13 of the
WUDIoFJULGORFNDQGLQFUHDVLQJO\XQFRPSHWLWLYH nation’s highest-tech, highest-R&D “advanced”
wage and salary costs. Meanwhile, in many of industries—contributes inordinately to regional
the “left-behind places,” the struggle to keep up DQG86SURVSHULW\,WVGLIIXVLRQLQWRQHZ
has brought stagnation and frustration. These SODFHVZRXOGJUHDWO\EHQHoWWKHQDWLRQ V
uneven realities represent a serious productivity, well-being. However, far from diffusing, the
competitiveness, and equity problem. sector has been concentrating in a short list
of superstar metropolitan areas. Most notably,
FINDINGS AND RECOMMENDATIONS MXVWoYHWRSLQQRYDWLRQPHWURDUHDVBoston,
San Francisco, San Jose, Seattle, and San
Assuming that nonchalance is no longer tenable, Diego—accounted for more than 90% of the
the present report presumes that the time has nation’s innovation-sector growth during
come for the nation to offset the pull-away of the years 2005 to 2017. In this fashion, they
the innovation superstars with a concerted have increased their share of the nation’s
intervention to support the emergence of new total innovation employment from 17.6% to
tech stars in new places. Along these lines, 22.8% since 2005. In contrast, the bottom
the report draws a number of conclusions and 90% of metro areas (343 of them) lost share.
Innovation sector job creation has been strongest in metro areas that already
FIGURE 2
FIGURE 2
have the largest sectors
Innovation sector employment index (2005 = 100), 2005-17
115
115
110
110
105
105
100
100
95
95
90
90 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
Bottom 75% Next 15% Next 5% Top 5%
Bottom 75% Next 15% Next 5% Top 5%
1RWH 3HUFHQWDJHkELQVyUHpHFWFRKRUWVRIPHWURDUHDVUDQNHGLQHDFK\HDU7KHkWRSyLQFOXGHVPHWURV
Source: Brookings and ITIF analysis of Emsi data
THE CASE FOR GROWTH CENTERS 5
Metros by change in share of total innovation sector jobs
Share of innovation sector
jobs change, 2005-17
0.4% - 2.0%
0.0% - 0.4%
0.0%
-0.1% - 0.0%
-0.7% - -0.1%
Innovation sector
jobs, 2005
Top 5% of metros
Next 5%
Next 15%
Bottom 75%
Source: Brookings and ITIF analysis of Emsi data
$VDUHVXOWWKH86LQQRYDWLRQLQGXVWU\ is the stark gap between the productivity of
has become heavily entrenched in just a the relatively few metropolitan areas with
few places. Fully one-third of the nation’s high shares of innovation industries and
innovation jobs now reside in just 16 counties, the many more with less. These patterns
and more than half are concentrated in 41 are hurting the country’s innovation-based
counties. competitiveness, since the skyrocketing costs
of the most successful tech hubs mean that
All of this points to the extent to which tech investment is often made in other places—
innovation-sector dynamics compound over but not in other parts of America, given the
time, leaving most places falling further shortage of vibrant lower-cost hubs. The
behind. UHVXOWLVWKDWLQYHVWPHQWVpRZWRSODFHVVXFK
as Bangalore, Shanghai, Taipei, or Vancouver,
2. Such high levels of territorial polarization rather than Indianapolis, Detroit, or Kansas
are now a grave national problem. At the City.
economic end of the equation, the costs of
excessive tech concentration are creating Equally concerning is the fact that the
serious negative externalities. These range nation’s divergence is unfair. So many
IURPVSLUDOLQJKRPHSULFHVDQGWUDIoFJULGORFN Americans reside far from the opportunities
in the superstar hubs to a problematic associated with the nation’s innovation
“sorting” of workers, with college-educated centers, undercutting economic inclusion
workers clustering in the star cities and and raising social justice issues. Regional
leaving other metro ares to make do with divergence is also clearly driving “backlash”
thinner talent reservoirs. As a result, whole political dynamics that are exacerbating the
portions of the nation may now be falling into nation’s policy stalemates.
“traps” of underdevelopment. Of concern here
THE CASE FOR GROWTH CENTERS 6
Localization economies make innovation industries clustered together more
FIGURE 4
productive
Avg. output per worker by industry group, 2017
$400,000
$371,306
$350,000
$300,000
$264,119 $260,658
$246,265
$250,000
$200,000
$150,000
$100,000
$50,000
$0
Bottom 75% Next 15% Next 5% Top 5%
Basic manufacturing Retail Healthcare Finance Innovation industries
1RWH 3HUFHQWDJHkELQVyUHpHFWFRKRUWVRIPHWURDUHDVUDQNHGLQHDFK\HDU7KHkWRSyLQFOXGHVPHWURV
Source: Brookings and ITIF analysis of Emsi data
3. Markets alone won’t solve the problem; Moreover, “bottom-up” technology-based
place-based interventions will be essential economic development efforts cannot
in ameliorating it. When the economy was VLJQLoFDQWO\FKDQJHWKHVHSDWWHUQVE\
“converging,” it was easy to assume that themselves, in part because the resources
any problems of regional unevenness would states and cities can bring to bear are limited.
naturally resolve themselves. Indeed, until $FFRUGLQJO\WKH86QHHGVQRWMXVWQDWLRQ
very recently, self-correction remained the scaled solutions for its regional imbalances but
expectation of mainstream economists, place-based ones.
with their embrace of traditional doctrines
RIkDOORFDWLYHHIoFLHQF\ykHTXLOLEULXPy 4. The nation should counter regional
and “welfare-maximizing” spatial results. divergence by designating eight to 10
However, the rise of newer innovation-oriented new regional “growth centers” across
economic theories has thrown more attention the heartland. The time is right for, among
onto the power of local “agglomeration” other initiatives, a 21st century comeback
HIIHFWVE\ZKLFKODUJHEHQHoWVDFFUXHWR and update of “growth pole” strategy—the
oUPVZKHQWKH\ORFDWHWRJHWKHULQXUEDQ 1960s and 1970s emphasis in regional
areas. Substantial evidence now suggests economic planning that called for focusing
that agglomeration brings with it strong transformative investment on a limited
self-reinforcing tendencies that not only do number of locations to catalyze the takeoff of
not support the “spread” of development, those regions and the nation. What is needed
but are likely to exacerbate its concentration. in this respect will be: Generous awards of key
THE CASE FOR GROWTH CENTERS 7
federal innovation inputs (including support 5. Numerous metropolitan areas in most
IRUVFLHQWLoFDQGHQJLQHHULQJUHVHDUFK regions have the potential to become one of
UHJXODWRU\EHQHoWVDQGVXSSRUWVIRUKLJK America’s next dynamic innovation centers.
