On Supercities, Economic Growth, and Income Inequality in a Post-COVID World

Scott Kupor

The pandemic has been devastating in many ways — from total number of deaths (in the U.S. alone, topping half a million — exceeding the number of deaths across World Wars I and II and Vietnam combined) to devastating economic burdens (exacerbating pre-pandemic record levels of income inequality). Yet amid this grim news, we also can’t underestimate the triumph of modern biology over a novel disease variant — enabling companies such as Moderna to develop a vaccine formulation in two days without physical access to live cultures, faster than even the virus itself traveled around the world.

As companies and cities wrestle with the future of work, future of cities, and future of tech made possible in a post-COVID future, the question is whether it also impacts — and presents an opportunity to address — one of the greatest problems of our time: the unequal distribution of economic opportunity across the United States. Economic inequality and opportunities for job growth are directly correlated with overall spending levels on R&D and with the geographic concentration (or distribution) or those dollars — both public funding (government) and private (as with VC). And the shift to remote work could be an initial gateway to other policy changes that could help distribute opportunity.

Network effects of cities and innovation — and how it ties to inequality

Just as there are network effects in online platforms, where a network grows more valuable to users the more users that join and use it, there are also network effects in the physical world. In fact, cities represent some of the earliest network effects.

Supercities, such as the San Francisco Bay Area, have traditionally attracted lots of newcomers into the network because that’s where the jobs, the labor pool, and the prospects for economic growth lie. This of course creates negative externalities too, such as traffic, burdens on infrastructure and schools, housing shortages, and increases in property values. But network effects tend to persist because the value of the network to its participants outweighs the costs. Just as people may complain about online platforms they use, leaving the platform would represent such a significant loss in other value that they stay, even if there are things they may not like about it. The same applies to cities.

The geographic concentration of network effects, jobs, and economic growth in cities is not a new phenomenon, but disparities between the haves and have-nots continue to expand: Workers in cities earned more than 50% higher wages relative to workers outside those cities in 2016 (it was 30% in 1980). Furthermore, as of 2016, the U.S. coasts housed 9 of the top 10 areas for average earnings (compared to only 3 in 1980).

A similar pattern exists for venture capital funding: In 2020, just three regions — the Bay Area, New York, and Massachusetts — accounted for 70%+ of total venture funding in the U.S. And since VC dollars tend to concentrate around high-tech R&D areas, it’s notable that more than 90% of the nation’s innovation-sector growth (2005-2017) is accounted for by just five metro areas — Boston, San Francisco, San Jose, Seattle, and San Diego.

Government spending on public R&D also correlates with the highest wage-earning metropolitan areas: Seven of the 10 highest-earning areas are in states that also are the top 10 recipients of public R&D spend. These R&D investments matter because they enhance productivity, which enables the formation of new (and often higher-wage) jobs.

Productivity enhancements can also help reduce income inequality. For every 1% increase in productivity, wages for low-skilled workers rise at roughly double the rate of those for high-skilled workers. But, overall government spending on publicly funded R&D has been declining precipitously: We spend less on public R&D today (about 0.7% of GDP) than at any other time in the last 70 years (government spending on R&D peaked at around 2% in 1964). For example, real spending from the National Institutes of Health — which comprises about 25% of all medical research spend in the U.S. — has been down about 10% over the past 15 years.

Why should we care about any of this?

It all comes down to economic growth — and jobs.

From World War II until the early 1970s, U.S. GDP grew roughly 2.5% annually with 2.5% real income annual wage growth. The distribution of income between top and bottom wage earners was also more distributed: The ratio of the top 20% of earners compared with the bottom 20% was 8.5x in 1947 and 7.5x by 1973. Fast forward to the 43-year period from 1973-2016 and we see a very different picture. As noted above, government spending on R&D has fallen and at the same time U.S. GDP has grown annually about 1.7% per year; real wage growth has grown only 0.4% annually. Not surprisingly, economic inequality has gone in the same direction: The ratio of the top 20% of earners compared with the bottom 20% has gone from 7.5x to 13.3x.

What if, post-COVID, we could reverse this trend and enable productivity enhancements to help drive reductions in income inequality?

What if we could help do this by reducing the concentration of labor, capital, and economic growth in supercities and spur economic development in other cities? The pandemic has already shown employers that many jobs can be de-coupled from location. Whether this lasts beyond COVID to a significant extent remains to be seen, but we have an unprecedented opportunity to try now.

Before we talk about how this might work, it’s helpful to take a look back at a case study of what gave rise to Silicon Valley as an innovation cluster and the San Francisco Bay Area as a modern supercity.

Ingredients of innovation clusters (which in turn lead to opportunity)

We’ve long known that innovation clusters thrive with a healthy combination of both bottom-up communities (startups, entrepreneurs) and top-down institutions and policies (universities with research and talent, government R&D, VC funding, favorable legal conditions, and more).

