NARA and Declassification

Statement of Steven Aftergood

Federation of American Scientists before the Public Interest Declassification Board

July 22, 2010

Thank you for the opportunity to address the Board. My name is Steven Aftergood and I direct the Project on Government Secrecy at the Federation of American Scientists.

I want to briefly discuss why I believe that the entire classification category known as Formerly Restricted Data (FRD) should be eliminated. The primary reason for eliminating FRD is that it adds needless complexity to an already baroque classification system and it poses an unnecessary obstacle to the efficient functioning of the declassification process.

Three Classification Systems is Too Many

It is sometimes said that we really have two classification systems – one for national security information (NSI) that is classified by executive order, and one for nuclear weapons-related information (RD and FRD) that is classified under the Atomic Energy Act.

But one could just as well say that we really have three classification systems because the classification rules for Restricted Data (RD) and for Formerly Restricted Data (FRD) are sufficiently different from each other that they practically constitute two separate systems all by themselves.

Information enters classified status differently in all three systems: RD is “born secret” under the Atomic Energy Act; FRD is created by being transclassified from RD; and NSI is classified under the executive order by a presidentially authorized original classifier.

What is even more important is that information is also declassified differently in all three cases. The Atomic Energy Act permits declassification of RD when it can be done "without undue risk." But strangely enough, the legal standard for declassifying FRD is when it can be done "without unreasonable risk." It is hard to say what the difference between those two standards is, but they are apparently different, which is a source of confusion all by itself. And both RD and FRD differ from the NSI standard of information that no longer could reasonably be expected to cause damage to national security. There are other declassification differences as well that are important in practice – notably, RD can be declassified by the Secretary of Energy, but FRD requires joint concurrence by DOE and DoD. Also, NSI can be reclassified after declassification, but RD and FRD cannot. And so on.1

All of this, I think, is ridiculous and unnecessary. In fact, it’s worse than that. It is counterproductive because it adds useless complexity to the declassification process. This is especially problematic when the three classification categories are intermingled in record collections and even in individual documents – necessitating multiple declassification reviews, raising declassification costs, or stalling declassification altogether.

The FRD Category Should Be Eliminated

Several reviews that considered the problem of FRD over the last two decades have concluded that this classification category should simply be eliminated. That was the recommendation of DOE’s 1992 Classification Policy Study,2 a nd of the National Research Council’s 1995 review of DOE classification policy and practice.3 This Board’s own 2007 report on Improving Classification also described problems with the treatment of FRD.4

I disagree somewhat with the Board’s 2007 discussion of FRD, and also with its resulting recommendations, for the following reasons.

The main problem with FRD, in my opinion, is not that the public cannot request its declassification, as the Board suggested. The real problem, as I’ve said, is that the mere existence of FRD complicates and slows the declassification even of NSI – since it needs to be screened out and handled under separate procedures. Even if procedures for the public to request FRD were improved, this multiplicity of classification categories would still be a problem.

I also doubt that this problem can be rectified by the President through an amendment to the Executive Order on classification, as the Board recommended. Even if the President ordered that FRD should be treated as “defense information” (as per the Board’s recommendation 1), it would still need to be handled and processed separately from other NSI defense information. That is true for several reasons: unlike NSI, FRD cannot be transmitted abroad without a special cooperative agreement, it can only be declassified jointly by DOE and DOD, and it cannot be reclassified. All of the “friction” generated by the existence of FRD would remain in effect.

So I believe the solution has to be a legislative one. The Atomic Energy Act needs to be amended to eliminate the FRD category and to clarify that FRD that cannot be immediately declassified should be treated as NSI in every respect from now on.

If it were up to me, I would go further and eliminate the RD classification category as well. At this moment in history, when there are multiple nuclear weapons programs around the world, including even in North Korea, I think it has ceased to be true that nuclear weapons information is so uniquely sensitive that it requires its own classification system. The classification system for “regular” national security information has proven itself sufficiently versatile that it can reliably protect the most intricate technological secrets along with military, intelligence and foreign policy matters. Three classification systems is two too many.

Others may disagree. In any case, if a repeal of the Atomic Energy Act’s system of controls on Restricted Data is too ambitious or too much to hope for, the elimination of FRD should not be. Almost everybody realizes that as a classification category it is useless, unnecessary and disruptive. And the time has come to act on that realization.


The Cold War national security classification system that we have inherited is under stress from multiple directions. It is challenged by the explosion of digital records, the imperative of information sharing, the growing financial and operational costs of secrecy, the backlogs of records awaiting declassification review, and public skepticism about the integrity of the entire process.

It may turn out that the politics of classification are too hard and that the classification system cannot be “fixed.” Maybe it just has to be allowed to “break,” either gracefully or catastrophically. But I would say that it’s still too early to reach that pessimistic conclusion. There are still some ways of fixing things that have not been tried and have not failed.

We should be working collectively to salvage what is useful in classification policy, and to jettison what is not. A workable classification system of the near future will be simple in design, easy to implement and to correct, and modest in scale.

But Formerly Restricted Data is not consistent with that goal, and so it needs to go.


1. The differences in the declassification procedures for NSI, RD and FRD were carefully delineated by Arvin S. Quist in "Security Classification of Information," vol. 2, chapter 11, prepared for the Department of Energy, April 1993; available at:

2. "Classification Policy Study," prepared by the Meridian Corporation for the Department of Energy, July 4, 1992, pp. 54-56; available at:

3. "A Review of the Department of Energy Classification Policy and Practice," National Research Council, 1995, p. 5, 47-48; available at:

4. "Improving Declassification," Public Interest Declassification Board, December 2007, Issue No. 13, pp. 32-33; available at: