11/01/2012 10:27 am ET Updated Jan 23, 2014

Why Economic Growth Happens in Some Places and Not in Others?

Have you ever wondered why, sometimes, poverty continues to increase in countries that grow super-fast? Take the case of India, a model of economic take-off. The number of Indians living on less than $1.25 a day (yes, $1.25 a day, that's how poor the extremely poor are) went up in the 30 years through 2005 -- from 420 million to over 450 million. Sure, the poverty rate -- that is, the proportion of poor people in the total population -- may have gone down. It did in India, by about a third. But falling rates don't mean much to you if you happen to be among those who don't have enough to eat. What explains these big masses of poverty inside otherwise successful economies? A new book edited by Ejaz Ghani of the World Bank looks at South Asia -- the part of the world with the largest concentration of poor people -- and concludes that the main culprit is, believe it or not, geography.

Economic activity tends to concentrate -- to "agglomerate," in technical speak. Producers want to be close to consumers. This allows them to cut transport cost. It also helps them acquire information, especially on tastes and technologies. More important, proximity allows them to take advantage of "economies of scale" -- you can produce cheaper TV sets if you make a million of them instead of a thousand. Why? Because the cost of your initial investment can be divided among a larger number of customers. [NB: The initial investment could be money you spent on a plant or on researching and developing an idea.] So firms tend to locate where the market is, workers follow firms, the market gets bigger, more firms move close to the market, more workers follow those new firms, and on it goes. You have a virtuous circle of development. Of course, the "market" can be just a large city, or a city with a port from which you can ship things abroad -- from where you can go global. Just think why New York, London, Mumbai or Shanghai became what they are.

The problem comes when, for some reason, people do not follow jobs -- when they do not migrate, or not in sufficient numbers, which seems to be the case in South Asia. Then, a country gets split into "leading" and "lagging" regions. The former become more productive, more globalized, and more technologically sophisticated. The latter get stuck at low levels of income, mostly relying on subsistence agriculture. In the aggregate and from the outside, the country may look very good (who would question India's shining future?), but a mass of poverty remains at its core. True, when the economy as a whole grows fast, everyone may benefit, even if indirectly -- after all, the income of poor farmers depends on selling their produce to rich urbanites. But what almost-never happens is that lagging regions catch up. There is no convergence -- if anything, there is divergence. You may think: so what if a few provinces remain backwaters, as long as the nation progresses? Well, experience suggests that regional disparity breeds violent conflict, and eventually everybody is worse off -- pick any recent civil war and chances are that it is rooted in grossly uneven development.

What should governments do, and not do? The first obvious tool is a progressive system to share among regions the taxes collected by the central government -- that is, to transfer more money to poorer provinces and municipalities. In a way, richer regions subsidize poorer ones. In practice, this rarely solves the problem. Governors and mayors in lagging areas usually lack the capacity to invest the money well, or to use it to improve public services that are critical for development, like education and health. Things get even more complicated if the transfers are paid for by selling oil, gas or minerals that are extracted from the less-developed parts of a country, whose residents may then feel that they do not profit enough from their own natural resources. You will find many cases of this in West Africa.

Second, help people migrate. This is less dramatic than it sounds. When the workers of a lagging region leave for better jobs elsewhere, they send remittances back home and wages improve for those left behind. Over time, everyone is better off. China is a good example of that. The problem is that, many times, governments get in the way. By granting subsidies to specific activities in specific places -- say, giving away fertilizers for rice cultivation in far-away villages where water is scarce -- they force people to stay put, producing something they may not be competitive at, lest they lose the subsidy. On the other hand, the kind of public investment that would help mobility -- like portable skills and employment services -- often is missing in lagging areas. Ergo, people get geographically "trapped" in poverty.

And, third, make it easier for agriculture to become "commercialized," that is, for larger enterprises to start working the land with the kind of equipment, technology and management that can make local produce exportable. This, in effect, connects lagging regions with the rest of the world. It could take the form of small farmers getting together into cooperatives. More likely, it means reform in land ownership -- giving rural families clear titles to their plots, which allow them to sell if they want and migrate if they wish. This was one of the keys to Brazil's success in transforming its "Cerrado" -- a giant savannah that was once known for its remote backwardness -- into a global agricultural player.

What if transfers from the central government, migration and agricultural transformation do not work? Are lagging regions condemned? Not necessarily. Technology may still change the geography of economic growth. The web has brought people together, wherever they live. It has also made information more accessible to firms, wherever they locate. Cheaper transport has fomented the break-up of the production chain -- the final product is assembled with parts made in many locations, usually across country borders, allowing less-developed regions to contribute the easier components. East Asia is a pioneer of this. Most promising, new production techniques -- keep an eye on "three-dimensional printing" -- have reduced the importance of "economies of scale". One day, size will no longer matter. [It is said that Detroit's car manufacturers can now break even in new models if they sell about two hundred thousand units, instead of over one million as in the past.]

So, the overall point is clear: With time, regional disparities may very well disappear by themselves. But, for countries that cannot wait -- and for political reasons most cannot -- government action is not only possible but also necessary.