Cities around the world want to become “the next Silicon Valley” and the race to be the next Silicon Something is already a bit of a joke. New York has Silicon Alley, Austin Silicon Hills, Portland Silicon Forest, London Silicon Square, New Zealand’s Silicon Welly, Louisiana’s Silicon Bayou, Israel’s Silicon Wadi, Scotland’s Silicon Glen, and Kenya’s Silicon Savannah testify to the power of the idea.
Why not? After all, we know the key ingredients: round up entrepreneurs and venture capitalists, stir in some startup lawyers, accountants, and angel investors, recruit engineers who want lower cost housing, and build ties with the local university. Boom — growth, innovation, and fancy restaurants. How hard can it actually be? And why do we not have more successful regions?
This matters because rich regions keep getting richer, as do the people living in them. The growing gap between urban professionals with nicely growing incomes and those living in rural areas or in marginal cities is at the heart of the problem of American income stagnation. Can low growth regions learn from high growth ones?
There are two forces that make this difficult. First, these regions experience increasing returns, meaning that winners are going to take a lot and sometimes all. As an increasing return region succeeds, it will begin to generate growth well above the national average due to well-understood agglomeration economics. We know that companies that cluster near each other often grow faster because they exchange information more easily, learn more quickly, share specialized services, grow trust, grow specialized talent and build specialized infrastructure. Proximity matters because cities and regions grow based on the economics of industry clusters. If the underlying technology grows rapidly, these regions grow quickly and develop enormous advantages.
Note that these advantages are not just to individual companies, or they would form their own cities. Amazon may become an exception in Seattle, but for the most part “agglomeration” economics benefit suppliers, competitors, funders, specialized services, talent from related industries, etc. — not just one dominant company. Even Amazon has always benefitted from its proximity to Microsoft.
Some of these advantages are offset by traffic congestion, high labor and housing costs, and other negative externalities. In regions focused on digital technologies (Bay Area, Boston, New York), media (New York and Los Angeles), online commerce (Seattle), finance (New York), and political lobbying (Washington, DC) the positives outweigh the negatives, but it does not always work that way. In general, increasing returns regions are likely to experience high housing prices as homeowners restrict new supply and property values rise. This creates a real barrier to those who want to migrate from less affluent regions. Anyone who has moved for the first time to one of the cities mentioned above understands this.
Historically regions have also become weaker thanks to the “core-periphery” dynamic — the tendency of large firms to outsource activities to lower cost regions. Again what matters is relative magnitudes. A city that outsources most of its work will obviously experience a decline, as happened in assembly manufacturing in much of the midwest. (There is a sound economic case for preventing this that we shall visit another time). An industry that continues to grow rapidly despite outsourcing lower value work (Apple and iPhones) can still grow employment and preserve a dynamic core.
The second problem after increasing returns is that nobody is in charge of regional economic growth. Cities and states can pretend to be, but economic regions are emergent, organic, unplanned, and uncontrolled. They are not engineered, linear, or guided. In their book Competing on the Edge: Strategy as Structured Chaos, Shona Brown and Kathleen Eisenhardt offer a useful metaphor: rebuilding a prairie. We know all of the ingredients of a prairie. We understand precisely the dozens of plant and animal species that comprise the ecosystem that once stretched from the Rockies to the Mississippi. They point out however, that even with perfect knowledge, if you were to acquire land near O’Hare airport, prepare the ground, and introduce the appropriate plants and animals, what you end up with would not be a prairie. Indeed, it would potentially be nothing like a prairie.
Even when we know all of the ingredients, it is very difficult to re-create an ecosystem. Emergent systems are grown, not assembled. They are not grown from scratch — and the actual starting point matters because emergent systems are highly path dependent: past choices shape and constrain future ones. That means that simply introducing seeds and prairie dogs into an acre of land is more likely to result in a patch of weeds than “amber waves of grain”.
Worse, we usually don’t quite know all of the ingredients of most organic systems. Some are highly contextual (meaning your required ingredients and mine may vary) and some are contingent (they only work some of the time, mainly because our understanding of them is imperfect). It matters what sequence you introduce ingredients — much like a souffle that collapses unless the beaten egg whites are added last.
Technology regions and prairies are two examples of complex, emergent systems. There are many others, including companies and markets as well as governments and polities. As they grow in complexity, these organizations increase exponentially the number of components and the number of interactions between their components. They become more complex, organic, and self-organizing — which means you cannot predict how these systems will evolve, much less reproduce this evolution once it happens. They become what Argentine writer and poet Jorge Luis Borges memorably termed “a garden of forking paths”.
Can systems like this be led? They can be guided successfully only by leaders with a deep appreciation of unintended consequences. In any emergent system, the second and third order consequences of any decision are likely to overwhelm the intended first order effects. You can even look at your life as an emergent system, which is why, as Steve Jobs famously noted, you can connect the dots into a coherent post hoc narrative looking backwards, but you cannot “connect the dots going forwards”, i.e., predict anything very meaningful about your life long before it happens.
Douglass North is a wonderful economist who understands the organic and emergent nature of economic systems better than most of his fellow practitioners. He shared the Nobel Prize in Economics in 1993 for documenting the confounding role that institutions, culture, and history play in economic outcomes. Shortly after receiving this award, he was asked whether, since institutions matter so much, he had any advice for Russia.
He thought for a moment and replied: “Get a new history”. That is a good starting point for any city or region looking to start the next Silicon Valley.