Christina Romer and manufacturing

In the his state of the union speech, President Obama stressed the importance of rebuilding a manufacturing economy. This idea is not popular among “liberal” economists. During the worst of the panic in 2009, Robert Reich  piped up attack the auto rescue  because we should all work in the service economy  ( exporting power point slide decks, I guess). Another “liberal” economist, Christina Romer who used to work for the President outlined her objections  to manufacturing policy in the New York Times recently - and her article shows the gap between reality and orthodox economics in liberal as well as conservative versions.

Romer  begins with an assertion that building “clusters” of related industries to produce a positive environment for new industry does not make sense. Advocates of industrial policy argue that “clusters” of related industries like Silicon Valley or Shenzen or Stuttgart produce a manufacturing ecosystem in which it is more likely for manufacturing enterprises to prosper. So if we want an advanced electric car industry, we need a cluster of companies that make things like advanced lithium batteries and the components that go into advanced lithium batteries. The idea is that once there is an operating ecosystem for batteries, private investment can take the lead role, but that establishing a new industry, especially in an area where the US does not have much of a presence,  the government must assist in order to get the industry going. Most people, especially people who have worked in industry find this to be plain obvious - it is, after all, how we got Silicon Valley. In fact, one of the founders of Intel points out that the disappearance of consumer electronics manufacturing in the USA meant that nobody was making and doing research and development for advanced batteries so we are at a huge disadvantage on electric cars - clusters of industry are not only key for a particular industry but for related and emerging industries.  None of this is visible in orthodox economics theory. Romer writes:

But large clustering effects have been hard to find. A study by Professors Glenn Ellison of M.I.T. and Edward Glaeser of Harvard showed that in many industries, businesses were only modestly more clustered than if they were allocated randomly — suggesting that the benefits, while real, may often be small.

Romer is incorrect about what that study shows. Ellison and Glaeser actually wrote that they were “reaffirming that geographic concentration is ubiquitous and there are many highly concentrated industries.”  What they said was that in some industries there don’t seem to be clusters. One of the examples they cited was television manufacturing which in the 20 years since the publication of their study has disappeared from the USA entirely. It turned out that the dispersed nature of the industry in 1992 was not proof that industrial clustering is unimportant, it was a symptom of the ill health of the industry and its reduction to a few isolated factories that lacked the ecosystem needed to survive. If you want to run a TV factory, being near similar companies means that you’ll be able to find someone to repair and maintain your manufacturing equipment, that there is someone around to sell you the parts you need (this is called a “supply chain”), that there are companies around that know how to distribute your products to markets, that bankers in the area understand your business and its capital requirements, that you can find civil engineers who understand how to build your plant, that there are skilled workers available for hire, that there may even be used equipment nearby which can reduce your startup costs and so on. How could Romer so totally miss the implications of the study she cites? Because in orthodox economic theory, it is “known” that manufacturing policy does not work and she does not question this starting assumption.

image

(from Business Week)

Here’s Romer with another counter-factual defense of theory:

A related argument for subsidizing manufacturing involves learning by doing. It takes time for a production process to become efficient. But whether learning creates a role for government depends on whether the eventual returns are captured by the company taking the risk. If the company that jumps in first and eventually succeeds reaps all the rewards, there’s not a market failure. The company needs to count the learning period as part of the investment cost. And with well-functioning capital markets, it should be able to find investors without government help.[bold added]

Consider the last sentence of this paragraph, written 4 years after world “capital markets” imploded and while Europe is still teetering on the edge of collapse because European “capital markets” allocated funds to speculation on obviously unsustainable debt instead of investing in long-term research and development. Despite orthodox economic theory, our capital markets do a very poor job of allocating investment - especially for projects that involve long-term research and development and/or markets that have not yet been proven. Capital markets in the United States in the first decade of this century allocated a lot of money to building shopping malls, to real-estate speculation, and to Bain Capital type private equity looting of profitable, functioning companies. A “well functioning capital market” that is willing to bet on long term returns is exactly what we do not have. And in US history, to government has often played a key role in developing industries. Railroads, aerospace, integrated circuits and the internet on which I read Dr. Romer’s words were all nurtured through infancy by public support. That’s because actual innovation does not work like in economics textbooks: anyone with experience in corporate R&D will tell you that there are many opportunities companies pass up because they know that they will not be able to recover costs even if the product is a success because others will enter the market.  Dr. Romer is operating with a simplified model of corporate behavior that economists like because it fits their schemas, but that does not correspond to real companies.

Notes

  1. cathrin-armstrong reblogged this from krebscycle
  2. krebscycle posted this