In the case of Silicon Valley, it was a combination of:

  • Universities like Stanford. Even before Fred Terman, then head of the Stanford computer science department, spearheaded the 1951 formation of the Stanford Industrial Park as a means to link academia and commercial activities, Stanford from its beginnings in 1885 saw its mission as both to be a leading academic institution and to facilitate the economic development of the West. Terman encouraged students and professors alike to commercialize technology originally developed in the university research labs, spawning such important institutions as Varian Associates and Hewlett Packard.
  • Technology transfer office. Stanford was one of the first universities to create an Office of Technology Licensing (OTL), in 1970, to facilitate the legal transfer of intellectual property from academic to for-profit commercial activities. And while the OTL has benefited directly in the form of royalties and equity appreciation it has earned over the years through its licensing programs, the university also took a long view that the indirect benefits — in the form of potential future donations to the university from successful entrepreneurs — of a commercially friendly licensing program would outweigh any more extractive patent capture, as has been the case with other universities.
  • Federal government. After World War II, Silicon Valley was a major recipient of government funding to support innovation. Post the 1957 Soviet launch of Sputnik, Eisenhower signed the National Aeronautics and Space Act and the government turned to the one company at the time — Fairchild Semiconductor — capable of building the transistors required for space exploration. Even before this, Terman had positioned Stanford as key to the nation’s Cold War technology research, landing the first government research funding immediately post-World War II. With this funding came significant expansion of competition and end-user markets for semiconductors, ultimately leading to its namesake, “Silicon Valley.”
  • Local/state governments. Among the most important acts of the California legislature and courts was to provide for a strong right-to-work culture in the state, including that employers couldn’t enforce non-compete clauses and non-solicitation agreements, other than in limited circumstances. Non-competes are contracts between employers and employees that prohibit employees from leaving to join competing companies, often for a specific period of time. Non-solicits are the same, but prohibit former employees from recruiting others they may have worked with at a given company to join them at a new company. In fact, the enforceability of non-competes in Massachusetts has been cited as one key contributor to the rise of Silicon Valley compared to Boston.
  • Corporate incentives: Fairchild Semiconductor in 1957 developed a broad-based equity incentive compensation program known as stock options. Simply put, stock options provide employees with the ability to own equity in a company and, assuming all goes well, to benefit from the upside appreciation of their work in helping to grow the company, thus aligning corporate and employee interests. Over time, the concept of broad-based equity ownership has become characteristic of most tech companies and embedded in Silicon Valley startup culture very early on.
  • Venture capital. Silicon Valley didn’t create venture capital, and venture capital didn’t create Silicon Valley. The modern form of the venture capital firm — dating to American Research and Development Corporation and JH Whitney & Co. in the mid-1940s — derived from East Coast institutions. It wasn’t until 1961 that Arthur Rock — after he had already helped the “Traitorous Eight” secure venture funding from Fairchild Semiconductor in 1957 — established Davis and Rock in Silicon Valley, just a few years after Draper, Gaither and Anderson was established out West in 1959. But the development of a permanent, risk-seeking form of innovation capital was central to the technological development of Silicon Valley.

(For those interesting in learning more, The Code is one of the best articulations of the full history of Silicon Valley.)

Some policy recommendations (to help create more opportunity)

How might we apply the lessons of what has worked in creating successful innovation clusters like Silicon Valley to help spur change in a post-COVID era? Again, we already have several tech employers willing to unbundle opportunity from location. So now what can the other players in the ecosystem do to help spur growth — outside of remote tech work and beyond just a few supercities — and into their regions?

State and local government

States can limit non-competes and non-solicits, which would help with job mobility and enabling entrepreneurship in more places. For states with state income taxes, creating state/local level Qualified Small Business Stock (QSBS) rules to incent company development and early-stage investing would help too. And then of course investment in special innovation zones — where regulatory experiments can happen alongside innovation, as with Nevada for autonomous vehicles and South Dakota for crypto — can especially help create opportunity for places where there isn’t already a strong innovation ecosystem.

Realigning K-12 educational systems around the areas of technology specialization that the state/local government is trying to foster is critical. Similarly important are private/public partnerships between local community college systems and private sector employers to create job retraining and certification programs, again tied to the area’s specialization. 


One way that many for-profit companies try to jumpstart a network effect is to start out small — to introduce some constraint into the network versus trying to be a broad horizontal venue trying to please everybody. Universities that are not in supercities could do something similar: Start with an area of specialization in which they can become the best.

Economists call this comparative advantage, but specialization enables universities to get to critical mass and an eventual flywheel faster, one where great professors in an area attract students interested in that area, which then creates enough critical mass of workers to join/build companies in that area. Rinse and repeat and in time you have the beginnings of a network effect. From there, you can build out concentric circles of additional specialization and expand the economic footprint in time.

Private capital

Seed, or very early-stage startup capital, is inherently local. All geographic ecosystems that have been able to attract larger sums of venture capital and growing numbers of entrepreneurs started with vibrant early-stage ecosystems funded by local capital. So, there is a very important role for private capital to play in many of these initiatives, but they need to start at the early stage.

