When it comes to secessionism, identity trumps economic development

India, aware of the separate identities problem, tried to solve it 72 years ago through articles 370-72 of the Constitution. That success should not be challenged by trying to forge a single identity

It has been an article of faith in India that economic development is the best way of "winning the hearts and minds" of the Muslim people of the Kashmir valley. Give them jobs, roads, schools, colleges, hospitals etc., and they will start loving India. 

Indeed, this has been the general solution tried in all countries that face a rebellion by a section of the population and their desire for secession. But it has never worked. 

So three economists, Klaus Desmet Ignacio, Ortuño-Ortín and Ömer Özak, tried to find out why. They conducted extensive surveys and presented their findings in a working paper uploaded on the NBER website. This paper is one of the best I have read in a long time. 

It "analyses whether the propensity to secede by subnational regions responds mostly to differences in income per capita or to distinct identities."

It's a fascinating paper in which they say, "Counterfactual analysis strongly suggests that identity trumps income in determining a region's propensity to secede."

They examined 173 countries for the secessionist impulse and found that "Removing identity differences reduces the average support for secession from 7.5 percent to 0.6 percent of the population."

In other words, the old adage, give the dog a bone to stop it barking, doesn't work. Identity is a powerful motivator of secession. 

The authors took 3,003 subnational regions worldwide —yes, 3003— and devised ways of measuring "the instability of countries and the propensity to secede of subnational regions." The results are most instructive. 

So we have them saying that the utility a person derives from some infrastructure, or any public good for that matter, is inversely proportional to that person's sense of separate identity. 

More importantly, the authors asked, "by how much income per capita would have to drop for secessionist regions to give up on their demands… large drops in income are needed, suggesting that economic differences only matter if they are large." 

The short message is to let the separatists suffer. We should stop pampering the Kashmiri separatists. 

There are other lessons, like language, ethnicity and religion being crucial for demands for separate sovereignty. Here language matters the most. 

One might recall Selig Harrison's book "India: The Most Dangerous Decade" about India breaking up after the 1956 linguistic division of the Indian states. India got it right. Harrison got it wrong. 

The authors conclude that "at this point, no good data exist that capture the multi-dimensionality of identity groups at a subnational level for the entire globe. Collecting such data should be the subject of future research."

Finally, they say that it is also possible for groups of people with separate but similar identities to seek separation, either to form their own country or to ask to join another country. 

The short point is that once a group starts thinking of itself as being different and that it could run its affairs better as a separate country, it's difficult to change its thinking. It will go on and on seeking separation without much thought to the economic consequences of separation. 

India, aware of the separate identities problem, tried to solve it 72 years ago through articles 370-72 of the Constitution. It harnessed diversity well. That success should not be challenged by trying to forge a single religious or linguistic identity because diversity exists even within a single religious identity.