There’s one line in the middle of a sweeping new study of the most controversial National Security Agency (NSA) program revealed by Edward Snowden that sums up the report’s central conclusion: “surveillance of American phone metadata has had no discernible impact on preventing acts of terrorism.”
That’s a bold claim, but there’s a lot in the research to back it up. The researchers at the New America Foundation, a DC think tank, compiled a database of 225 counterterrorism cases where U.S. citizens or residents were either indicted on terrorism charges or killed before they could be charged. They then used court documents, public reporting, and any other relevant research to figure out what alerted the government to the case and where the government got the evidence it needed to go to court.
They found that the searches under the two provisions that purportedly authorize NSA “bulk collection” of U.S. citizens’ metadata, either under Section 215 of the PATRIOT Act directly authorizing bulk collection or the more indirect Section 702 of the 2008 FISA Amendments Act, were far less helpful in catching alleged terrorists than the NSA and Obama administration may have said. Here are the five main takeaways from their research and what this should tell us about the debate over reforming the NSA going forward:
1. The NSA metadata programs rarely find terrorists on their own. The New America researchers found that only 14 cases out of the 225 — under 6.5 percent — were launched because of either metadata program. The majority of indictments grew out of more normal law enforcement tools — an informant (16 percent), a tip from family members or communities (17.8 percent), intelligence gathered by a non-NSA intel agency (8 percent), and so on. The public record, it seems, just doesn’t support the claim that the NSA collection of U.S. citizen metadata plays a crucial role in helping the government find terrorists that otherwise would have flown below the radar.
2. NSA metadata collection isn’t totally useless; it’s just not very useful. When you start to look at the specific cases that metadata surveillance played a key role in, the picture gets grimmer for the NSA. The New America researchers found that metadata searches never led to an arrest that actually prevented a terrorist attack. Take, for instance, the arrest of Basaaly Moalin — the only case, according to NSA Director Keith Alexander’s sworn testimony, metadata surveillance may have sparked an investigation that stopped “terrorist activity.” All the government found on Moalin was $8,500 donated to Somali terrorist group al-Shabaab; “the case involved no attack plot anywhere in the world,” according to New America’s review. In almost all of the cases profiled by New America, metadata surveillance played a supporting role, unnecessarily displaced traditional investigative methods, or failed to uncover a serious plot.
That’s not to say it does nothing. Perhaps the best case for the NSA examined in the report is the arrest of Khalid Ouazzani and his two co-conspirators. The government argued 702 surveillance uncovered their $67,000 in contributions to al-Qaeda and affiliated individuals, as well as conversations about bombing the New York Stock Exchange with al-Qaeda handlers. New America said “little evidence is available to contest the government’s assertion” that metadata “played a role in this investigation,” and, though the bombing plan “was not a serious threat,” the financial contributions were “certainly worrisome.” Metadata programs appears to have caught low-level activity in the Ouazzani case, but there’s no evidence that they ever played a major role in preventing more concrete, dangerous terrorist plots. That’s consistent with the conclusion of an Obama Administration review board, which found that at least one of the NSA programs was “not essential to preventing attacks.”
3. But there’s still a lot we don’t know about counterterrorism arrests. In the plurality of cases (27.6 percent), the New America report was simply unable to identify the cause of the investigation. It’s certainly possible that in some of those 62 cases, metadata surveillance led to real breakthroughs. In around a third of the unclear cases, New America’s research uncovered enough of a role for an informer to suggest it was unlikely that NSA metadata kicked off the investigation. However, the fact is we don’t know a lot about many terrorism arrests, and it may be metadata is playing a bigger role behind the scenes. In three cases New America identified, there was good reason to believe the NSA played a role in the arrest even though the U.S. government doesn’t officially claim them as “NSA successes.”
4. Government claims lack credibility. President Obama, General Alexander, and other U.S. government officials have repeatedly claimed that NSA metadata surveillance has headed off 54 terrorist “events” around the world, 13 of which targeted at the United States. No one knows what an “event” actually means, and the New America research uncovered at most one case in which metadata surveillance might have aided in preventing “an operational al-Qaeda plot to conduct an attack within the United States.” So when people like Rep. Mike Rogers (R-AL) say the NSA “thwarted terrorist attacks” 54 times around the world, it’s worth noting there’s almost no public corroboration of these claims, particularly as they related to the U.S. homeland.
This pattern of “repeatedly exaggerat[ing] the role of NSA bulk surveillance programs in preventing terrorism,” in New America’s words, means that U.S. officials have torched their credibility on the question of whether metadata surveillance prevents terrorism. If they want us to believe that the NSA needs the powers it has, they need to provide hard, public evidence in favor.
5. It’s time for reform. The NSA surveillance debate keeps coming back to the question of whether Edward Snowden should have released the documents detailing the documents that he did and, relatedly, whether he should be prosecuted or pardoned. These conversations are certainly important, but they need to be bracketed from the question of what to do about NSA metadata surveillance. The New America report reveals that we’ve got enough public evidence to start making preliminary judgments about the NSA programs’ utility in stopping terrorism; we can also start making judgments about the degree to which these programs threaten civil liberties. Making this assessments, and sussing out what they mean in terms of needed policy reforms, is a task that can be done without any need to come to judgments about Snowden himself.
Both the Obama Administration, which is privately deliberating on how to reform the NSA, and public commentators would do well to treat NSA and intelligence reform more broadly (New America’s work implies the intelligence community still has a serious intelligence sharing problem) as policy questions in their own right.