America’s Unnecessary Secrets

Published in The New York Times.

The danger of excessive government secrecy is a lesson we should have learned over the last decade. Although the proper classification of information is vital to keeping the nation safe, “overclassification,” as the 9/11 Commission found, jeopardizes national security by inhibiting information sharing within the federal government and with state and local agencies.

Overclassification also corrodes the democratic process by leaving the public and even Congress to debate the issues of the day without full information — as happened, for example, in the lead-up to the Iraq war. And it forces the government to waste billions of dollars on security measures to protect information that doesn’t need protecting.

Unfortunately, overclassification continues to be rampant. In fiscal year 2010, officials made 77 million decisions to classify information. Even the most security-minded government officials — including Donald H. Rumsfeld, the former defense secretary, and Porter J. Goss, the former director of national intelligence — have said that far too much information is classified. Defense Department and National Security Council experts have estimated that anywhere from 50 percent to 90 percent of classified documents could safely be made public.

Why is there so much overclassification? Because there are so many incentives, unrelated to national security, to classify. Classifying documents indiscriminately is easier than giving each decision careful time and thought. Officials fear sanctions for mistakenly releasing sensitive information. It is easier to get things done in government when there are fewer people involved. Information is a key weapon in turf wars between agencies, and hoarding information increases officials’ sense of importance. Finally, officials who are involved in government misconduct have a powerful incentive to hide the evidence.

By contrast, while officials face harsh sanctions for failing to protect sensitive information, they are rarely, if ever, penalized for classifying documents improperly. In the ordinary course of business, classifiers are not required to justify their decisions, and no one reviews their judgments. In short, there is no accountability.

President Obama has taken some steps to try to reduce overclassification. In a December 2009 executive order, he directed officials not to classify information if they had significant doubts about whether it needed such protection. He also declared that no information could be classified indefinitely and increased the training requirements for officials who classify documents. These were important first steps, but they are unlikely to solve the problem because they do not address the skewed system of incentives.

The president has announced his intention to undertake a “more fundamental transformation” of the classification system, and his advisory committee on classification is currently drafting recommendations for a new executive order. To be effective, that order must rebalance the existing incentives by introducing accountability. As a report co-authored by one of us recently concluded, a well-designed accountability mechanism would include three elements.

First, classifiers should be required to explain their decisions. E-mail and word processing programs on classified computer systems generally include drop-down menus that allow officials to classify documents. These programs should include prompts that request basic information about the classification decision, including the national security harm that could result from disclosure.

Second, agencies should audit officials’ classification decisions. There are too many classifiers (approximately four million) to permit yearly audits of all of them, but agencies could conduct “spot audits,” selecting classifiers at random and reviewing samples of their decisions. The agencies’ inspectors general, whose offices operate with relative independence, should conduct the reviews. They could use the classifiers’ explanations as their starting point, but they should have access to any agency information they deem necessary to evaluate those explanations.

Finally, audits should be coupled with consequences. A classifier who performed poorly on an audit should be subject to repeat audits every six months. A continuing pattern of improper classification should trigger mandatory escalating consequences, beginning with remedial training, proceeding to a note in the classifier’s personnel file and culminating in revocation of the official’s classification authority.

By holding classifiers accountable for their decisions, the president could make great strides toward solving a problem that imperils national security, weakens our democracy and needlessly saps the treasury. And 10 years after terrorists threatened to shake our commitment to our values, the United States would show the world that an open government and an informed public are among this nation’s greatest sources of strength.

Elizabeth Goitein co-directs the Liberty and National Security Program at the Brennan Center for Justice at New York University School of Law. J. William Leonard was the director of the Information Security Oversight Office from 2002 to 2007.