Thursday, June 26, 2014

Single-industry clusters and deregulation



At Bloomberg, I express some pessimism about Marc Andreessen's idea of creating single-industry tech clusters via local deregulation:
Why does Silicon Valley exist? Why do software engineers go there to get a job, even though rents are high? Why do entrepreneurs go there to start companies, even though wages there are high? Why do so many big technologies and products come out of that one tiny region of the world? 
These are the kind of questions that urban economists ask when they study so-called industrial clusters. But unfortunately, the quest for the secret of how to create the next Silicon Valley has so far met with frustration. We still don’t really know whether companies go where the talent goes, or talent goes where the companies are, or both. We don’t know if government can bring clusters to life by building infrastructure, or by luring big firms to the area, or by encouraging firms to form relationships, or by creating a cool youth culture...
Read the whole thing here!

16 comments:

  1. Sigh...

    Noah,
    This is textbook urban and regional econ stuff. Go read Krugman's Geography and Trade (I have been critical of him on some of these issues, but in this book he is excellent). Better yet, go read Alfred Marshall. Most of it is in there (I quote a bunch of this stuff in my 2011 book, which I shall not advert further by mentioning its name; it is obscure anyway, although right on this, among other things). The fact is that there are a variety of external localization economies that tend to lead to the development of local industrial clusters, which in effect reinforce local comparative advantages, which dynamically evolve with such things as local labor force compositions and skills. This is also Brian Arthur Santa Fe stuff, been around for awhile.

    Of course, as Krugman notes, it is hard to pin down how a particular area gets to have a particular cluster. There is a lot of path dependence here given these externalities and increasing returns. Locality that gets ahead with a particular industry can stay ahead with that industry, although it is not impossible for other locales to overtake it/them, just hard.

    I think you are right that proposing single industry deregulation is silly. What is to keep the other locales from meeting this with their deregulation. This just opens up a "race to the bottom" of deregulation. I do not know this guy's work or views at all, but this is simply consistent with a general libertarianism or pro-laissez faire view. If what you really want is general deregulation, well, calling for local amounts of it may lead to it happening generally. But if there is some good reason for having the regulations, then deregulating locally in some possibly vain effort to steal an industry away from some other locale where it has gotten entrenched without deregulating strikes me as not necessarily such a great strategy.

    Of course, the same sorts of issues show up for all other sorts of local subsidization schemes, whether through tax cuts or direct subsidies. This is the old enterprise zone approach, once popular back in the days of Jack Kemp, but now largely passe, with some such as Kemp seeing it as a foot in the door to general tax cuts and deregulation (those enterprise zones also emphasized deregulation of such things as local zoning rules and so on).

    Barkley Rosser

    ReplyDelete
    Replies
    1. This is textbook urban and regional econ stuff. Go read Krugman's Geography and Trade (I have been critical of him on some of these issues, but in this book he is excellent).

      I have read all of that, and really like it. It motivated my thinking in this piece, and I was going to mention it, but cut it for length. The fact that clusters depend on entire patterns of trade is an argument against the power of governments and firms to create new clusters by will alone.

      I think you are right that proposing single industry deregulation is silly. What is to keep the other locales from meeting this with their deregulation.

      I pointed this out in my article.

      But if there is some good reason for having the regulations, then deregulating locally in some possibly vain effort to steal an industry away from some other locale where it has gotten entrenched without deregulating strikes me as not necessarily such a great strategy.

      I pointed this out in my article.

      Delete
    2. Gosh, Noah, you are so darned reasonable, I guess you deserve a gold star, :-).

      JBR

      Delete
  2. Anonymous3:12 PM

    Many countries have special economic zones. They can work quite well if structured appropriately, and can lead the way on policy as "laboratories of productivity".

    http://elibrary.worldbank.org/doi/pdf/10.1596/1813-9450-5583

    It could be that the US already effectively has quasi-SEZs due to a more decentralized political system: local governments can do a lot to make self-organizing business clusters happier. If the problem is opening up new clusters, then perhaps the solution is to give less power to low-information bureaucrats who live 2,500 miles away.

    The FAA just now put the commercial drone industry in the US back a few years. For what gain? A nation-wide preemptive regulatory regime spells stagnation. Ask Admiral Zheng He.

    Let a million flowers bloom, then solve and regulate problems locally. Let national regulation wait for national problems. You'll get your clusters.

    ReplyDelete
    Replies
    1. Most of those SEZs are in nations with really high amounts of state intervention in their economies, such as China. Not really relevant to US.

      Delete
    2. Anonymous5:10 PM

      I'd suggest that you are underappreciating the amount of federal, state, and local intervention in the US and its drag on productivity. Bluntly, your views seem a bit outdated.

      But if US SEZs could boost productivity, create clusters, and help test new regulatory costs and benefits, then why not? Detroit may be a good start.

      http://www4.ncsu.edu/~jjseater/regulationandgrowth.pdf

      Delete
    3. Sorry, Anonymous, but this "been there and done that," with very mixed results. So, US and UK have been doing these since the 80s. In US NJ, CA, IN, MD, MA, KY have been doing them. Some have claimed to be successful, perhaps most vigorously, one in Evansville, IN, but the general story is a mixed bag. You have to keep in mind that development in one location is often at the expense of it somewhere else, with little to no net increase in jobs or growth, but a loss of tax revenue. For US experience, see Alan H. Peters and Peter S. Fisher, _State Enterprise Zones: Have They Worked?_, 2002, Kalamazoo: Upjohn Institute.

