Publications Review Board, basic review guidelines,
Reviewing the Work of CIA Authors
CIA's Publications Review Board (PRB) and its small staff perform a balancing act more than 300 times a year, navigating a process sanctioned by the US Supreme Court to clear the writings of Agency authors for nonofficial publication. The challenge: to balance CIA's secrecy agreement with the Bill of Rights. Business is brisk, as a growing number of former CIA employees seek to become published authors--especially former operations officers reflecting on their clandestine careers abroad.
The variety of material the PRB has reviewed for publication in recent years has encompassed former President Reagan's memoirs, the Brown Commission Report on the Roles and Capabilities of the US Intelligence Community, and broadsides from timeworn Agency antagonist Phillip Agee. Former employees submit manuscripts directly to the PRB, as do some nonemployees--such as former Defense Secretary Weinberger; Judge Lawrence Walsh, the Iran-Contra Independent Counsel; and former members of Congressional oversight committee staffs--who, because of their special access to CIA information, need to seek PRB review before publishing.
The daily "take" logged in by the five-person PRB staff ranges from 1,000-page book manuscripts to one-page letters to the editor. There are speeches, journal articles, theses, op-eds, book reviews, and movie scripts. There are scholarly treatises, works of fiction, and, recently, a cookbook featuring a collection of recipes acquired and served by Agency officers and spouses around the world. Perhaps the most novel review (no pun intended) involved an interactive CD-ROM video spy game co-authored by former Director of Central Intelligence (DCI) William Colby and KGB Gen. Oleg Kalugin.
The reason all this work is reviewed lies in the Agency's need--and its employees' contractual obligation--to protect sources and methods of collection and analysis.
The authority, for both the contractual secrecy agreement and prepublication reviews, rests on the statutory responsibility of the DCI to protect sources and methods and is found in the National Security Act of 1947 and the CIA Act of 1949 as amended, as well as in Executive Orders 12333 and 12958.
The sole purpose of prepublication review is to assist authors in avoiding inadvertent disclosure of classified information which, if disclosed, would be damaging to national security--just that and nothing more.
What is involved in each review is neither censorship nor a declassification process, but rather a determination of the absolute minimum of deletions, if any, that would uphold both the DCI's authority and the individual's constitutional right to free speech under the First Amendment, a right the courts take especially seriously.
Permission to publish cannot be denied solely because information may be embarrassing to CIA or critical of it, or inaccurate. People have a right to their opinions, and they have a right to be wrong. People also have a right to write; our reviews are not aimed at discouraging them. My goal as the Chairman is to be an honest broker, not merely identifying problems but suggesting solutions. Our purpose is to help people to publish in a way that will not cause a problem for them, the Agency, or for the country.
It usually is not hard to write around a required deletion, even when it involves a paragraph or more, in a way that enables the author to retain the point of a passage and the flow of the text. Small changes often can do the job, such as referring to the "office" rather than the "CIA station," or describing a liaison official by an actual government or military title, or in general terms as a senior official, without a specific connection to an intelligence service.
Especially sensitive subjects can be more complicated--for instance, if they have an impact on ongoing operations, identify particular cover or liaison arrangements, or involve a country with which relations are particularly delicate. But over the past two years, only a few authors have thrown in the towel in the face of such major review problems and decided not to publish, at least for the time being. One had to give up on the book he had taken two years to write because he had written of a cover life that could not be revealed. Sadly, he was left to ponder the dubious prospect of redoing--and marketing--the book as the life of an investment consultant rather than as a spy.
But success is the rule. The process works. Negotiations with authors almost always result in finding acceptable substitute language and a meeting of the minds: of more than 600 reviews during the past two years, only three were appealed.(1) Not all submissions sailed through smoothly, to be sure. Some experienced delays, especially in the case of one embarrassingly protracted, seemingly snakebitten review which a prospective author endured with remarkable good grace, and which happily came to a successful conclusion with his manuscript approved.
Reviews Are Not Optional
Reviewing the writings of former employees is not simply an option for the Agency; nor is it a service offered as a convenience. The Federal courts have approved the process, which stems from the DCI's statutory obligation to protect sources and methods. The courts have ruled, in effect, that prepublication review is the only way to carry out the DCI's statutory mandate consistent with the First Amendment. In addition, prepublication review is essential if the Agency is to uphold the validity of the secrecy agreement CIA staff employees and contractors sign as a condition of employment.
