Publius has moved!

The conversation has moved. Please join us at

no wal-marts in malawi

In a previous post on Charles Kenny’s book draft, “The Success of Development,” I highlighted the immense quality-of-life gains enjoyed in the past century.  Inventions, such as the polio vaccine and the green revolution, have made it much cheaper to provide quality of life, allowing societies who’ve seen no wage gains in 50 years to live longer, healthier, better educated lives.  The other side of the coin, however, income growth, depends not on inventions but on ‘sticky technologies.’ Kenny makes the argument that national policy, be it collectivist or capitalist, is fairly lousy predictor of income growth, and joins Douglas North in pointing to institutional development – be it a legal system or factory management system – as determinative of economic growth. Also, notice the insight of Paul Romer, recently in the news for revealing his plan to create Hong Kong-like exchange zones  across the world. A future post will highlight Kenny’s recommendations for effectively catalyzing income and quality-of-life improvement:

As Aisbett notes, further supporting an explanation that downplays the effects of government policies on long term growth is the weak reaction of African countries to a range of different policy environments. From state-led dirigisme in the 1960s to structural adjustment programs in the 1990s, average African economic performance was pretty grim. If the right kind of government was really the key to fast growth in Africa, one would expect that Ghana–interventionist in the 1960s and 1970s, liberal-reform oriented in the 1980s and 1990s—would have seen significant growth in at least one of the two periods. In fact, GDP per capita growth averaged –0.4 percent per annum in the first two decades and 0.1 percent in the second two.

This evidence on investment in human and physical capital does suggest that models which emphasize technology over capital may have the upper hand in explaining longterm outcomes. Given evidence of global income stickiness, the technologies that matter must also be ‘sticky‘ in some way. They don‘t flow easily across borders as suggested by the models of Solow, so that some countries have considerably larger stocks of technology than others. What we appear less able to pin down is exactly what types of technology matter where and when –and how we can speed their adoption. This becomes all too clear if we examine the cross-country growth literature covering potential influences on growth other than investment. A list of factors seen to influence economic growth that have been put into modern crosscountry growth regressions would include well over one hundred overlapping economic, policy, structural, sociological, geographical and historical features. Yet a brief survey of surveys appears to suggest that the results of all this computation have been disappointing –even when holding to standards lower than those of Levine and Renelt, who suggested that only investment was significantly linked to growth.
Once again, this is not to say we know nothing about the growth process. First, there are a number of factors that appear fairly robustly associated with growth–if likely as symptoms rather than cause. It is a near-universal phenomenon that the proportion of the population working on a farm declines as countries become richer and that manufacturing and services contribute more to GDP.105 We know that expenditure on food as a percentage of total expenditure drops rapidly as people become wealthier. Furthermore, we have learned that technology is central to per capita GDP growth –given that countries can invest in considerable physical capital (roads, factories, power plants) as well as human capital (school and university graduates) and still not see rising average incomes. We‘ve also learned that some technologies may be more important than others in promoting output and –given the divergence of global incomes– these technologies must be ‘sticky‘.

Paul Romer, the father of technology-driven growth theories, suggests that process technologies are the key to growth. He argues that WalMart‘s management of inventory data has had a bigger impact on economic growth than inventions such as the transistor, for example. Or take the example of Toyota, the auto industry‘s most profitable firm. Toyota does not produce the most innovative or exciting cars (compare the Tercel to the Mustang). But it does have the Toyota Production System –which reorganized factory floors and pioneered just-in-time parts delivery, amongst other things. There is strong evidence to support Romer ‘s contention that process technologies are more important to per capita income growth than ‘traditional‘ invented technologies. Not least, we‘ve seen the absence of a relationship between research and development expenditures in ‘traditional‘ or invented technologies and growth rates. And the evidence from Vietnam, Germany and global studies regarding the impact of war on long term growth outcomes is a sign of the overwhelming relative importance to long-term development of things that can’t be blown up. Process technologies can‘t be blown up, cars and computers can be.

More fundamentally, traditional invented technologies aren‘t sticky –they flow across borders. Transistors (followed by microchips) have spread to every country in the world –and very rapidly. Take transistors in televisions and microchips in computers. Over one half of households in the developing world own a television and there are 219 million computers in low and middle income countries. Per dollar of GDP, developing countries have far more televisions and computers than rich countries. Similarly, you can be stuck in a traffic jam the world over, and some of the worst are in poor countries (Kabul, for example, has hideous traffic). Cars, televisions and computersdon‘t appear to be like the kind of ‘sticky‘ technologies that must be behind income per capita performance.

On the other hand, there are no Wal-Marts in Malawi. And despite many attempts to copy Toyota‘s Production System model, most have ended in failure. Televisions and computers –along with cars, buses and smelters—work pretty similarly worldwide. Inventory control systems and production management systems do not. They are highly context-specific. It takes the same skills to fix a television in New York or Nairobi, it is likely to take considerably different skills to be a good inventory manager. Furthermore, improvements in process technologies have to take account of the existing institutional context. Toyota‘s Production System approach is built around innovation as a constant but incremental process based on small improvements to the existing system, for example. This suggests a long-term and context-specific path of improvement in process technologies that would have the sticky characteristics we are looking for. Institutions such as inventory management techniques, regulatory structures or regime type might be central to the growth story. But the type of institutional innovation which spurs growth in a particular country at a particular time might be highly context dependent. And given the interlocking nature of process technologies, it may be difficult to predict the impact of altering a particular process technology in a particular setting. This was, pretty much, the insight that won Douglas North his Nobel Prize in 1993, for what was termed ‘new institutional economics.‘


Filed under: Economic Policy, General Welfare, World

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: