Does What you Export Matter?

Does what economies export matter for development? And, even if it does, can governments improve on the export basket that the market generates through industrial policy?   These questions were the topics at the recent launching of Does What you Export Matter: In Search of Empirical Guidelines for Industrial Policy a book I co-authored with Daniel Lederman at the World Bank’s INFOSHOP last week. 

A large literature has answered affirmatively to the first question.  Certainly, theory suggests that certain goods may have externalities (benefits to society not captured by their price) that may make it possible for governments to improve on the market outcome.  But to date, measuring such effects at the level of goods has proven extremely difficult and the economics profession has yet to produce a handbook that would tell policy makers which sectors are most desirable.  Instead, we’ve developed some empirical shortcuts or rules of thumb based on characteristics that are thought to be correlated with these externalities. Some schools of thought are best known by their colorful metaphors: natural resources are a “curse”; “high tech” goods promote high levels of human capital and  the “knowledge economy;” a “product space” made up of “trees” (goods) from which “monkeys” (entrepreneurs) can more easily jump to other trees fosters growth, to give just three examples. 

The book made two broad points. 

First, these empirical shortcuts do not, overall, offer reliable guidance for policy makers.   For instance, on balance, there is little evidence for a resource curse -- that on average exporting natural resource intensive products makes countries grow slower;  nor does producing what rich countries produce appear to offer growth benefits once we control for  the country’s investment effort or the diversification of the export basket.   

Second, what we find striking and necessary to understand is the extraordinary heterogeneity in development experiences found among countries exporting similar products.  I’ll give four examples here.
 
• First, on the one hand, Africa and Latin America certainly offer disheartening experiences with mineral exports.  On the other hand, as a previous Bank publicationNatural Resources: Neither Curse nor Destiny documented, Norway has done extremely well with its recent oil wealth, not to mention the historical experiences of the US, Canada, Australia, Finland and Sweden whose development depended critically on resource exports.  
• Second, Korea certainly has managed to generate a knowledge cluster out of exporting computers, measured by patents granted, but Mexico has done far less well over the same 30 year period. 
• Third, recent work looking at the quality of goods (measured by unit values) finds that within very finely disaggregated goods categories (Men’s white cotton shirts, for example, or red wine), there is an extraordinary range of quality.  You can pay two bucks for a bottle of Charles Shaw, or $US14,000 for a bottle of Henri Jayer Richebourg Grand Cru.  Presumably the latter reflects partly a higher level of accumulated expertise and hence quality. 
• Fourth, as Dan shows, about half the differences in industry skill premia at a country level can be attributed to the composition of the export basket, but the other half are due to country specific factors.  This means that, to the degree that higher premia are beneficial because they stimulate private investment in education, a particular country in mining, or agricultural services, sectors which on average show the among the highest skill premia, may not necessarily share the average benefit. Nor is the reverse the case for a particular country in   manufactures of radio, television and communication equipment, which show among the lowest average premia. 

In each of these cases, very similar export baskets can yield very different outcomes and whatever benefits exporting a particular good may potentially confer, it doesn’t happen automatically.  

Hence, we probably need to spend more time thinking about How rather than What goods are produced.  For example, Nokia started as an innovative forestry company and became an innovative cellular technology company. The emphasis is on being innovative, not on the particular product.  To return to the monkey-tree metaphor, it may be the quality of the monkeys, rather than the nature of the forest that matters most. 

Pushing the argument further, in world of globally fragmented value chains, we probably need to focus less on goods, and more on what tasks a country performs along the value chain.  For example, China’s trade statistics document that it exports the i-Pod.  But what it is really exporting are assembly services which, as a former Minister of Finance of Singapore noted at an earlier stage in his country’s development, may require less skill than cutting hair.  By contrast, the tasks that Apple contributes to US i-Pod exports are the more sophisticated parts of the value chain, even though falling under the same general product category.  Whether engaging in these assembly tasks leads China to more sophisticated production processes over time, again, is more a function of the how than the what-some countries have been able to leverage the rents arising from inexpensive labor into sophisticated economies.  Others have not.  But saying that both the US and China are producing i-pods is, at this point, telling us relatively little.

In sum, there is evidence that products differ in their contribution to development along various dimensions, but the profession has given policy makers little solid guidance on how to choose among them.  What is clearer is the vast heterogeneity in development experience of countries with very similar export baskets.  Hence, we advocate “horizontalish” policies that will help get the most out of existing products and lay the foundations for the emergence of new ones.  These may include raising the level of human capital, promoting product quality upgrading, reforming the national innovation system, and improving infrastructure, which benefit a wide range of existing and potential products but which wouldn’t require choosing particular sectors to support.

The commentators  at our launch event added a dimension which the report, in an effort to “give industrial policy (IP) a chance,” deliberately avoided.  Throughout, our analysis assumed  governments, once identifying the right sectors, would engage in optimal policies, thereby  waving away the common argument that state failures may be worse than market failures.  However,  both Carlos Felipe Jaramillo,   Director  for Central America and former Head of  Colombia’s  Trade negotiation team with US, and Jose Guilherme Reis, Lead Economist in the Bank’s trade group and  Secretary for Economic Policy with the Ministry of Finance in Brazil noted  that the costs incurred in the past in pursuing industrial targeting -- never enforced sunset clauses, poorly directed and non transparent subsidies, downstream sectors prejudiced by  protected poor and expensively domestically produced inputs -- should give us pause when contemplating supporting individual sectors, even if we could know which ones should be favored.