A cows plus strategy, and clustering

Some thoughts prompted by ‘The New Geography of Jobs‘, by Enrico Moretti

This book is a great description of the economic geography of the modern United States and an explanation of how it got that way. It is also a primer for policy-makers interested in how to ensure more good “innovation sector” jobs. Mr Moretti talks about how the “brain hubs”, like San Francisco, Seattle or Boston, came about, and what to do if you find yourself (like New Zealand and most United States cities), wanting to transform your local economy over time while avoiding the cul-de-sacs and mistakes that others have fallen into along the way.


Mr Moretti divides cities into places with good jobs, those without, and the places that could go either way. “Good” jobs are involved in the innovation sector, which is loosely defined as high-tech plus any other occupation that makes intensive use of human capital and ingenuity. Places without good jobs are mostly declining or declined manufacturing centres in the Rust Belt where the book starts.

Wider economic benefits from innovation sector jobs are very big. Mr Moretti reports that for every innovation sector job, e.g., a new software engineer at the Googleplex, there are an additional five jobs created, two professional jobs (e.g., doctors, lawyers, teachers, nurses) and three non-professional jobs (e.g., waiters, carpenters, taxi drivers, shop workers). Most sectors in the economy have these multiplier effects but those in the innovation sector are particularly big. So while employment in the innovation sector will never account for the majority of jobs (at the moment it accounts for around 10%), it has a disproportionate and positive impact on the economy overall. Although Boeing employs twice as many people in Seattle as Microsoft, it ultimately creates fewer local jobs.

There is not just a jobs effect but also a big income effect. Everybody earns more in a brain hub than s/he does doing the same job in a Rust Belt city. Your neighbour’s skill level affects how much you earn, regardless of your own skill level. This result is not necessarily intuitive, but Mr Moretti (on page 99) says there are three reasons for it. First, skilled and unskilled workers complement each other: an increase in the former raises the productivity of the latter. Second, a better educated labour force encourages the deployment of new and better technologies, raising productivity. And third, the economic magic of human capital externalities, i.e., when people interact they learn from each other, and those who interact with better educated peers ultimately become more productive and creative.

The result holds for employees of all skill levels. So a college graduate in Boston, where 47% of residents have a college degree earns $20k a year more on average than a college graduate in Yuma Arizona, where 11% of residents have a college degree. But the biggest impact is for the least skilled. A high-school graduate in Boston earns 34k more a year than a high-school graduate in Yuma. Part of this difference, says Mr Moretti, is compensating for the higher cost of living in Boston. But using a nifty longitudinal dataset that follows the same individuals over time, he finds that the same individual can make a very different salary depending on how many skilled workers s/he has around.

Learning from history

The early part of the book charts the decline of American manufacturing jobs and the rise of jobs in innovation economy. It explains the hollowing out of the American labour market, as new technologies favour high-skilled workers, reduce the need for many occupations that call for medium-level skills, and have little effect on occupations at the low end of the skill spectrum: jobs that involve non-routine tasks have not been particularly hurt by computers.

Since every city, of course, wants to be a brain hub, the book also explains how brain hubs developed, by way of sketches of the development of Seattle, Silicon Valley and Hollywood. This part is particularly interesting because it goes back far enough in time to show what Seattle was like in 1979 when Bill Gates and Paul Allen moved their then fledging business there. They did this not for professional reasons but for personal ones. Indeed Seattle seemed like a terrible place at the time: The Economist had recently labelled it the “city of despair”. And especially compared with Abuquerque, New Mexico where Microsoft was founded and which already had the beginnings of what might have become a tech hub. Nevertheless, they moved to Seattle and the rest, together with the decisions of many others to set up shop there over time, is history. Jeff Bezos started his new firm, Amazon, in Seattle in 1994 because by then Seattle had the engineers, programmers and venture capitalists that he needed to get started.

The new picture of economic geography

I saw this picture recently on Twitter (created by Reddit user atrubetskoy). The blue areas are responsible for 50% of US GDP. So are the orange areas. Twenty three successful cities.

Mr Moretti talks about why this clustering happens, and why is it that new software firms, which might be just two guys in their garage, decide to locate in San Francisco or Silicon Valley or Seattle despite the fact that these are amongst the most expensive places to live in the United States. The contrast with industrial economy firms is stark: they go where resources plus transport are cheapest, so paper mills are near forests, milk factories are near farms, and steel mills are near coal mines. The lobster industry is in Maine, because that is where the lobsters live, and the oil industry is in Texas, because that is where the oil is.

There are three factors that create clusters First, firms go where they think they will find good employees, which happens to be where all the other employers are. Second, they also want to find good suppliers, and customers for their products,hence the ecosystem. The venture capitalists of Silicon Valley are far more helpful to a growing firm looking for funding than the venture capitalists of other places with less developed ecosystems. And third, it turns out that innovation, collaboration and the creation of new ideas are highly influenced by proximity. Even small physical distances are enough to deter collaboration and information sharing. This is seen from networks of patent citations: inventors are significantly more likely to cite other inventors living nearby than inventors living far away. Similarly, I recall a study from Google some years back of information flows in the workplace that said that what you know is determined largely by where you sit (sorry I can’t find the reference just now). Mr Moretti says (p141), “being around smart people tends to make us smarter, more creative and ultimately more productive”. These cluster effects are intensifying because they are self-reinforcing, and because, once a cluster is established, the costs of moving it are overwhelming.

The policy of inevitability

The question for policy-makers, especially local economic development agencies, is therefore how to turn their city into a brain hub.

Mr Moretti explores a few options. Basically it is hard. The best bet is to have already got lucky and have had some innovation sector businesses set up shop. If that has not happened, then you could look at supply side ideas or at demand-side ideas to get things started (this is from the perspective of the labour market, so “supply side” means increases the number of workers, and “demand side” means increasing the supply of jobs).

On the supply side, the basic strategy is to attract highly skilled, young, creative types, and then expect that high-tech employers will follow, hoping to take advantage of the ready local labour supply. Mr Moretti is down on this approach, arguing that Berlin is the best example of this strategy but that Berlin continues to have amongst Germany’s worst unemployment problem and weak economic growth. That said, Berlin is widely considered to be one of the most interesting and vibrant places in Europe at present. Conclusion: there may still upsides for residents in adopting the supply side strategy, but it is far from a slam dunk case from an economic point of view.

In some industries, and Mr Moretti says biotechnology might be one, attracting stars is also important to encourage both other workers and to encourage employers. Qatar does this. So does Singapore. And probably other places with which I am even less familiar. New Zealand’s efforts to attract over-achievers from other countries to come and live here might also be an example, so when James Cameron or Julian Robertson come live here, they can help attract other people and establish an ecosystem. Mr Moretti talks about the cautionary tale of the University of Washington, whose efforts to build a classy economics department foundered when resources ran scarce before the star attraction programme was completed.

The demand side idea is basically about attracting high-tech employers to come set up shop. The main tool is subsidies or tax breaks as part of so-called “place-based policies”. Mr Moretti says around $60 billion a year is spent on them in the United States. The idea is to provide subsidies for the first firms to come along, and then stop the subsidies after enough firms have arrived that economic development is self-sustaining: the “big push” strategy aims to break the impasse that keeps high-tech firms from locating outside of existing high-tech areas.

Mr Moretti says the push needs to be really big, decisive and sustained, and it has to target the right people. And the big problem is this last point: local government needs to be in the business of picking some winners.

The track record of these types of policies is mixed, to put it nicely. Many hubs have nothing to do with government action: Silicon Valley, the biotech cluster in San Diego, and the movie cluster in Hollywood are examples. Even in smaller places, Portland Oregon being one that I am familiar with and that seems actively to encourage the inward migration of young, skilled people, the heart of the high-tech hub is a private employer: Intel’s semi-conductor facility in 1976 was the start of things there. More locally, it is not obvious that government had much to do with Weta, nor with Xero, both of which have substantial local ecosystems around them.