quality placemaking) coupled with a rigorous Skeptics may doubt that eight to 10 metro
and competitive selection process to identify areas worthy of growth center investment
the most promising locations for intervention. FDQEHLGHQWLoHGDQGFDWDO\]HGIRUkWDNHRIIy
and self-sustaining growth. However, even
Along these lines, the federal government a fairly restrictive list of eligibility criteria
should: yields plenty of potential candidates. Based
on a demonstration in this report, some 35
• Assemble a major package of federal potentially transformative metro areas surface
innovation inputs and supports as candidates for growth center designation.
for innovation-sector scale-up in Candidates are situated in at least 19 states,
metropolitan areas distant from existing lie in multiple regions (especially the Great
tech hubs. Central to this package will be a /DNHV8SSHU6RXWKDQG,QWHUPRXQWDLQ:HVW
direct R&D funding surge worth up to $700 and exist often at far remove from the coastal
million a year in each metro area for 10 superstars.
\HDUV%H\RQGWKDWZLOOEHVLJQLoFDQWLQSXWV
such as workforce development funding, Many more promising metro areas exist.
WD[DQGUHJXODWRU\EHQHoWVEXVLQHVV There is likely a score of “up-and-coming”
oQDQFLQJHFRQRPLFLQFOXVLRQ, and federal metro areas that hold a solid capacity for
land and infrastructure supports. The countering the nation’s regional divides by
increasing preference of innovative people bringing tech-based development closer to
and companies for mixed-use downtowns, the nation’s left-behind places.
waterfront areas, and urban “innovation
districts” means that federal contributions *
to urban placemaking also should be
prominent. To be sure, there will be objections. Some will
say the present proposal goes way too far, while
Overall, a rough estimate of the cost of such others will say it doesn’t go far enough.
a program suggests that a growth centers
surge focused on 10 metro areas would cost 7RWKHoUVWSRLQWPDQ\FRQYHQWLRQDOHFRQRPLVWV
the federal government on the order of $100 will argue that any such push to promote
billion over 10 years. That is substantially regional equity will come at the expense of
OHVVWKDQWKH\HDUFRVWRI86IRVVLOIXHO HIoFLHQF\+RZHYHUEHFDXVHRIERWKWKHQHJDWLYH
subsidies. externalities from growth in booming tech
hubs and the positive externalities of growth in
• Establish a competitive, fair, and rigorous targeted emerging hubs, intervention can help
process for selecting the most promising underperforming metro areas turn the corner,
potential growth centers to receive escape a cumulative causation trap, and add to
the federal investment. To distribute its the nation’s total welfare, including its global
supports, the proposed growth center competitiveness. Other critics will deny the
program would select for awards the eight ability of the federal government to effectively
to 10 metropolitan areas that had best pick regional “winners” or reject that the
demonstrated their readiness to become a emergence of existing clusters had anything to
new heartland growth center. The process do with government efforts. But one has only to
would employ a rigorous competition H[DPLQHWKHKLVWRU\RI86WHFKQRORJ\KXEVVXFK
characterized by an RFP-driven challenge, as Boston, the Bay Area, and North Carolina’s
goal-driven criteria, and an independent Research Triangle to see that the federal
selection process.
THE CASE FOR GROWTH CENTERS 8
Strong potential candidates for Growth Center designation appear across the
country
Source: Brookings and ITIF analysis of Emsi data
government has often played important, if not supply chain relationships, commuting, and other
decisive, roles in helping new tech centers attain interdependencies with the growth centers.
critical mass. In that spirit, then, the present initiative is best
viewed as but one component of the full federal
To the other point, others may say that a agenda needed to ameliorate the nation’s
JURZWKFHQWHUVSXVKGRHVQRWVXIoFLHQWO\ unbalanced economic geography.
“change capitalism” or address the full crisis
of America’s smaller cities, towns, and rural As such, a concerted growth centers surge—while
areas. And certainly that is true. There is much not a total solution of the nation’s now-acute set
more that needs doing, especially for the of regional imbalances—would represent a major
most deeply struggling communities. But the break with past inaction and demonstrate that
proposed innovation surge would absolutely federal action can not only bring technology-
begin to transform the nation’s spatial malaise. based opportunity to more parts of the nation,
Most notably, it would bring new vitality closer EXWDOVRVSXUPRUHLQQRYDWLRQDQGLQFUHDVHG86
to more struggling communities, allowing for economic competitiveness.
VPDOOHUWRZQVDQGFRXQWLHVWREHQHoWWKURXJK
THE CASE FOR GROWTH CENTERS 9
Introduction
T
he stark facts of the widening regional the 20th century, after all, market forces tended
GLYLGHLQ$PHULFDoQDOO\KLWKRPHLQWKH to reduce regional disparities in employment,
aftermath the 2016 presidential election. wages, investment, and business formation. By
narrowing these divides, the economy ensured a
As nothing before it did, the election exposed a welcome “convergence.”2 Midsized and smaller
deep fault line: the widening gap between two cities, as well as once-lagging regions such as
separate Americas, one centered in the nation’s the South, were catching up with richer, bigger
dynamic, high-tech “superstar” metropolitan places.
areas, and the other encompassing pretty much
everywhere else. However, in the 1980s, that trend began to break
down.
Rather than growing together, the nation’s
communities are growing apart—and that has As innovation moved to the center of economic
been a shock, especially for an economic and activity, the advance of digital technology
policy mainstream that has long trusted the increasingly rewarded the most talent-laden local
self-regulating nature of the market.1 For much of FOXVWHUVRIVNLOOVDQGoUPV,QWHQVHQHZGHPDQGV
THE CASE FOR GROWTH CENTERS 10Figure 1. Since the 1980s, wage and employment growth convergence among
metro areas has broken down
Indexed average annual wages Indexed employment level
1969 - 2017 (1969 = 100) 1969 - 2017 (1969 = 100)
180 280
170
250
160
220
150
140 190
130
160
120
130
110
100 100
Bottom third of metro areas Median metro areas Top 2% of metro areas
Source: Brookings analysis of BEA data
for talent and insights increased the dominance hubs such as Boston, San Francisco, and Seattle—
of “agglomeration” economies, unleashing forces along with smaller hubs including Austin,
WKDWEHQHoWHGELJFRDVWDOFRUHUHJLRQVRIWHQWR Denver, Raleigh-Durham and San Diego, as well
the detriment of the midsized cities and smaller DVoQDQFLDODQGFRQWHQWRULHQWHGPHJDPDUNHWV
“heartland” towns that had found manufacturing- such as New York and Los Angeles—have pulled
based prosperity in the 20th century. away and secured themselves as America’s core
domain of advanced industry activity.4 These
Amid these conditions, convergence gave way to SODFHVHQMR\WKHEHQHoWVRIZKDWHFRQRPLVWVFDOO
“divergence.” Over time, a fortunate upper tier cumulative causation, through which their earlier
of big, techy metro areas (about 20 in all) began NQRZOHGJHDQGoUPDGYDQWDJHVQRZDWWUDFWHYHQ
to consistently grow faster than the median and more talented workers, startups, and investment,
less-prosperous ones. creating a gravitational pull toward the nation’s
critical innovation sectors while simultaneously
By the 2010s, a clear hierarchy of economic draining key talent and business activity from
performance based on innovation capacity had other places.
become deeply entrenched.3 Large innovation
THE CASE FOR GROWTH CENTERS 11TERMINOLOGY
“Advanced industries” are America’s 50 high-R&D, high- “Superstar cities” are the 20 metropolitan areas with the
STEM manufacturing, energy, and services industries, largest absolute number of jobs in innovation industries.
ranging from aerospace and automobile manufacturing to These mostly coastal, high-tech metro areas represent
solar electricity generation to internet publishing to bio- about 5% of the nation’s metropolitan areas and are almost
tech. all growing relatively rapidly.
“Innovation industries” are a subset of advanced Source: Mark Muro and others, “America’s Advanced
industries that include the 13 most STEM- and R&D- Industries: What They Are, Where They Are, and Why They
intensive industries. Matter.” (Washington: Brookings Institution, 2015).
In contrast, most midsized cities and smaller For the many metro areas that are going
towns (let alone rural areas) have struggled to sideways or declining, the struggle to keep up has
make progress in amassing critical innovation brought a mood of desperation as their fears of
industries, with many falling farther behind. For VWDJQDWLRQKDYHEHHQUHDOL]HG3XEOLFRIoFLDOV
most of them, “innovation” or advanced-industry- working families, and business and development
sector employment and incomes have declined. leaders in these places are often deeply
Many such places have been left to cope with frustrated and at a loss for what to do to get
brain drain, the hollowing out of the labor market, onto the prosperity track—in part because many
and industrial decline. of the most well-intentioned local interventions,
such as tech-based economic development
As a result, few can now deny that the programs, have proven incomplete on their own.
imbalanced geography of America’s current Indeed, since these efforts are all chasing a
economy is spawning disturbing negative limited amount of innovation activity, they almost
externalities—or side effects—that cry out for all are falling short in their quest to reach the
response. needed “critical mass” to succeed, while the less
fortunate of them face the corrosion of drift,
Among the superstar metros, the “winner-take- brain drain, and lost capacity.
most” dynamics of the innovation economy
have led to dominance thanks to their deep $QGIRUWKHQDWLRQDVDZKROHoQDOO\WKH
SRROVRIWDOHQWDGYDQFHGoUPVDQGFDSLWDO%XW juxtaposition between a few dominant
those dynamics have also spawned tumultuous innovation stars and the many other places
OLYDELOLW\FULVHV VSLUDOLQJUHDOHVWDWHFRVWVWUDIoF trending sideways or downward raises broader
gridlock, and homelessness.5 Firms in these problems. For the aggregate economy, there
metro areas increasingly move activity elsewhere, is the likelihood that while the importance of
or launch it in other places, although too often star agglomerations remains unquestionable,
the move is to lower-cost overseas tech hubs, WRRPXFKLVEHLQJVDFULoFHGWKURXJKWKH
rather than equally cost-competitive centers underperformance of so many less successful
LQWKH8QLWHG6WDWHV6XFKG\QDPLFVUHSUHVHQW regions.7
WUXHGLVHFRQRPLHVRIVFDOHDQGGUDJVRQ86
competitiveness.6
THE CASE FOR GROWTH CENTERS 12At the same time, civic and political leaders have
grown deeply concerned about the social and
political side effects of such unevenness.8 For
Few can now deny that the
some, the isolation of millions of people from imbalanced geography of America’s
the dynamic innovation economies of superstar
metro areas is a justice issue.9 For others, the
current economy is spawning
fear looms that the widening gap between the disturbing negative externalities—
superstars and everybody else is eroding the
national ideal of equal access to opportunity.10 or side effects—that cry out for
response.
All of which raises the question of what ought
to be done about the concentration of the
innovation sector, and the regional imbalances it
entails. tech cities and the drift of most everywhere
else seem tenable. Instead, a growing number
Historically, the convergence trend of the of voices are concluding that a new set of
economy minimized worry about the nation’s viewpoints and policies are needed to actively
uneven economic geography. Wage convergence— respond to the imbalances of today’s innovation
companies in high-wage areas moving to lower- map—not only to heal geographic divides, but
wage ones—supported the belief that regional WRDGYDQFHWKH86 VLQQRYDWLRQEDVHGJOREDO
imbalances would naturally even out, thanks to competitiveness.
