12 June 2009

Silicon Valley and Savile Row

dispatches from TJICistan | More on biotech / industrial policy idiocy

Skeptics cite two major problems with the race for biotech. First, the industry is highly concentrated in established epicenters like Boston, San Diego and San Francisco, which offer not just scientific talent but also executives who know how to steer drugs through the arduous approval process.

“Most of these states probably don’t stand much of a chance to develop a viable biotech industry,” said Gary P. Pisano, a Harvard Business School professor and the author of “Science Business: The Promise, the Reality and the Future of Biotech.”


The state of Florida and Palm Beach County used $510 million as bait for a research institute that will employ 545 people

Wow - only $1 million spent per job created. If those jobs pay $100k/year each, and local government gets 10% of that in tax revenue, and the jobs persist for a whole decade, they’ll have destroyed $460 million.

Or, if you factor in the cost of money, something like $480 million.


Cities like Shreveport, where public and private money have built the InterTech Science Park, remain steadfastly optimistic, though a biotechnology manufacturing center at the park was occupied for only six months in 2001 before the tenant went under.

Paul Graham had a good explanation about why tech start ups fail in bad locations, but it's just as good an explanation for why these let's-build-a-high-tech-industry-right-here! industrial policy plans never work.

What are the top cities in the world for fashion design? London, Paris, Milan, New York. Maybe Tokyo. LA? What's the 10th best fashion city in the world? Or the 20th? No one knows. Like Graham says, it's probably so marginal that's it's misleading to even call it a center of fashion. For all practical purposes there's no difference between being the 20th and 100th best city for fashion design. No economic development committee would dream of trying to establish themselves as centers of global fashion because the top spots are already taken and there's no reward to being middling-high on the list.

So what are the top US locations for software design? Silicon Valley, Seattle, Austin, Boston. New York and Denver, but they're pretty far behind. Are Northern Virginia and Southern California worth mentioning? It doesn't make any more sense for Shreveport or Palm Beach to try and oust one of these cities than it does for them to try to oust Milan or Paris as fashion centers. And it also doesn't help to half-ass yourself up to the 25th best high tech business location in the country. And yet bureaucrats toss billions of dollars away trying every year.

Economic development councils try these tricks (always unsuccessfully) all the time. They would never dream of trying it with fashion or finance or publishing, I don't know why they try it with technology, besides that they read some breathless and uninformed article in USA Today and then heard that their neighbor has a nephew who does something they don't understand with computers or test tubes and makes a lot of money doing it. There is nothing magic about biotech or software start-ups that make it any easier to become the new Silicon Valley than to become the new Savile Row.


  1. Is it just me, or is that entire argument by Graham just pulled out of his ass? He doesn't appear to have any actual numbers, it's all strictly his perception.

    Just as a for-instance, my field is CAD software design. The biggest US hubs I'm aware of for it are the greater Detroit area (due to proximity to the auto industry) and Washington (due to proximity to Boeing). I suppose there might be some, but I can't think of any CAD software companies in those top four software locations you name. Yet I know of CAD software companies scattered across the country.

  2. I think that's been the knock on Graham generally - it's all just his observations. I don't think that's a huge problem if you keep it in mind that he's just one dude making observations, and not some oracle like some people make him out to be.

    And yes, I'll admit to simplifying things down to "software companies" (or even "technology companies") when there's much more nuance. Part of that is because Graham (and I, and local industrial planners) are really talking about the presence of software start-ups, not established companies, and part of that is because planners don't really have specific plans. They build big complexes for a software companies, or genetic labs, generally, and hope someone -- anyone -- shows up with some jobs.

    I have some friends in digital effects and their industry seems to be centered about LA, NY, SF and Chicago. I'll admit that every sub-discipline will have it's own centers. But even so, what are the chances that Shreveport could make itself a center of CAD or FX or some other sub-discipline that doesn't already flourish there?