That said, internationally Israel and Ireland would be held out as examples of success, the former more so than the latter these days, and even there the intention of government spending was not to create a local tech sector but instead to develop innovative defense technologies. Taiwan is another relevant example, transforming from a rural economy to an advanced one courtesy of government-sponsored research in the 1960s and 70s. Policy-makers bet on several failed technologies, but also on semi-conductors, which turned out extremely well. The history of efforts by governments is not studded with success, however. Mr Moretti reviews the experience of Fremont California, where Solyndra, a major employer, maker of solar panels, and recipient of significant federal government support, went broke in 2011. It seems that the industry of making solar panels does not exhibit strong forces of agglomeration, although it is hard to find this sort of thing out without trying.

Smaller scale efforts to attract employers, e.g., Twitter’s recent move to central San Francisco, are very common and, in some cases, seem to pay off for the communities involved, i.e., the benefits from spillovers can be bigger than the cost in tax foregone. That does not make the residents happy necessarily, since they appear to be giving city tax breaks to enormous profitable companies.

A programme that successfully encourages development, that does not try to pick winners too directly, that targets incentives carefully, and that incentivises private investment is more likely to be a winner, says Mr Moretti.

So what to do

Sadly Mr Moretti does not offer too much encouragement. There are no straightforward answers it seems, and you will only know you have succeeded once you have succeeded. His policy agenda is fairly broad, mentioning vouchers to encourage people to move to areas where they are more likely to find a job, a boost to RnD, a major improvement in the quality and quantity of education especially in technical subjects, and a loosening of immigration policy, since immigrants seem to be a lot more inventive than locals in the United States.

What all this means for New Zealand, a place that is not a brain hub in many industries and is far from the world, is not especially clear. The picture is even less clear for regions within New Zealand, which are still further from brain hub status by comparison with the major centres of New Zealand.

But let’s say New Zealand wants to be a brain hub, i.e., we want to adopt some policies that will attract or create high-tech companies that will hire high-tech workers, and this will generate jobs and boost everyone’s income. Assuming we have not got lucky, i.e., we are not a brain hub already, and we have good education, immigration and RnD policies already, some possibilities include the following:

* Understand the situation, i.e., figure out what New Zealand is good at and not good at, what is useful and can be built upon and what is difficult and will need to be worked around. Thinking about how the whole hangs together (“the city of four million people” to quote Shaun Hendy) and how we can compete with much larger Australian cities that attract more people is useful.
* Back some local employers or employment initiatives with public money, and over time expand the ones that seem to be working (i.e., the example of Taiwan, but not of Solyndra). Bear in mind that tech venture capital firms expect only one of every ten investments to succeed, a few to muddle on, and the rest to sink without trace. Politically you are going to have to be resilient to failures with public money. Not easy.
* Connect with educated locals and encourage them not to leave. The experience of Otorohanga might be inspirational as well as educational at the level of the nation.
* Connect with your diaspora, and try to encourage them either to come back with their businesses and networks, or to take an active role in supporting local initiatives from wherever they are in the world.
* At a national level, I think it would really helpful if New Zealanders went to a more diverse set of places (and not just focusing on the UK and Australia). We are not sufficiently well connected to China and the coasts of the USA.
* Make some localised improvements to amenities. Perhaps just giving talented people a place to run into each other could be a useful step forward. ATEED is building an innovation hub to house high-tech firms: a cluster of them is clearly already developing on Viaduct Harbour Avenue, with Vodafone, HP, Microsoft and others already in residence, just to name the brands you can see on the door. Fonterra is moving in next door.

You could also get some useful ideas on talented people and how to attract, retain, develop and connect them from the ever-interesting McGuinness Institute, and their project TalentNZ.

Do not expect speedy miracles. Sustainable economic development takes a long time. You can see New Zealand’s exports and imports broken down by type below, courtesy of the MIT observatory of economic complexity. First is 1990. Second is 2010. See how many colours have changed?

NZ Exports 1990


NZ Exports 2010