10 November 1998
Remarks by Roslyn A. Mazer, Chair
Interagency Security Classification Appeals Panel
November 2, 1998
I am delighted to have this opportunity to discuss the work of the ISCAP at the outset of this important conference. I thank Thomas Curtis, Program Manager of the Declassification Productivity Initiative at the Department of Energy, and Edmund Cohen, Director of the Office of Information Management at CIA, for inviting me to speak this morning.
I'd like to give you a brief overview of ISCAP: our responsibilities, our accomplishments, and how we fit in to the overall declassification program under Executive Order 12958, to which all of you have devoted so much effort. Then, drawing on ISCAP's experiences with a few difficult, recurring issues, I'll suggest how we could generate a more efficient declassification program and greater public access to the historical record, without compromising the current functioning of the intelligence community or any other aspect of our national security. In that connection I will discuss the real costs of secrecy -- not just the dollars and cents -- though these are astronomical -- but the opportunity cost to our policymakers, historians, scientists, and citizens, and the cost in terms of growing cynicism about government. I will conclude with three observations about how, together, we can propel the Clinton Administration's openness agenda into an even more energetic gear in the next few years.
As you know, in issuing Executive Order 12958 in 1995, the President made profound changes in the declassification program throughout government. As President Clinton noted when he announced the new order: "Protecting information critical to our nation's security remains a priority. In recent years, however, dramatic changes have altered, although not eliminated, the national security threats we confront. These changes provide a greater opportunity to emphasize our commitment to open government."
One of the most significant, 180-degree turns made by the Executive Order was to reverse the resource burden for over 25 year-old information. Unlike the prior systems, in which agencies had to expend resources in order to declassify older information, agencies are now required to expend the resources to retain classification -- to demonstrate how older, historical information falls within one of the narrow exceptions to automatic declassification. The ultimate goal was to mandate the maximum responsible disclosure of older classified information.
Among the several innovations of the Order was the creation of the ISCAP. Before ISCAP, there had been no interagency body to hear appeals of classification decisions since 1978. Under the predecessor Executive Order (12356), appeals of agency classification decisions regarding presidential materials were taken to the Information Security Oversight Office, or ISOO.
While serving as ISCAP's chair, I also serve as the Justice Department's voting member on the ISCAP. Today the other five voting members are Jennifer Carrano, appointed by the Director of Central Intelligence, who is Chief, Requirements, Plans, and Policy Office, DCI Community Management Staff; Sheila Dryden, appointed by the Secretary of Defense, who is Principal Director, Security and Information Operations, Office of the Deputy Assistant Secretary of Defense (Security and Information Operations); Frank Machak, appointed by the Secretary of State, who is Information Management Reorganization Coordinator for the State Department; William Leary, appointed by the Assistant to the President for National Security Affairs, who is Senior Director for Records and Access Management at the NSC; and Michael Kurtz, appointed by the Archivist of the United States, who is Assistant Archivist of the United States.
Each member has a liaison to the executive secretary and staff. Steven Garfinkel, as director of the Information Security Oversight Office (ISOO), serves as ISCAP's executive secretary, and members of the ISOO staff serve as ISCAP staff members as well. The staff and liaisons do the preparatory work, and meet as a group at least once before each meeting of the ISCAP members. ISCAP could not function without ISOO's extraordinary assistance and dedication, particularly the indispensable contributions and expertise of Steve Garfinkel.
The ISCAP has three primary functions: first, to hear appeals from an agency head's decision not to declassify information in response to a mandatory review request by a member of the public; second, to hear appeals from an agency head's denial of a challenge to classification; and third, to approve, deny, or amend an agency head's exemptions from the Order's provisions for automatic declassification of permanently valuable information when it becomes 25 years old. (ISCAP does not have a role in the functionally related, but legally distinct procedures to avoid inadvertent declassification and release of Restricted Data or Formerly Restricted Data established by the recent Defense Authorization Act.) To date the ISCAP's work has focused exclusively on mandatory review appeals.
Under our bylaws, at least five of the six voting members must be present to produce a quorum for voting; and the votes of a majority of those present are necessary in order to overturn an agency head's decision. With six voting members, we have had several 3 to 3 votes, in which case the agency head's classification decision is upheld.
An agency head has 60 days to seek a review of an ISCAP decision by the President. To date, despite dozens of decisions to overturn agency heads, no member has sought Presidential review of an ISCAP decision. (Because it exists solely to advise and assist the President, the records ISCAP generates are Presidential records, subject to the Presidential Records Act, rather than federal records, subject to the Freedom of Information Act.)
Since our first meeting in May 1996, ISCAP has voted on a total of 119 appeals. Of these, we have fully declassified 75 documents, or 63 percent. We have declassified significant new information in an additional 26 documents, or 22 percent, while entirely upholding agency decisions to retain classification in the case of 18 documents, or 15 percent. Thus, we have declassified new information in 85 percent of the documents we have voted on -- information that was kept classified at the highest levels of administrative appeal within the agencies. Perhaps more significant, I am confident that had just about all of these appeals been brought before the federal courts under the Freedom of Information Act, the appellant would not have prevailed, and the information would have remained classified in perpetuity unless the agency itself decided to declassify it.
But given the volume of historical information in government archives, what real difference can ISCAP make? ISOO estimates that the 400 million total pages declassified in fiscal years 1996 and 1997 approach just one-quarter of the total universe of classified pages subject to automatic declassification by April 2000.
The question is entirely well-founded. Indeed, it is self-evident that the work of ISCAP would go down in history as a footnote -- perhaps a successful footnote but a footnote nonetheless -- if its work were carried out in isolation, and not communicated to people, such as yourselves, who are responsible for carrying out the broader declassification mandate, and if its decision rationales were not incorporated into agency decisionmaking. That is why I am here today.
Before addressing the multiplier effect I believe ISCAP decisions should have in the next phases of classification management, I would like to touch briefly on the public interest in information contained in the documents that ISCAP has directly acted on. Information in our cases has often proven to be of significant interest to historians, journalists, and other researchers.
For example, several of our appeals involved documents from the Eisenhower, Kennedy, and Johnson presidential libraries discussing the deployment and potential use of nuclear weapons in Europe. Among the subjects discussed were the targeting of weapons against the nations of the Warsaw Pact; command and control of nuclear weapons in emergency circumstances; and the relative authorities of the chairman of the Joints Chiefs of Staff vis-à-vis the Supreme Allied Commander in Europe. With the exception of a small portion or two, the ISCAP voted to declassify this information.
Other cases that have generated interest from reporters and researchers include: six State Department messages from June and July 1967 (the aftermath of the "Six Day War" in the Middle East) discussing the situation in the West Bank and Israeli capabilities and intentions concerning the acquisition of nuclear weapons; and seventeen documents from the Ford presidential library, dating from 1974-76, addressing nuclear material processing and reprocessing by the Republic of Korea, and the potential development of nuclear weapons and missiles by that nation.
Even though ISCAP's cases often involve information of high public interest, the broader point remains: in the grand scheme of things, ISCAP's immediate output is statistically insignificant.
It is for this reason that my fellow ISCAP members and I place such great emphasis on communicating our decisions to others with significant stores of classified documents. We try to raise awareness of the ISCAP and its work through several means:
I now would like to give you the flavor of the ISCAP's deliberative process. From the outset, the challenge to ISCAP has been quite daunting: to apply very different legal standards in a very different, a very dynamic, and a very uncertain geopolitical environment. Moreover, the structural make-up of ISCAP posed an initial challenge: each member was called upon to represent his or her agency but at the same time consider appeals in an objective fashion, without merely carrying the water of the agency's particular equities.
It became quite clear early on that, under standards such as "seriously and demonstrably impairing relations between the United States and a foreign government" or "seriously and demonstrably undermining on-going diplomatic activities," an agency coming to the table with over 25 year-old information would bear a heavy burden. The proponent of classification would be subject to searching scrutiny. I believe that our track record of declassifying 85 percent of some or all of the appeals we have voted on reflects this balanced skepticism toward continued classification.
