Yishan Wong's answer really got me thinking about this question a lot more seriously, so I am posting a slightly broader take building on his answer, and my comment on it (http://www.quora.com/Silicon-Val...).

I agree that most people vastly underestimate the role of Big Government in catalyzing growth of creative capital. But I think there's more going on. In particular, I think it isn't an either/or between the more superficial sounding structuralist formula that the naive believe in (culture, universities...), and a deeper historical chain of inevitable causation.

Why the Question Matters

Anyway, this is one of the subjects I am thinking the hardest about these days, partly because it interests me from a writing/technology history perspective, and partly because I am furiously trying to rationalize staying in Las Vegas (one of the major underdogs in the "imitate Silicon Valley" race) instead of moving to the Bay Area.

I'd like my reluctance to be about more than ridiculously depressing cost of living equations for anyone involved in technology, but too old/tired to put up with rent-and-ramen slumming. If tech is only about poor-and-energetic young gamblers, and older winners who can actually afford to root themselves in SV for the long-term, well to be blunt, 35+ married people like me with a history of wandering around the rest of the country are basically out of the game. So I have a serious stake in good models around these questions.

Even if I end up in the Bay Area tech scene sometime soon, I still need to understand the region differently from its current mainstream inhabitants (today's gamblers+yesterday's winners).

If I stay in the circuit of wannabe-SV cities (I've lived in Austin, Ann Arbor, upstate NY, DC and now Las Vegas, all regions with wannabe-SV ambitions), I need a better theory of these regions that has genuine grounds for optimism.

As a result, I've been having these "tech innovation everywhere" conversations with lots of people who kinda want to have the SV cake without moving there.

First, in support of Yishan's general thesis, a couple more data points:

Data: Big Government Matters

1. The American Industrial Revolution was triggered by military investments in the Connecticut Valley region starting in the 1820s, and centered around the Springfield and Harpers' Ferry armories. America solved the big problem in manufacturing: interchangeability. The effect was comparable to Moore's Law, and there is an excellent book on this history of the impact, Charles' Morris' The Tycoons (which is ostensibly about the Robber Barons, but is really about the broader sweep of technology history between 1820 and around 1920). I wrote a post about this recently, and speculated that an analog to Moore's Law for interchangeability, which I call Hall's Law, drove innovation during that period. http://www.ribbonfarm.com/2012/0...
2. Bangalore is today synonymous with the Indian IT revolution. Guess what? Bangalore started as a sleepy traditional village, which then turned into a major military town during the British Raj. Post-Independence, it became the center for exactly the same kind of high-tech that Silicon Valley between 1950-80: Aerospace (the bulk of the Indian aerospace industry is in Bangalore, including ISRO, the space agency, and HAL, the main aircraft manufacturing corporation). There are a ton of defense labs. I interned at the DRDO (kinda like DARPA) AI and robotics lab there in 1996, while staying at my uncle's house -- he retired from LRDE, the major radar electronics lab. Military tech is everywhere in Bangalore. And oh yeah, the top Indian graduate school, the IISc (which started as the "Tata Institute"... i.e., private sector roots like Stanford, later nationalized) is in Bangalore. I got the internship because my undergrad thesis adviser at IIT Bombay had, as his adviser, an IISc professor, who had strong contacts with the defense research labs. Pune is another Indian city with a very similar story.
3. The German (and later, the Japanese) industrial revolutions were probably the most top-down affairs ever. Both, a century apart, were based not just on strong government support at the foundations, but extremely aggressive foreign policy support abroad, and domestic cartels so that domestic high prices could subsidize export wars.

Data: Big Government is not Enough

Now for a couple of non-data points.

1. I am currently in Las Vegas. The entire city was effectively created by the Manhattan project. Before that, it was a sleepy mining town/railroad hub built around Comstock silver mining. The Hoover Dam started the town on its current course, but it was nuclear weapons testing that made it take off. Later Howard Hughes invested heavily in the city and for a while it was a minor aerospace hub (and related to the LA aerospace boom). But despite these initial conditions, the region never took off as a tech hub. Tony Hsieh is trying once more.
2. I am just starting to read the history of the British Industrial Revolution, and as far as I can tell, this was a counter-example (probably because the "military industrial complex" did not really exist then). It was more of a true bottom-up revolution in steam, iron, steel, textiles etc. There was some impact from the one highly organized "military industrial complex" industry though: naval architecture. But overall, it was bottom-up and trade/guild driven in merchant-dominated regions (similar to early Renaissance developments in Italy)
3. The French, as far as I can tell, tried hard, but mostly failed to achieve top-down catalysis. There were two periods, one during the time of Jean-Baptiste Colbert, minister to Louis XIV, who managed to trigger a brief era of French superiority in naval technology in the 17th century, but was unable to make the temporary leadership turn into a sustained competitive advantage. The second was a brief period around Napoleon (haven't read up on this period much). That also failed. In both cases, French superiority in basic science and mathematics failed to get translated into industrial superiority.

