Eight Steps to Reduce Overclassification and Rescue Declassification by Elizabeth Goitein, The Brennan Center for Justice

A White Paper Submission to the PIDB:

In a political climate where consensus is rare, there is remarkably little dispute about the need for classification reform. Officials from Democratic and Republican administrations agree that far too much information is unnecessarily classified and that the current declassification process is wholly inadequate to handle the oncoming wave of classified digital information. It is also clear that our dysfunctional classification system harms national security, as it inhibits information-sharing and invites carelessness.
Below are eight measures the next executive order could include that would go far toward reducing overclassification and rescuing declassification.

Reducing Overclassification
Overclassification is enabled by a lack of objective criteria to guide original classification decisions. While officials must be able to exercise discretion and judgment, these should not be unbounded. The classification categories listed in section 1.4 of the order are too broad to provide meaningful constraints. Moreover, the concept of “damage to the national security” is not defined and is extremely elastic. The next executive order should create a White House-led commission of senior agency officials charged with (1) tightening the criteria for classification, and (2) providing a definition of “damage to the national security” that sets an appropriately high bar.

Ensuring that the criteria are sufficiently specific will require addressing the treatment of “intelligence sources and methods.” Some agencies interpret the National Security Act to categorically bar public disclosure of sources and methods, without any further inquiry into their sensitivity. In fact, the law requires the Director of National Intelligence to protect against “unauthorized” disclosures only, and makes clear that the classification of sources and methods must be under appropriate legal authority. The next executive order should clarify that sources and methods may be classified only if they otherwise meet the criteria for classification.

Another problem lies in the guidance provided to derivative classifiers, whose role is not to exercise independent judgment but merely to carry forward original classifiers’ decisions. The topics in some agency classification guides are broad or vague, deputizing derivative classifiers to make their own assessments of national security harm. Many guides are outdated and thus unreliable. And classifiers are understandably overwhelmed by the sheer number of guides and topics. The next Fundamental Classification Guidance Review should be directed at addressing these problems.

A major source of overclassification is the absence of disincentives. In theory, agencies may penalize officials for improper classification, but there is little appetite to do so and no system in place to identify individual offenders. The executive order should direct agencies to implement an auditing system, such as the spot audits recommended by the Brennan Center, with mandatory consequences for repeated or intentional misuse of classification.

The classification of rules or legal interpretations that set binding standards for government conduct – i.e., “secret law” – engenders unique harms and constitutional concerns. The next executive order should impose a higher substantive bar for classifying legal authorities and implement additional procedural safeguards to ensure they are classified only in rare cases of clear necessity. The Brennan Center has developed specific recommendations along these lines.

Rescuing Declassification
The current executive order requires “automatic” declassification of information after 25 years unless it falls under certain exemptions. In fact, declassification is anything but automatic. Multiple agencies perform lengthy “equity” reviews, and the so-called “Kyl-Lott amendment” requires line-by-line review absent certification that the information is “highly unlikely” to contain nuclear information. As long as these impediments remain, declassification has no chance of keeping pace with classification. The next executive order should authorize the National Declassification Center to declassify records at 25 years without agency referral, and empower agencies to make broad categorical decisions about which records are “highly unlikely” to contain nuclear information.

Inexplicably, there is no system in place to identify and declassify information classified for less than 25 years, unless it is the subject of a Freedom of Information Act (FOIA) or Mandatory Declassification Review (MDR) request. Such information often languishes until the 25-year mark is reached. The next executive order should direct the implementation of a two-tiered system for declassifying information classified for less than 25 years. Where the declassification date is tied to the completion of specific operations or events, the document should be marked accordingly, and classification should be “self-cancelling” – i.e., declassification should happen electronically and without review. In all other cases, an electronic alert should trigger review when a document reaches the declassification date.

Finally, MDR has been highly effective in achieving declassification of information classified unnecessarily or for too long. However, it operates too slowly to be useful when the information sought relates to immediate public controversies. MDR lacks an expedited review mechanism like the one that exists under FOIA. The next executive order should rectify this and establish an expedited review track where the requested information would contribute to a significant current public conversation or debate.