Speaking as the ISCAP representative of the Department of Justice and not as the ISCAP's Chair, I believe that the ISCAP has voted to keep some information classified that could have been declassified without jeopardizing the national security interests of the United States, and should have been declassified in the interest of the historical record. Given the focus of this conference on the Intelligence Community, I thought it would be helpful to address in a non-classified format two categories of intelligence-related information that are on the fault line, if you will, between the old ways of assessing national security risk, and the more pragmatic standards on the 1995 Order.
In this context, I am struck how information provided by a foreign intelligence service - or even a reference that suggests that a particular foreign service may have provided information - is treated as though it remains in the control of the foreign government forever after. For information that is twenty-five years old, this is an anomaly under the Executive Order. One of the innovations of E.O. 12958 was that it no longer treats foreign government information as categorically exempt from historical declassification. Information that was originally received in confidence through diplomatic, military or other non-intelligence channels may be declassified without prior consultation with the originating government. The current sensitivity of the information is the focus of the decision, rather than simply its foreign origin.
In my view, two elements of the above arguments for continued classification warrant a fresh look:
"For many years the CIA has been working with the Israeli Government and the Palestinian Authority to combat terrorists in their midst. . . . There is nothing new in this role for the CIA. . . . In the past, the CIA has worked to support agreements to end wars in the Middle East, to monitor arms control agreements with the former Soviet Union and to lower tensions between India and Pakistan." (NYT, 10/27/98)
To be sure, not every recurring category of information lends itself to a categorical declassification approach. Some decisions are unavoidably context-specific, even if they involve the application of a fixed rule. Consider, for example, declassification decisions that involve human intelligence sources. The long-standing, generally-accepted principle is that the identities of human intelligence sources warrant continued classification for an indefinite period of time. But, fairly applied, this principle is often only the starting point for the declassification review of older, historical documents. Unless the document names or precisely describes the source, there is a good chance that it can be declassified, in full or substantial part, without posing any reasonable risk of disclosing the source's identity. And, I would contend, the passage of many years diminishes the probability that any reader could, entirely through circumstantial evidence, surmise the identity of an unnamed source.
Consider the following hypothetical example: an intelligence source reports on a political indoctrination session that he and 50 other persons attended. That report should be classified at the outset, even if it does not name the source or otherwise indicate which of the 51 provided the information. Disclosing the report could foreseeably compromise the source; for example, by prompting the targeted organization to look for an informer in its ranks. But consider the same report after 30 or 40 or 50 years. If, as is possible, the targeted organization has survived and retained an institutional memory and motivation to identify the source, then disclosing the report could still result in his compromise. But, under other circumstances, couldn't the substance of the report be released after so many years, without revealing the source's identity? In my view, the damage-to-national-security standards of E.O. 12958 require us to assess such factors carefully in deciding whether and how to make a redacted release of documents presenting this issue.
For ISCAP's part, our membership has consistently voted to retain classification of information that would identify a human intelligence source, even after thirty-plus years. In all but one case, which I will discuss separately, I believe this principle is virtually inarguable. Where there is play in the joints is in situations where particular information provided by the source could be released because it was not source-identifying.
In the one exceptional case I mentioned, the names of the intelligence sources were ultimately kept classified. This case involved an appeal from a classification decision of the Defense Intelligence Agency. This agency had denied a mandatory review request brought not by a researcher, but by an Executive Branch entity, The National Archives' Franklin D. Roosevelt Library. The Library asked DoD to declassify information now in its collection that was maintained in the White House Map Room during World War II. At that time, the White House Map Room had served the function of an intelligence headquarters.
Specifically, the library wanted us to declassify the names of East Europeans who had provided the Allies with information about the occupying Axis armies. The information provided by these individuals had long ago been declassified, but their identities had remained secret. Thus, the only question was whether the sources still required protection; no additional foreign relations or other concerns were implicated.