You can do similar analyses for multiple other famous regions around America (Upstate NY, SE Michigan, Ohio valley, Seattle-Portland, North Carolina research triangle...) but I'll stop here.

The Richard Florida Structuralist Formula

Most people seem to believe the formula involves the following necessary conditions (which is sort of a Richard Florida inspired structuralist model, a sort of Porter's Five Forces thing applied to urban geography).

This is the model that Yishan is (correctly) arguing is somewhere between irrelevant and insufficient:

• Major anchor university(ies)
• Great weather
• Great cultural/arts scene
• Diversity
• Major anchor companies to drive both a talent exodus at bureaucratic maturity and to serve as M&A vacuum cleaners and provide counter-cyclic inertia for the business cycle as far as talent markets go

A lot of people sincerely believe in this, especially social-science types who only believe in Truth By Regression. The evidence is apparently compelling. For example, the DC tech scene struggles because it has no major tech university (Johns Hopkins is really more medical), and major anchor companies (AOL came closest). Vegas, lacks a culture scene and has a terrible summer (though the weather is lovely outside of summer).

But there is something deeply unsatisfying about this very regression-analysis type approach to the question. Whenever I read "tech scene catalysis" op-ed's from non-Bay-Area regions (Austin, Boulder, Seattle, Chicago, Detroit, Vegas, LA) that are either trying to start from a blank slate, or trying to revitalize a culture that was once strongly innovative in some other technology, I feel like there is no "there there."

Besides a few ritual remarks about "we shouldn't attempt to copy Silicon Valley but play to our strengths" people stick to Florida-type models. Usually they settle on either a hot topic (biotech in SE Michigan for post-auto growth for instance) du jour or make up some deeply unsatisfying and naive extrapolation from the past (like the DC scene tries to be about government-sector innovation, or LA about media/entertainment, an argument I am myself guilty of making in a sloppy way).

So maybe a deeper historicist approach is the answer?

The Historicist Model

But a purely historicist approach also feels unsatisfying. If everything is contingent and path-dependent and unpredictable, it suggests there is nothing we can do. Your region is the sum total of its past decisions. It also has too much of a "manifest destiny" sense of inevitability. The conclusion "you cannot copy SV, you might as well give up and move here" hangs in the air.

I enjoyed Blank's series, which Yishan summarizes, but I remember feeling, while reading it, that the story had the inevitability of Greek drama. Good for drama, bad for history. That sense kinda ruined the story for me. I am never happy with a historical narrative that does not suggest a larger space of counterfactuals and a thread of more robust, non-path-dependent causal reasoning. I don't like that way of telling technology history stories, to be frank. It borders on hagiography -- a regional version of the self-satisfaction found in CEO autobiographies and commissioned company puff-piece books. There is too much confirmation bias going on.

The historicist model also shows us the importance of history without saying what to do about it.

For example, the French case shows that while Big Government may be necessary, it is not sufficient. In that case, Mahan argues in Sea Power and its Influence on History that France simply lacked a deep common culture around seafaring that existed in Britain and the Netherlands. Colbert's top-down push created a temporary success, but ultimately, nothing took root. It was a nation-scale version of Vegas.

But this tells us exactly nothing about how France might have discovered and built on themes with top-down catalysis that could have succeeded. Today, France has charted a wobbly course to leadership in some areas (like being the hub of the ESA) thanks to the Marshall Plan and a strong nationalist sentiment (read: "Concorde") but I still feel sad for France. It is the great could-have-been in technology history, that achieved only a fraction of its potential, going by its deep scientific culture etc. The best theory I've heard for the failure involves the nature of its institutions, but that's a story for another day.

Synthesis: Model A, Pristine Tech-Hub Formation

So how do you put the two together and go beyond? Uncritical story-telling and regression-happy structuralism both seem deeply inadequate.

Here's a Frankenstein's Monster style synthesis: a 3-phase model of growth for technological cultural capital growth. Let's call this Model A. Pristine Tech Hub Formation.