WKHLQKHUHQWO\HIoFLHQWDQGZHOIDUHPD[LPL]LQJ
nature of the market economy. However, more $QG\HWIRUDOORIWKDWIHZVSHFLoFDQG
recently, other voices have noted the powerful substantial ideas for countering the winner-take-
positive links of innovation and agglomeration, most dynamics of the innovation sector have
and warned that efforts to reduce variation been forthcoming. To the extent any are offered,
across places might be detrimental to the they mostly resemble lectures on self-help, based
innovation sector’s ability to drive national on the opinion that if local leaders were just a bit
productivity.11 more creative, they could overcome the powerful
forces of cumulative causation on their own.14
Accordingly, mainstream economists have until
recently remained largely unphased by the Which is where this report comes in. Concerned
nation’s spatial divides and thus skeptical of ideas WKDWWKHGLYHUJHQFHRI86PHWURDUHDVKDV
to counter it.12 Such scholars would have the become self-reinforcing and destructive, the
nation rely mainly on spurring migration from discussion that follows presumes that the time
lagging to leading regions, ignoring the fact that has come for the nation to offset the domination
most leading regions are already bursting at the of innovation superstars with a concerted
seams—not to mention that many Americans intervention to support the emergence of at
retain deeply rooted ties to their home places. least a few new stars in new places. As such, this
brief urges the federal government to undertake
In recent years, however, the heightened a major new effort to counter divergence with
awareness of regional imbalances, especially a robust set of carefully targeted innovation,
in the innovation sector, has prompted new business-development, placemaking, and
responses. Economists, investors, politicians, related investments of a type only it can deliver
journalists, and economic development leaders systematically. This effort would at once add to
are all beginning to reassess the costs of inaction the nation’s net innovation effort (as is sorely
after decades of exactly that.13 No longer does needed), spur global competitiveness, and push
nonchalance about the pull-away of superstar back against the nation’s dangerous economic
divides.15
THE CASE FOR GROWTH CENTERS 13markets acting alone won’t solve the divergence
problem, and proceeds to argue why robust
The transformation of even a short intervention will be necessary.
list of metro areas in new, interior From there, the report plays out one possible
UHVSRQVHWRWKHH[FHVVLYHFRQFHQWUDWLRQRI86
areas could be expected to improve
innovation activity in superstar cities: a plan for
the fortunes of whole regions building up dynamic new growth centers in a
set of promising metropolitan areas that have
modest but not dominant existing strengths.
Section 6 of the report articulates a vision for
In pursuing this vision, the report assumes that updating the concept of “growth poles” as a
while the innovation economy can be more theory for creating more innovation centers
geographically dispersed, not every place can across the nation, and Section 7 lays out what
become an advanced industry hub. In fact, a federal, innovation-based growth centers
the following discussion suggests that only a program might look like, including how the
modest number of places are likely to be able to government might select eight to 10 promising
transform themselves into self-sustaining tech metro areas and provide them with a host of
hubs with large-scale (but temporary) help from innovation inputs in order to become strong, self-
the federal government, if for no other reason sustaining technology hubs.
than the innovation economy is not big enough
to act as the engine of growth for an extremely After that, Section 8 shows how one possible set
large number of places. At the same time, the of selection criteria for growth center contestants
transformation of even a short list of metro presents an array of up-and-coming locations
areas in new, interior areas could be expected to whose accelerated emergence could improve
improve the fortunes of whole regions, opening on the nation’s imbalanced geography. Section
up new possibilities for intraregional worker 9 anticipates and responds to the possible
mobility and deeper supply chain ties. This objections, and Section 10 concludes the report.
transformation should be vigorously pursued,
because it will affect not only the targeted cities To be sure, ours is not a strategy for mitigating
but surrounding communities as well. the full enormity of the nation’s regional
imbalances, or for jump-starting scores of worthy
Accordingly, the paper urges the federal metro areas (that is the even more expansive
government to update a forgotten 1960s and vision espoused in Jonathan Gruber and Simon
1970s strategy—growth poles—for transforming Johnson’s book, Jump-Starting America).16 Our
promising places by fusing it with modern “growth centers” proposal, rather, is focused on
methods. Metropolitan areas would compete accelerating the growth of the most promising yet
WRZLQPDVVLYHLQpRZVRIIHGHUDOUHVHDUFK lagging metropolitan areas in the nation’s interior,
business development, and placemaking inputs rather than on saving the most distressed cities
DQGEHQHoWVLQH[FKDQJHIRULPSOHPHQWLQJEROG and towns across the vast totality of America’s
innovation, economic growth, and transformation economy. That too needs doing, but that work
strategies. lies beyond the scope of this proposal. Instead,
we focus here on the dangerous core problem
In that vein, the discussion here begins by of the hyperconcentration of the innovation
reviewing the entrenched nature of America’s and advanced industries economy, accepting
concentrated innovation sector and why it is that even on that narrower front it is simply not
a problem. After that, the report explains why possible to “target” everywhere at once.
THE CASE FOR GROWTH CENTERS 14The focus of this report is on distributing high- strategy and promising potential participants
quality economic development more widely exist for spreading innovation-driven growth
DFURVVWKH8QLWHG6WDWHVLQSDUWE\VWLPXODWLQJ farther across the nation’s map. Individual metro
new innovative activity in the most promising areas, whole new regions including smaller
metro areas proximate to more places in the nearby communities, and the nation as a whole
country’s interior. Such an innovation surge is ZRXOGEHQHoWIURPVXFKDQLQLWLDWLYH0RUHRYHULI
only one part of the needed push to counter America is to avoid ceding its innovation lead to
the nation’s regional divergence, but it should China, it is vital that we create more innovation
be counted as an important element of such a hubs in America, if for no other reason than to
campaign. Such a campaign has a good chance JLYHDGYDQFHGWHFKQRORJ\oUPVLQWKH8QLWHG
of beginning to stem the epidemic of inequality States an alternative to places like Shanghai and
in the nation by helping places beyond the Shenzhen for their future growth.
immediate targets of action, which would likely
EHQHoWIURPWKHSURJUDPWKURXJKVXSSO\FKDLQ The same is true if we are going to cease
labor market, and other spillover gains. growing apart and begin to grow together again.