      For UK, they have had even more extensive such zones since the 80s. Peter Hall has written much on this. A book by him in 2002 found that for a loss of 300 million pounds there appeared to be a net increase in jobs of 13,000. Is that a good deal?

      JBR

      Delete
  3. Anonymous4:45 PM

    Interesting you didn't mention the most famous example of all - the financial industry and the amount of new products it engendered (and of course the subsequent costs).

    ReplyDelete
  4. Are you really sure these tech clusters are so attractive? There used to be a substantial aerospace industry in the Los Angeles area, with firms like North American Rockwell, McDonnell Douglas, Hughes, Lockheed Aircraft, and a host of smaller outfits, busy building bombers and fighters and spacecraft and satellites, I'd guesstimate that about 3/4 of the employment of those firms and demand for their products has evaporated in the last 20 years. And yet the local folk have borne this loss quite placidly, and not an economist in the nation has ever voiced the slightest regret.

    ReplyDelete
    Replies
    1. Mike,
      This is true for specialization of any sort based on comparative advantage, the gains from specializing and trading according to which is one of the few ideas out there that the vast majority of economists agree with, although there are some dissenters. In any case, the hard fact is that whatever industry a nation or region or local area specialize in, it can go down the tubes either because there is a fall in demand for it, or if it is a depletable natural resource, it can get used up. So, yes, that happened to LA with defense industry, but the logic of specialization remains, although it does not hurt to have some backups. In some cases, having good human capital can help a transition, such as in Pittsburgh, which lost its steel industry, but has expanded a specialty metals industry that its skilled metal-working labor force has been able to get at least some jobs in, with support for this coming partly from Carnegie-Mellon University and some research and tech development spinning off of it that aided this. But there is no guarantee of such transistions happening.

      I might note that Alfred Marshall's interest in this came late in his career, driven especially by the decline of UK's industrial clusters in the early 20th century due to competition from clusters in Germany and the US.

      Barkley Rosser

      Delete
  5. Anonymous8:20 AM

    Noah writes,

    "Detroit is the best example we have of a single-industry cluster, and I don’t think many cities want to go down that path."

    Yes, we all know that Detroit is an easy target but perhaps you should reconsider the cheap shot. Using a photo of the infamous Packard Plant that closed in 1958 is a nice, albeit predictable, touch.

    Why is it that every time an automaker announces that it is considering building a new factory in the U.S., state governments, particularly in the Southeast, fall all over themselves to throw big money incentives at the target? It would seem that there are plenty of cities in the U.S. that "want to go down that path". Among them recently: Chattanooga, Montgomery, San Antonio, and, yes, even sort-of-Siliconny Fremont, CA is offering big bucks to Tesla just this month.

    The automotive industrial cluster continues to survive and thrive in "Detroit" it just doesn't and hasn't ever existed in the urban core of Detroit. A Wolverine surely knows what drives the considerable wealth in Washtenaw County and places such as Novi, Plymouth, and West Bloomfield.

    ReplyDelete
  6. I don't get it. Why are geographically localised clusters of particular industries considered to be a good thing? Why are we focused on creating such clusters? Might not silicon valley be not viewed as a glittering gem of technological innovation, but rather the product of adverse laws suppressing technological development in the rest of the country? Would it not be more efficient to have tech industries be distributed more evenly across the country, so that technologically minded people would be able to find tech jobs without having to travel halfway around the country?

    ReplyDelete
  7. somalia is completely deregulated. I suggest mr andreesen opens a new startup there.

    or maybe he needs highly educated and skilled workers, that (generally) only come from stable countries with well organised property rights that can bring about good education systems that produce skilled workers. nice weather helps.

    ReplyDelete
  8. Andreessen is either ignorant or cynically greedy. Silicon Valley and other technology clusters are the result of massive government intervention in the economy. Silicon Valley was built on billions in speculative government spending through such agencies as ARI, DARPA, ORD, and a host of others. Andreessen's should know this. He developed the MOSAIC browser at the government funded NCSA and then commercialized it. Talk about scorning the lesser rungs by which one ascends. NASA Ames alone dispensed bales of money to get Silicon Valley going, so Andreessen's arguments are at best ingenuous.
    Clusters like this also require massive regulation and high taxation. Strong anti-trust regulation back when Silicon Valley was founded led to speculative corporate funding on research at places like Bell Labs, Xerox PARC, and IBM Watson. High taxes meant that companies spent more on R&D rather than on executive perks and salaries, so even middling companies had real R&D departments that would now and then take a flier and try something actually innovative.
    This high taxation, high spending, high intervention, high regulation regime has repeatedly developed technology clusters. Look at medieval Germany developing its mining and metal processing which still provides that nation with a lead in industrial and chemical products. Look at Dayton, Ohio, home of army procurement which led to the late 19th and early 20th century industrial revolution and such products as the airplane.
    Technology and innovation clusters basically develop where the government wills them to.

    ReplyDelete
    Replies
    1. The first US technology cluster was in the Connecticut River valleys. Eli Whitney scammed the Continental Congress, even before there was a US, to fund a bunch of rifles made with interchangeable parts. It took until the 1830s before his son and Colt succeeded, so you can imagine the whining we'd hear about this kind of thing these days. Still, the US came out ahead.
      Another technology cluster involved the British Navy which first pulled ahead under Queen Elizabeth I. Not only did England get a good return from its navy in the form of a colonial empire, but it also got all sorts of high tech stuff like soda crackers and lime juice to fight scurvy.

      Delete