Thursday, July 19, 2018

Conflict Minerals and Unexpected Tradeoffs

The cause seemed worthy, and the policy mild. Militia groups in the Democratic Republic of the Congo (DRC) were taxing and extorting revenue from those who mined like tin, tungsten, and tantalum. Thus, Section 1502 of the Dodd-Frank Act of 2010 required companies to disclose the source of their purchases of such minerals. The hope was to reduce funding for the militias, and thus to benefit people in the area. Human rights advocacy groups supported the idea. The Good Intentions Paving Company was up and running.

But tradeoffs are no respecters of good intentions. Dominic Parker describes some research on the tradeoffs that occurred in  "Conflict Minerals or Conflict Policies? New research on the unintended consequences of conflict-mineral regulation" (PERC Reports, Summer 2018, 37:1, pp. 36-40). Parker writes:
"First, Section 1502 initially caused a widespread, de facto boycott on minerals from the eastern DRC. Rather than engaging in costly due diligence to identify the sources of minerals—and risking being considered a supporter of rebel violence—some U.S. companies simply stopped buying minerals from the region. This de facto boycott had the intended effect of reducing funding to militias, but its unintended effect was to undercut families who depended on mining for income and access to health care. The decreases in mineral production rocked an artisanal mining sector that had supported an estimated 785,000 miners prior to Dodd-Frank, with spillovers from their economic activity thought to affect millions.
"Second, the legislation changed the relative value of controlling certain mining areas from the perspective of militias, who changed their tactics accordingly. Before the boycott, the militias could maximize revenue by taxing tin, tungsten, and tantalum at or near mining sites. They therefore had an interest in keeping mining areas productive and relatively safe for miners. After the legislation, the militias sought to make up for reduced revenue in other ways. According to the evidence, they started to loot civilians who were not necessarily involved in mining. They also started to fight for control over other commodities, including gold, which was in effect exempt from the regulation."
One result of the economic losses in the area was a sharp rise in infant mortality rates: "The combined evidence suggests that Dodd-Frank increased the probability of infant deaths (that is, babies who died before reaching their first birthday) from 2010 to 2013 for children who lived part of or all of their first year in villages targeted by the legislation and mining ban. The most conservative estimate is that the legislation increased infant mortality from a baseline average of 60 deaths per 1,000 births to 146 deaths per 1,000 births over this period—a 143 percent increase."


The level of violent conflict affecting civilians actually seemed to rise, rather that fall: "At the end of 2010, after the passage of Dodd-Frank, looting in the territories targeted by the mining policies became more common and remained that way through much of 2011 and 2012, when our study period ended. ... The incidence of violence against civilians also increased in the policy regions after the legislation ..."

One economic insight here is the "stationary bandit" theory that when a bandit remains in one location, there are incentives for the bandit to keep local workers and companies safe and productive.

The political insights are fuzzier. One can't rule out that if the Dodd-Frank provisions had been better thought-out or better targeted, maybe the effects would have been better, too. Or maybe this is a case where long-run benefits of these provisions will outweigh short-run costs. But it's also possible that an alternative strategy for bolstering the economy and human rights in the area might have worked better. And it's quite clear that those who supported this particular conflict mineral policy did not predict or acknowledge that their good intentions could have these adverse consequences.