Thus, public/private partnerships that can help foster seed investor ecosystems are critical. That includes education — in concert with the local university and community college systems — and potentially tax-related incentives (e.g., local/state QSBS) to encourage high-risk, early-stage investing. Later-stage capital of course ultimately matters, but that capital can be “borrowed” from players in other states at the right time (and even countries; capital is global after all). Simply put, later-stage capital travels easier; seed-stage capital providers are homebodies.

Federal government

The U.S. government has long played a role in innovation — going back to Vannevar Bush and the development of the National Defense Research Committee in 1940. We did the same post-Sputnik with the 1958 founding of the Advanced Research Projects Agency (now known as DARPA). And the government sponsorship of the Human Genome Project starting in 1990 arguably led to the many developments we are now starting to reap in terms of modern drug development. A more recent incarnation of this is the 2007 development of ARPA-E, designed to fund basic science related to advanced energy technologies.

The federal government has also played the role of “kingmaker” in where important grants and contracts go. In many cases, this has led to creating centers of excellence (e.g., NASA in Houston; SLAC and Lawrence Livermore labs in California).

But instead of just increasing basic R&D funding overall, the federal government can also be investor-like, directly benefitting from the returns generated on this capital. If we thought of these dollars as part of a federal “Office of Technology Licensing” akin to what major research universities have built, the government could create a potential endowment pool of research dollars — funded in part by royalty returns on technologies generated by its funding (and even potentially by small equity grants in companies licensing the technology too).

On the tax policy side, federal government could include preferential tax treatment for private investments in areas deemed to be developing important technologies. This could be viewed as an extension of the QSBS rules that today permit many entrepreneurs and some early investors to avoid paying federal capital gains taxes when they sell appreciated equity.

Countries such as Canada have also adopted aggressive payroll tax holidays for engineering roles, effectively reducing the cost of employing an engineer by about 50%. The federal government could do the same here, but tied to specific, otherwise underserved geographic areas, to stimulate economic development in those regions. And of course, more liberal immigration policies are always a way for governments to help jumpstart targeted investment into special areas.

Finally, as technology transforms the workplace, the government could create more tailored student loan programs by tying loans, and even loan forgiveness, to certain educational/job categories and geographic areas, much as they did for programs like Teach for America and others, only in this case targeted to specific regions and domains.

To be clear: When I talk about the government targeting domains or reinvesting gains, I am not talking about the government picking winners and losers among established companies, or about government-owned enterprises. The free-market system is still best at determining which technologies to develop and commercialize.

But the government, as a funder of foundational technologies, could be a very small beneficiary of the economic returns generated and turn it into something more generative. Investing in basic R&D is also very different from the government funding applied research: The latter is for the primary market to determine winners and losers, whereas in the former case the government is filling a role otherwise abandoned by the free market. 

Starting new companies, creating new jobs, reducing income inequality, engendering economic development — these all are hard things and ripe for failure. Startup founders and venture capitalists know this all too well; most efforts fail and some achieve good (not great) outcomes. If we are going to try to tackle really hard problems with the help of government, we need to adopt measures of success that orient government “investments” like a portfolio. Then the measure of success would be the total economic value generated by the portfolio, not the individual wins and losses of each individual investment.

Some have interpreted this with respect to Silicon Valley as a celebration of failure, but nothing could be further from the truth: Nobody likes failure and certainly nobody sets out to fail. But all hard problems — whether starting a company or trying to get back to 2.5% annual GDP growth — may end in failure. If, however, a portfolio of hard problems can ultimately generate uncapped economic upside from one or two successes, most people would vote with their wallets in favor of experimentation.

A number of people have made policy recommendations like these before. We know what works, just as we know what doesn’t work. But there’s an unprecedented timing opportunity now. If there’s one thing any of us who have studied Silicon Valley and the history of innovation know, it’s that timing matters: some ideas are too early. But when platform, market, or other inflection points present themselves, it’s a good signal that something has changed or is about to change in the market that enables the adoption of a new technology.

We are now at an inflection point with COVID and the opportunity to address income inequality and stagnant economic growth. Many employers entered the pandemic never believing remote work would happen, and many are already issuing policies that have shifted the new normal. Technology enabled us to make this work, and while it won’t solve everything, it certainly helped accelerate the new normal.

If we don’t get it right, our competitors — in the form of foreign governments that also recognize that economic growth, national security, and international competitiveness are converging — certainly will. As countries such as China develop centralized mechanisms to invest in core enabling technologies (e.g., artificial intelligence), more than ever the future economic competitiveness of the U.S. is inextricably linked with our own national security and sovereignty. Creating jobs and reducing economic inequality by investing in technological development are the keys to our long-term success.

It will not be easy. It will not happen overnight. We will fail many times along the way. But the stakes are high enough and the uncapped upside big enough that now is the time to try.