Taken together, the results imply that anti-opium efforts substantially increased the opiate-industry resources flowing to the Taliban. For each kilogram of opium removed from the market, the estimates imply that only one-sixth of a kilogram would have come from Taliban-heavy areas. Demand is sufficiently inelastic that, absent supply responses, the value of remaining opium would rise substantially, increasing the net income of farmers in Taliban-heavy areas by around $120. Supply responded quickly, and did so largely in Taliban-heavy districts. All told, the opium-source income of farmers in these districts rose from $240 million in 2004 to $580 million in 2010.
Check his Figure 1 on page 4. I didn't know these facts but they make sense. The U.S. military was less able to suppress production in Taliban-friendly areas than in areas that the U.S. military controlled.
When I asked General Stanley McChrystal, at a Hoover event, why he had wanted to suppress one of the few parts of the Afghan economy that seemed to thrive, I had in mind the problem that the U.S. government would hurt its relations with the people in the area it controlled. I didn't think of what Clemens found. McChrystal gave a flippant answer about how he ought to have known he would get a question like this given that he was speaking in California.