1. First stage: A native, chaotic culture around some sort of risk-taking activity. Sea-faring for early England, gold mining for California, collapse of eastern-seaboard agriculture due to westward expansion for the Connecticut river valley. This is like a nursery stage.
2. Second stage: the first stage creates the chaotic, bottom-up, resilient, robust, anti-fragile foundation. The second stage: Big Government involvement, sculpts it in broad-strokes ways that require levels of capital investment that only a government can drive (R&D funding driving universities, public utilities etc.). This is like an aggressive fertilization/green revolution type stage.
3. Third stage: Rooting. The training wheels come off, the region either works or doesn't. This is where all the Florida type structuralist factors come in. Capital markets either thrive or don't (rich people like nice weather, even if miners can handle roughneck frontier conditions), smart-but-poor people can either survive on very little, or not. There is either enough law-and-order or there isn't.

I might speculate on a fourth stage: maturity and decline (what happened to many parts of the Eastern tech hubs): the rich people get so rich that they focus on creating beautiful Edens for themselves that then makes it impossible for the poor talent to live in. The vitality diminishes and the economy degenerates to serving the first-world problems of the ultra-rich. You get regions like Hamptons, the richer counties of NY/Connecticut etc., universities that are more about prestige than creativity etc.

I visited IBM's T. J. Watson in Hillary Clinton's Westchester county a few years back and the airport had a NetJets hanger and more private jets taking off and landing in 10 minutes than I've ever seen in my life. I got to the airport early when I was heading out, and spent an hour on the observation deck just watching the damn private jets. Planespotting and musings on wealth competed for my attention. You can actually tell wealth in fine-grained ways based on the planes, if you know your planes.

By this argument, by the way, Silicon Valley is actually under very severe threat. Once the Facebook IPO hits, there may be just too many people with too much money for vitality to survive. It may go the way of the Connecticut Valley.

Model B: Competitive Tech-Hub Formation (Leapfrogging)

In political science, academics distinguish between pristine and competitive state formation. The first state-like entities took root when there were no states. But once a layer of states had formed, future states could only form via competitive creative destruction.

You can apply that idea to tech hubs as well. Look around the world. I guarantee you won't find a new candidate region for a Silicon Valley or Bangalore. The pristine-tech-hub-formation (Model A) opportunities are all gone. Every available patch of great land with great weather already has good universities, military infrastructure, nascent-to-mature capital markets etc.

Which means all the new action will be around competitive tech-hub formation. I don't yet have a theory about this, but political science provides some good indications. Some of the major things to think about are:

Regions perfect for Model A will either decline like upstate NY or turn into enclaves of the rich like a lot of the East Coast. Model B tech hubs cannot really take root there directly (standard center-periphery dynamics). But neighboring regions have a great shot at stardom. So that might mean Santa Cruz or something for the Bay Area.

Unrelated regions trying to pull a Model B will have to deal with less than ideal conditions: there may be no cultural blank slate conditions for Phase A, no big government moneypots for Phase B, and bad weather or other problems for Phase C.

But the good news is that if the Model A regions are facing maturity/decline, then all the Model B candidates are hamstrung in one way or another. The new rules of the game handicap everybody somewhat.

There is also a HUGE disruptor in place. Silicon Valley did NOT enjoy the Internet as one of the advantages of its origin story. It created the modern Internet. Model B candidates can all approach the potential of the Internet in clean-sheet ways, a true open frontier. Ironically, Silicon Valley is the MOST dependent on non-Internet enabling mechanisms like physical concentration of talent, nice weather for rich VCs, etc.

Which means every Model B hub should try and figure out how to use the Internet to beat Silicon Valley instead of trying to join it.

One way to think about this situation is the idea of Leapfrogging. This was first applied in the development economics sphere: Africa being able to leapfrog the development of wired telephony for instance.

We need Tech Hub leapfrogging doctrines and grand narratives. Regions ranging from Boulder to Vegas can use the Internet to leapfrog the development path of Silicon Valley and other Model A regions. But pure structuralism or pure historicism won't get you there. You need MORE than a sense of enabling conditions and a sense of history. You need an imaginative sense of the future, and opportunities available to the rest of the world that were not available to SV, with its locked in cultural investment in pre-Internet innovation conditions.

Anyway, I'll end this loose discursive thinking-out-loud thing here, since this is a subject where I am tempted to go on and on. Sadly, there is little demand for rigorous study of this sort of thing, otherwise I'd be spending a lot more time on this than I do.