By Fareed Zakaria, CNN
Since 9/11, a decade ago, Washington has given the government of Pakistan more than $20 billion in aid. Two- thirds of that has gone to the military to fight the war on terror. The other third, about $6 billion, has gone for development of Pakistan's civilian economy and society.
The theory behind many of those billions of dollars is that by bringing Pakistan's poorest out of poverty and despair, fewer young men will be seduced by radical Islam.
As President Obama said, "We must stand with those who want to build Pakistan." And Secretary Clinton adds, "Providing this assistance is not only the right thing to do, but we believe it is essential to global security and the security of the United States."
This is an admirable approach for sure; it's a strategy the U.S. uses all over the world, from Baghdad to Bali and back.
But what if you discover that aid money does not necessarily make Pakistanis less likely to turn to terror? What if you learn that there's actually no correlation between being poor and supporting Islamic extremism?
Well, that's what a new serious academic study seems to prove. It's a robust survey by four academics from Princeton, Georgetown and the University of Pennsylvania that conducted extensive field research in Pakistan, interviewing 6,000 people across a broad spectrum of income groups and geography. Their findings could challenge the way we approach fighting terror, not just in Pakistan but around the world.
First, they find that in general Pakistanis don't like militant groups. Not just al Qaeda, but the other ones like the Pakistani or Afghan Taliban.
Second, contrary to conventional wisdom, poor Pakistanis dislike militant groups more than the middle classes.
Third, the people who hate militants the most are the urban poor, probably because more than any other group they're the ones who are affected by terror attacks - bombs in subways or cafes or whatever.
It's an interesting conclusion. The people we've long considered the likeliest candidates for extremism are actually the ones most against it.
The study points out that this goes against most of the existing policy literature on the subject.
It cites both the U.S. State Department and the UK's Department for International Development as saying poverty motivates people to extreme violence.
Now, giving aid to poor people is good in and of itself. But if we've been doing that to prevent them from becoming Islamic fundamentalists, then this study suggests we've been aiming at the wrong target. Perhaps our focus should be on the middle classes or on secular education.
Research shows that members of the IRA in Northern Ireland or Hezbollah's militant ring are more likely to come from economically advantaged families and with a relatively high level of schooling.
These are important issues for Washington to consider. Who does it want to give economic aid to? What exactly is the best way to create a climate less conducive to extremism? Those are long-term issues.
For now, what I want to say in the short term is: Let's at least focus on accountability for this aid – for both hard and soft aid. We need to demand results. What is the money achieving?
A CNN poll from last week shows that nearly half of all Americans think all aid to Pakistan should be stopped. Another quarter thinks it should be reduced.
It seems like 10 years and $20 billion later, the American people understand basic lessons in accounting that Washington has been learning the hard way. Pentagon documents now show that we are rejecting nearly half of Islamabad's claims for expenses over the last two years.
But the more important question is: Will the Pakistani military in return for all this money, finally move against the terror organizations like the Haqqani Network and Lashkar-e-Taiba that they claim to be willing to battle?
It's long delayed, but it is the right message that we should be sending.
American purse strings are important and necessary from Afghanistan to Zimbabwe, but they don't have to remain open at all costs.