Accelerating the emergence of 10 new growth
In sum, this report argues that the moment is centers in the heartland would help with that.
urgent and the prospects favorable for launching It is time, then, for Congress and a farsighted
a major national push to counter regional administration to embrace the present challenge
divergence—and with it, the nation’s current crisis and opportunity.
of economic and social inequality. Both a feasible
THE CASE FOR GROWTH CENTERS 152. The entrenched geography of America’s innovation
industries
T
he steady narrowing of economic $VWKH86SURGXFWLRQV\VWHPJUHZDIWHU:RUOG
GLVSDULWLHVDPRQJ86UHJLRQVWKURXJK War II—powered by nationwide telephony,
most of the last century licensed a strong the interstate highway system, air travel
FRQoGHQFHDPRQJHFRQRPLVWVWKDWUHJLRQDOJDSV and, importantly for the South and West, air
would “naturally” close.17 Initially as a fact and conditioning—manufacturing and corporate
then as a reassuring orthodoxy, the conventional functions that had historically been centralized
wisdom assumed that even seriously lagging in the Midwest and Northeast now had many
regions would “catch-up” to leaders through more options for where they could locate. Given
natural equilibrium processes, as the “costs” of that many of the mass production industries of
success—such as increased labor or real estate the time had evolved to compete primarily on
costs—accumulated and eventually motivated costs, it made sense that as costs increased in
oUPVDQGZRUNHUVWRUHORFDWHWRORZHUFRVW some regions, industry would grow faster in less-
regions. expensive regions, which would then “catch-up.”
THE CASE FOR GROWTH CENTERS 16For years, regional trends comported with such DIVERGENCE AT WORK IN THE
neoclassical theory. From 1880 to 1980, incomes INNOVATION SECTOR
across states “converged” at a rate of 1.8% a
year.18 In other words, low-income states grew The geography of the nation’s most important
faster than high-income ones. Wages in poorer industries—its core “innovation industries,” which
metropolitan areas likewise grew 1.4% faster are a key portion of its 50 most “advanced”
than those in higher-wage metro areas between industries—epitomizes the new dynamic.22
1940 and 1980.19 These trends, coupled with the America’s innovation sector has, in truth, seen
simplistic, cost-based version of neoclassical few triumphal “catch-up” stories and very little
economics that many economists subscribed to, diffusion into new places. In fact, despite four
licensed great optimism about the self-regulating decades of state and local technology-based
nature of the nation’s regional dynamics. economic development policies, what we call
WKH86kLQQRYDWLRQVHFWRUy UDQJLQJIURP
Nor has that hands-off view of regional pharmaceutical manufacturing and aerospace
imbalance fully dissipated, even as convergence products to software publishing and data
broke down in recent decades. Right up to the services) has remained strikingly immobile and
election of President Trump, most economists highly concentrated in a short list of “superstar”
seemed to have assumed that market forces metropolitan areas for many years.
would still correct for harmful variations among
places and lead to greater regional balance, at This concentration is important because the
least in time.20 At most, they held, the federal innovation sector—consisting of 13 of the
government could provide free bus tickets for nation’s highest-R&D industries (at the four-
auto workers in Flint, Mich. to move to Silicon digit NAICS code level)—matters inordinately
Valley, where they could be retrained as software for the nation’s competitiveness and prosperity,
engineers. given that innovation industries encompass
the nation’s “tech” sector at its most dynamic,
And yet, market forces haven’t reduced such competitive, and valuable.23 As a group, these
gaps in recent decades, and especially not 13 industries—which also include chemicals,
when it comes to the geography of the nation’s computer equipment manufacturing, telecom,
innovation sector. Nor has the problem been a DQG5 'VHUYLFHVLQRUGLQDWHO\FRQWULEXWHWR86
lack of bus tickets. Instead, the welcome reality prosperity. While innovation industries account
of convergence has been faltering and in fact for just 3% of the nation’s jobs, they generate 6%
shifting towards divergence—the polarization of of the nation’s GDP, a quarter of its exports, and
places. Several careful researchers have looked two-thirds of business R&D expenditures. They
across the economy and tracked dramatic also support solid economic multipliers in their
declines of income and other measures of regions (and nationally), and provide especially
prosperity across states and metropolitan area well-paying jobs even for workers without a
in the years 1980 until now.21 This should not be bachelor’s degree.24
surprising, given that the innovation industries of
the last 40 years largely do not compete on cost :KLFKLVZK\LWZRXOGEHH[WUHPHO\EHQHoFLDO
but rather on the richness of the local innovation if more of the nation’s innovation activity was
ecosystem, where success begets more success. diffusing outward, so that more metropolitan
But the change, while slow to be fully recognized, areas were joining the nation’s assortment of
has at last been recognized as an emergency.
THE CASE FOR GROWTH CENTERS 17“INNOVATION INDUSTRIES” AND “SUPERSTAR METRO AREAS” DEFINED
The “innovation sector” as discussed here is an especially “Superstar metro areas” are the 20 metropolitan areas
high-tech subsector of the “advanced industries” sector, an with the largest absolute numbers of jobs in innovation
earlier delineation of America’s highest-value industries by industries. These mostly coastal, high-tech places represent
the Metropolitan Policy Program at Brookings. about 5% of the nation’s metropolitan areas and are almost
all growing relatively rapidly. The 20 superstar metro areas
“Innovation industries” encompass America’s 13 highest- are:
tech, highest-R&D industries. Selected from among the 50
advanced industries, the 13 innovation industries represent • New York-Newark-Jersey City, NY-NJ-PA
a cohort whose R&D expenditures exceed $20,000 per • San Jose-Sunnyvale-Santa Clara, CA
worker and have a STEM-worker share of 45%. The 13 • Los Angeles-Long Beach-Anaheim, CA
innovation industries include: • Seattle-Tacoma-Bellevue, WA
• Boston-Cambridge-Newton, MA-NH
• Basic chemical manufacturing • San Francisco-Oakland-Hayward, CA
• Pesticide, fertilizer, and agricultural chemical • Dallas-Fort Worth-Arlington, TX
manufacturing • Washington-Arlington-Alexandria, DC-VA-MD-WV
• Pharmaceutical and medicine manufacturing • San Diego-Carlsbad, CA
• Computer and peripheral equipment manufacturing • Chicago-Naperville-Elgin, IL-IN-WI
• Communications equipment manufacturing • Philadelphia-Camden-Wilmington, PA-NJ-DE-MD
• Semiconductor and other electronic components • Phoenix-Mesa-Scottsdale, AZ
manufacturing • Minneapolis-St. Paul-Bloomington, MN-WI
• Navigational, measuring, electromedical, and control • Houston-The Woodlands-Sugar Land, TX
instruments manufacturing • Portland-Vancouver-Hillsboro, OR-WA
• Aerospace product and parts manufacturing • Atlanta-Sandy Springs-Roswell, GA
• Software publishers • Austin-Round Rock, TX
• Satellite telecommunications • St. Louis, MO-IL
• Data processing, hosting, and related services • Denver-Aurora-Lakewood, CO
• Other information services • Miami-Fort Lauderdale-West Palm Beach, FL
• 6FLHQWLoFUHVHDUFKDQGGHYHORSPHQWVHUYLFHV
Source: Mark Muro and others, “America’s Advanced
Industries: What They Are, Where They Are, and Why They
Matter.” (Washington: Brookings Institution, 2015).