The arguments presented in favor of classification were the conventional reasons for protecting human sources. First, it was noted that revealing the name of an intelligence source can endanger the source and, even if he or she is dead, endanger members of the source's family. It was argued that such risks were still inherent in disclosure, even after more than fifty years. Second, it was argued that, if a government demonstrates that it may eventually reveal the identity of an intelligence source, no matter how much time has intervened, potential new sources of information may be frightened off.
Against these general principles, the Roosevelt Library argued that the unusual circumstances of this specific case warranted declassification. The passage of some 53 years had in fact diminished the probability of retaliation against, or embarrassment to, the sources and their families. Moreover, the governments against whose interests these individuals acted no longer exist, and are universally reviled.
I would ask you to consider: isn't there some point in time after which we cannot responsibly conclude that the damage-to-national-security requirements for continued classification are met? If so, isn't it relevant to consider the rather obvious realities of a given case such as this -- to recognize that no existing government is poised to exact vengeance on behalf of the Nazi regime. We will be revisiting these and similar questions for as long as we keep classified parts of the official records of the great conflicts of our century.
The recurring nature of these categories of information means that adherence to hide-bound thinking will have a substantial effect on an agency's entire declassification program. The resulting costs of keeping these secrets might be justified if their disclosure would in fact harm national security. But if there is no clear and demonstrable damage to our national security interests, maintaining these secrets that sometimes are secrets to no one could severely retard progress under the Order -- with no compensating benefit. And maintaining secrecy is a very costly proposition.
Let's contemplate the costs of secrecy. The direct government costs of the classification system in FY 1996 were placed at approximately 2.7 billion dollars by the Report of the Commission on Protecting and Reducing Government Secrecy (which was chaired by Senator Moynihan and is often referred to as the Moynihan Commission) (Report, p. 9). This figure does not include CIA or private sector costs, so the total bill is obviously much higher.
But the cost in terms of dollars is the least of it. There are two other, more profound costs: first, the cost in hobbling or blinding policymakers, historians, scientists, and others who would benefit from the opportunity to study intelligence archives, and, second, the cost in contributing to public cynicism about secrecy without purpose -- secrecy merely for secrecy's sake.
This second consequence bears examination. Attorney General Reno spoke to this issue when she addressed the annual convention of the American Society of Newspaper Editors in 1996. She noted that in the past unnecessary classification distorted history, warped intelligence estimates, hid government waste and inefficiency, retarded scientific and academic research, and widened the gulf between the government and its people. How government secrecy breeds public distrust was best captured in the extraordinary report by the Moynihan Commission. I know you will hear more about that tomorrow from John Podesta, the President's Chief of Staff, who of course was a member of the Commission and is one of the Administration's great champions of openness in government. I will highlight just one of the Commission's observations:
"Where government activities have stayed shrouded in secrecy, sometimes for many years, that secrecy at times has contributed to widespread public speculation of government wrongdoing. Sometimes this has resulted in the eventual declassification records, but often the perception that the Government is using classification to hide its misdeeds has already taken root and is difficult to dispel." (Report, p. 52)
I would like to conclude my remarks this morning with three observations that I encourage you to bear in mind as you consider the challenges of classification policy and implementation:
"The concept of loyalty necessarily involved the notion of secrecy. Disloyal employees revealed secrets; loyal employees would not. In such a setting apprehension rose, and so did the dimension of secrecy. More and more matters became classified." (Report, A-48)
Thus, a perceived weakness aroused fear, which resulted in unnecessary secrecy.
But the Cold War reality is that the closed regimes found themselves hopelessly and fatally outpaced by open societies, and ultimately collapsed from exhaustion. This is the reason why our democracy endures, why we live under the oldest living constitutional democracy, and why we cannot export democracy like bananas to formerly closed societies. We prevailed over those societies because of our passion for openness, for trusting our citizens more than we empower our leaders. We celebrate our openness. In fact, it is unnecessary secrecy that is timid and cowardly. Openness is courageous.
Be courageous. Be as open as you responsibly can.