THE CASE FOR GROWTH CENTERS 18innovation hubs. However, that’s not what has (localization), and take advantage of dense air
been happening. Instead, strong centripetal and ground transport links (urbanization).
forces—the so-called “localization” and
“urbanization”25 dynamics of agglomeration— Map 1 illustrates the strength of agglomeration
have produced a remarkably entrenched in the innovation sector. Fully one-third of the
LQQRYDWLRQVHFWRUJHRJUDSK\LQWKH8QLWHG nation’s innovation jobs reside in just the 16
States. Two industry-level examples are the counties (those shaded dark blue) that contain
nation’s world-class life sciences and digital 1% or more each of the nation’s innovation
services sectors, which have remained extremely employment.
concentrated in a relatively short list of
metropolitan areas.26%HFDXVHoUPVLQVXFK For that matter, more than half of the nation’s
innovation industries compete on the basis of innovation jobs are concentrated in just the 41
product and process innovation rather than cost- counties with at least 0.5% of the jobs in the
minimization, and are more reliant on knowledge innovation sector. That these counties account for
than other industries, they tend to cluster in less than 27% of the nation’s aggregate job total
large metropolitan areas where they can tap underscores the strong tendency of innovation
specialized workers, suppliers, and institutions industries to cluster in select large regions where
related activity is already taking place.27
Map 1. U.S. counties by share of total innovation sector jobs, 2017
1.0% - 5.9%
0.5% - 1.0%
0.0% - 0.5%
Source: Brookings and ITIF analysis of Emsi data
THE CASE FOR GROWTH CENTERS 19Figure 2. Innovation sector job creation has been strongest in metro areas that
already have the largest sectors
FIGURE 2
Innovation sector employment index (2005 = 100), 2005-17
115
110
FIGURE
URE 3 3 105
2.0%
1.5% 100
1.5%
1.5%
1.0%
95 1.5%
1.5% 0.7%
0.5%
90 0.7%
2005 0.7%
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
0.0%
0.7%
Bottom 75% Next 15% Next 5% Top 5%
-0.5%
1RWH 3HUFHQWDJHkELQVyUHpHFWFRKRUWVRIPHWURDUHDVUDQNHGLQHDFK\HDU7KHkWRSyLQFOXGHVPHWURV
Source: Brookings and ITIF analysis of Emsi data
1.0%
-1.0%
1.0%
% Figure 3. Innovation industries became more geographically concentrated since
FIGURE 3
-1.5% the start of the century
Change 2001 - 2005
in share of all innovation jobs in 2005 - 2012
metros, 2001-17 2012 - 2017
2001 - 2005 2005 - 2012 2012 - 2017
2.0% Bottom 75% Next 15% Next 5% Top 5%
2005 - 2012 2012 - 2017
Bottom 75% Next 15% Next 5% Top 5%
05 - 2012 2012 - 2017
Next 15% 1.5% Next 5% Top 5%
Next 5% Top 5% 1.5%
1.0%
0.7%
0.5%
0.0%
-0.5%
1.0%
-1.0%
-1.5%
2001 - 2005 2005 - 2012 2012 - 2017
1RWH 3HUFHQWDJHkELQVyUHpHFWFRKRUWVRIPHWURDUHDVUDQNHGLQHDFK\HDU7KHkWRSyLQFOXGHVPHWURV
Bottom
Source: Brookings and 75% of Emsi data
ITIF analysis Next 15% Next 5% Top 5%
THE CASE FOR GROWTH CENTERS 20Looking beneath the snapshot furnished by Map widening—driven in particular by the pull-away
1, Figures 2, 3, 4, and 5 explore the underlying of the top 10% of innovation-job-rich metro
innovation sector dynamics driving the nation’s areas. For the 90% of metro areas with fewer
remarkable degree of agglomeration at the innovation jobs, however, total employment has
metro-area level. effectively remained unchanged in the last 12
years.
To begin with, the line chart in Figure 2 makes
clear the innovation sector’s concentration has 7KHIROORZLQJEDUFKDUW )LJXUH SXWVDoQHU
become self-reinforcing and divergent over point on this divergence by looking at changes in
time, to the extent that the innovation “rich” are the share of total innovation jobs by metro-area
getting richer as the metro areas with the largest echelons.
sectors steadily gained additional innovation
employment while most metro areas went The top 5% of metro areas by innovation jobs,
sideways. depicted in dark blue, have increased their share
of the nation’s total innovation employment
Far from catching up, most metro areas have since 2005 while all other size cohorts have
been slipping farther behind. This tendency drifted or lost share (the 1 percentage point loss
has been strong and sharpening in the last 15 of innovation employment share in the 2001 to
years. Since 2005, as the line chart indicates, SHULRGODUJHO\UHpHFWVWKHEXVWRIWKH
the geographic divergence of tech-sector tech “bubble”).
employment across metro areas has been
Figure 4. Localization economies make innovation industries clustered together
FIGURE 4
more productive
Avg. output per worker by industry group, 2017
$400,000
$371,306
$350,000
$300,000
$264,119 $260,658
$246,265
$250,000
$200,000
$150,000
$100,000
$50,000
$0
Bottom 75% Next 15% Next 5% Top 5%
Basic manufacturing Retail Healthcare Finance Innovation industries
1RWH 3HUFHQWDJHkELQVyUHpHFWFRKRUWVRIPHWURDUHDVUDQNHGLQHDFK\HDU7KHkWRSyLQFOXGHVPHWURV
Source: Brookings and ITIF analysis of Emsi data
THE CASE FOR GROWTH CENTERS 21Figure 5. Urbanization economies benefit innovation industries located in larger
FIGURE 5
markets
Avg. output per worker by industry group, 2017
$450,000
$400,000 $382,291
$344,717
$350,000
$300,000
$263,958
$250,138
$250,000
$200,000
$150,000
$100,000
$50,000
$0
Bottom 75% Next 15% Next 5% Top 5%
Basic manufacturing Retail Healthcare Finance Innovation industries
1RWH 3HUFHQWDJHkELQVyUHpHFWFRKRUWVRIPHWURDUHDVUDQNHGLQHDFK\HDU7KHkWRSyLQFOXGHVPHWURV
Source: Brookings and ITIF analysis of Emsi data
The top 20’s share of all metro-area innovation Among metro areas, the average output per
jobs rose from 59% to nearly 61% in the period worker in innovation industries rises with both
since 2001, even though they accounted for only the absolute number of jobs in the sector
45% of metro-area jobs across all industries. in a given metro area (Figure 4UHpHFWLQJ
Moreover, the gains among second-tier metro “localization” economies) and the total number of
areas in the early 2000s were more than offset jobs in the metro area overall (Figure 5UHpHFWLQJ
by innovation sector employment concentration “urbanization” economies).
in the top 20 between 2005 and 2017—a 2.2
percentage point increase among the top 20 at Average innovation sector labor productivity
the expense of the bottom 90% (equal to 343 in the top 5% of metro areas with the most
metro areas). By 2017, two-thirds of innovation innovation sector jobs—plotted to the far right,
sector output originated in those same top 20 UHpHFWLQJDPHUHSODFHVLVPRUHWKDQ
metro areas, once again pointing to the clear higher than it is in the 75% of metro areas
productivity advantages they continue to confer plotted to the left, which have the fewest
RQoUPVWKHUH innovation jobs and encompass 287 different
metro areas. The magnitude of this sector-level
Figure 4 and Figure 5 show the underlying disparity is similar when metro areas are grouped
productivity dynamics driving this remarkable by total employment.
degree of agglomeration at the metro-area level.
THE CASE FOR GROWTH CENTERS 22By contrast, this relationship appears much GHFDGH$VVXFKWKHoYHPHWURDUHDVFROOHFWLYHO\
weaker for less innovation-centric industry increased their aggregate share of the nation’s
groups such as basic manufacturing and health innovation jobs by 5.2 percentage points, from
FDUH ZLWKWKHH[FHSWLRQRIoQDQFHDQRWKHU 17.6% in 2005 to 22.8% in 2017. More broadly,
traded service sector). some 40 of the largest 100 metro areas increased
their share of the sector, although the gains were
Mapping some of these patterns further PLQLVFXOHRXWVLGHWKHWRSoYH
highlights how a truly small set of “superstar”
cities has been determining the nation’s By contrast, 60 of the largest metro areas
innovation geography. In this vein, Map 2 shows lost ground, with many seeing quite dramatic
that a very short list of large coastal metropolitan shrinkages of their participation in America’s
areas has substantially increased its preeminence innovation economy. In this respect, no fewer
in innovation industries since 2005, while a very WKDQRXWRI86PHWURDUHDVKDYHVHHQ
long list of metro areas actually lost ground. (For their share of the national innovation sector
metro-area statistics see Appendix A). decline since 2005. What’s more, 191 metro areas
actually shed innovation sector jobs during
2YHUDOOMXVWoYHkVXSHUVWDUyPHWURDUHDV the time period. For instance, absolute local
Boston, the San Francisco Bay area (San employment in the innovation sector fell by
Francisco and San Jose), Seattle, and San over 20% between 2005 and 2017 in Colorado
DiegoDFFRXQWHGIRUVRPH RIDOO86 Springs, Colo., Providence, R.I., Sacramento,
innovation-sector growth between 2005 and Calif., Albuquerque, N.M., and Wichita, Kan.
2017. These gains mostly transpired in the digital To be sure, several interior metro areas such as
and biopharmaceutical expansion of the last Madison, Wis., Raleigh, N.C., Atlanta, Denver,
Map 2. Only a handful of superstar metro areas have seen their share of
innovation jobs increase since 2005
Metros by change in share of total innovation sector jobs
Share of innovation sector
jobs change, 2005-17
0.4% - 2.0%
0.0% - 0.4%
0.0%
-0.1% - 0.0%
-0.7% - -0.1%
Innovation sector
jobs, 2005
Top 5% of metros
Next 5%
Next 15%
Bottom 75%
Source: Brookings and ITIF analysis of Emsi data
THE CASE FOR GROWTH CENTERS 23Salt Lake City, and Provo, Utah have increased WKHLUVKDUHRI86LQQRYDWLRQMREVHYHQDVWKH superstars boomed. But even so, most metro 1RIHZHUWKDQRXWRI86 areas have gone sideways or lost innovation share. In the Midwest and South, multiple metro areas have seen their share metro areas with solid industry, university, and of the national innovation sector workforce assets such as Des Moines, Iowa, Charlotte, N.C. and Minneapolis have added decline since 2005. innovation jobs, but failed to increase their share of the national sector, while metro areas such Chicago and Wichita have both shed innovation jobs and national share. More broadly, most of the nation’s major business hubs—including Los Angeles, Dallas, Washington, D.C., Philadelphia, New York, and Houston—have lost purchase in the sector. All of this points to the extent to which the pull of agglomeration compounds over time in the innovation sector, leaving most places behind and putting those metro areas that lack a self- sustaining critical mass of innovation sector activity at an increasing disadvantage. THE CASE FOR GROWTH CENTERS 24
3. The costs of hyperconcentration
W
hy, though, is this intense concentration and services-led growth.29 At the same time,
of the nation’s innovation sector such a ameliorative subsidies or other interventions
problem? Might the nation’s unbalanced for places that lacked the right mix of size,
degree of spatial concentration be the optimal, skills, infrastructure, and amenities were seen—
PDUNHWRUGDLQHGJHRJUDSKLFDOFRQoJXUDWLRQIRU at least by federally oriented economists and
maximizing innovation? SROLF\PDNHUVDVLPSHGLPHQWVWRWKHHIoFLHQW
movement of capital and labor into dense urban
To be sure, intense, even extreme, geographic markets where they would receive their highest
concentration has long been viewed as inevitable, return.
benign, and mostly desirable for advanced
economies.28 In this regard, the work of building tech
agglomerations has fallen mostly to state
In recent decades, agglomeration economies—a and local policymakers, who have made do
feature of regional economics for more than a with only small denominations of funding and
century—have been understood to be the focal programs that often ended with each political
points of national prosperity in an era of tech- administration. At the federal level, meanwhile,
THE CASE FOR GROWTH CENTERS 25the proper role of policy was simply to facilitate workers now commute more than an hour to
factor mobility by focusing on eliminating work in San Francisco, San Jose, Seattle, and
frictions and lowering costs (especially housing) Boston, compared to just 6% in all metropolitan
in the country’s most productive metro regions.30 areas. Home prices have more than doubled in
Given the assumed big “trade-off” between the Bay Area over the same period, with median
HIoFLHQF\DQGHTXLW\WRGRDQ\WKLQJHOVHZRXOG rents for a one-bedroom apartment exceeding
be to compromise growth.31 $1,500 a month in both San Francisco and San
Jose.33 Indeed, such high housing prices mean
The present decade has, however, seen the LWLVQRZQRWXQFRPPRQWRoQGHYHQVRIWZDUH
EUHDNGRZQRIWKLVFRQoGHQWQDUUDWLYHDVVFKRODUV developers in the Bay Area living out of their cars
and policymakers have begun to recognize the or Winnebagos.34 Such externalities represent
untenable economic, social, and political costs of VL]DEOHGUDJVRQUHJLRQDODQGQDWLRQDOHIoFLHQF\
sustained territorial polarization and industrial
concentration. At the same time, the current reconsideration
of the laissez faire orthodoxy on geographical
ECONOMIC COSTS trends has been further driven by frustration with
economic stagnation in so many of America’s left-
At the economic end of the equation, the costs behind places. 35 In these places, shuttered plants,
of hyperconcentration in large innovation faded downtowns, and depopulated residential
agglomerations are impossible to ignore. Since neighborhoods exemplify the economic and
2010, average commute times have nearly social costs of regional imbalance. As such, the
doubled in San Jose, Calif. and increased by “winner-take-most” ascent of the superstar
two-thirds in San Francisco.32 Over 30% of
Map 3. Spiking housing costs are one major consequence of the hyperconcentration
of innovation sector jobs
Metros by housing affordability
Ratio of median household income
to median housing costs, 2017
12.8% - 15.0%
15.0% - 18.0%
18.0% - 21.0%
21.0% - 24.0%
24.0% - 28.3%
Population, 2017
Over 5,000,000
2,500,000 - 5,000,00
1,000,000 - 2,500,000
500,000 - 1,000,000
Less than 500,000
Source: Brookings and ITIF analysis of Emsi data
THE CASE FOR GROWTH CENTERS 26metro areas has increasingly been accompanied workers increasingly remain in places with
by the decline of virtually everywhere else. diminishing job prospects (both at home and
in the distant, more vibrant cities) and slow or
7KLVGHFOLQHLVQRZLPSRVLQJkHIoFLHQF\y stagnant wage growth.37
costs both on communities and the nation. For
starters, the divergent, striated economic map These patterns ensure that the nation’s overall
is beginning to affect the geographic sorting of ODERUPDUNHWLVEHFRPLQJLQHIoFLHQWO\DQG
workers, with negative impacts on the nation’s stubbornly ill-sorted. Metro areas heavily
overall welfare. While some economists believe oriented toward innovation work are able to
worker migration from low- to high-productivity attract and retain high-skill workers from out-of-
places will increase the nation’s aggregate state and abroad, adding to their stock of human
HIoFLHQF\UHFHQWVFKRODUVKLSHPSKDVL]HVWKDW capital. By contrast, less innovation-oriented
geographic mobility has slowed, limiting its ability metro areas are losing their more mobile, high-
WRHUDVHUHJLRQDOSURGXFWLYLW\GLYLGHV6SHFLoFDOO\ skill workers (and the positive externalities they
economists Elisa Giannone, David Autor, Peter provide to other workers) while lower-skill, less-
Ganong, and Daniel Shoag have all shown that mobile workers remain behind. Over time, these
what migration has been occurring is now crosscurrents have locked in place a dangerous
sharply segmented by education, with serious VWULDWLRQRIWDOHQWSRROLQJWKDWZLOOEHGLIoFXOWWR
implications for non-college-educated workers’ alter absent national-scale intervention.
employment prospects.36
Additional market problems result from
Consequently, a problematic “sorting” of workers the spread and persistence of economic
has developed, as suggested by Table 1. College- underperformance across large swaths of the
educated workers have been clustering in fast- nation, raising the possibility that whole regions
growing agglomerations with large innovation may be falling into “traps” of underdevelopment,
sectors (leaving behind other metro areas with whereby underperforming regions begin to lose
thinner talent reservoirs), while noncollege the capacity to catch up to frontier regions at all.
Table 1. Innovation hubs draw in highly educated workers, while less-educated
workers face declining employment prospects everywhere else
Share not in
Share of adults with at least a BA, 2017
labor force, 2017
Prime-aged
All Out-of-state Foreign
Innovation sector jobs adults without
residents migrants migrants
a BA
Top 5% of metros 39.0% 56.2% 55.6% 20.4%
Next 5% of metros 34.9% 47.7% 51.5% 20.7%
Next 15% of metros 32.0% 43.8% 44.7% 21.2%
Bottom 75% of metros 26.4% 35.7% 40.1% 23.4%
Note: Migrants relocated sometime in the preceding year. Prime-aged adults are ages 25 to 54.
6RXUFH %URRNLQJVDQG,7,)DQDO\VLVRI,380686$$&6\HDUPLFURGDWD
THE CASE FOR GROWTH CENTERS 27You can also read