Nor is it optional for the individuals who sign a secrecy agreement to seek a review. It is specifically required to protect the sources and methods of collection and analysis they will learn about and which, if revealed, could cost heavily in lives, resources, and continued access to critical national security information.
The review requirement is spelled out clearly in the secrecy agreement, in which the signer agrees to submit for review any material that "I contemplate disclosing publicly or that I have actually prepared for public disclosure, either during my employment...or at any time thereafter, prior to discussing it or showing it to anyone who is not authorized to have access.... I further agree that I will not take any steps toward public disclosure until I have received written permission to do so from the Central Intelligence Agency."(2)
The courts have held that this signed agreement is a lifetime enforceable contract.(3) The courts also have noted that the secrecy agreement is a prior restraint of First Amendment freedom. But they ruled it a legitimate restraint, provided it is limited to the deletion of classified information and so long as a review of a proposed publication is conducted and a response given to its author within 30 days.(4)
If an author seeks to publish without having obtained PRB approval, the Agency can go to court to block publication or can seize the profits if publication already has occurred, even if there is no classified information involved.(5) In deciding to recommend litigation to the Office of the General Counsel (OGC), the Board must be convinced that the Agency can articulate harm to national security flowing directly from the disclosure. And the Board must weigh the risk of going to court, which not only may result in an adverse ruling but which also means identifying and calling attention to damaging information and providing publicity for the book that contains it. Each review thus requires a policy judgment that weighs damage and the prospect of litigation, with judicial precedent in mind.
The court decision in the Marchetti case--growing out of an appeal by former CIA employee Victor Marchetti against a district court injunction requiring him to submit for review his book entitled The CIA and the Cult of Intelligence--became a benchmark for the Agency's prepublication reviews. In its injunction, the court stated that Marchetti [read any CIA author] "may not disclose classified information obtained by him during the course of his employment which is not already in the public domain" (emphasis added). The court also stated that "Information, though classified, may have been publicly disclosed. If it has been, Marchetti [that is, any former CIA author] should have as much right as anyone else to republish it." This begs the question of how information still considered classified got into the public domain in the first place. But it places squarely on the Agency the burden of proof that additional damage will be caused by repeating the disclosure.
The courts also have held that an author may republish information if he or she can cite open sources for the same information.(6) This is a "fig leaf" the PRB may ask an author to wear as a way of indicating to readers that the information is in the public domain and does not necessarily come from unique, inside knowledge. For the PRB, keeping abreast of what is in the public domain on myriad subjects related to intelligence is a major and continuing challenge. It is a determination that comes into play with every review.
Through the mid-1970s, it was CIA's Office of Security that usually reviewed manuscripts intended for nonofficial publication, in association with the OGC and appropriate substantive components. But the marked increase in former employees writing on aspects of their Agency experience, and the Marchetti case in particular, made manifest the need to establish a more systematic review process in CIA. (It seemed clear that the Federal courts presumed such a process was in place and that, if not, it had better be.)
CIA Headquarters Notice 178, dated 10 June 1976, formally established a Publications Review Board and designated the Assistant to the DCI for Public Affairs as chairman. The PRB began to function immediately, and its membership and responsibilities became a matter of regulation.(7)
Putting the PRB Chair and supporting Executive Secretariat in the DCI's area as part of public affairs was seen as a logical fit and is comparable to where the State and Defense Departments and the FBI locate their review functions. The PRB, moreover, is the Agency office that deals above all with the public and almost exclusively with unclassified information. Another consideration was that the PRB often gets called on to assist other Intelligence Community organizations and to address publication issues raised by high-level officials, including former presidents, cabinet officers, and ambassadors who have had access to Agency information. Having the PRB in the DCI area seemed to facilitate such Community use of it.
The House and Senate intelligence oversight committees affirmed this organizational location, as did CIA's Inspector General in 1981 and 1991. Subsequently, a task force on the release of information, commissioned by DCI Robert Gates in 1992, proposed creating an information "czar" to bring together the Agency's information release programs and "be assisted by" the staff of the PRB. Dr. Gates signed a decision memorandum to this effect in January 1993 but departed as DCI the same month, and the concept of a "czar" and concomitant consolidation did not come to pass as envisioned.
Later that year, the PRB staff was placed alongside two major release programs in the Directorate of Administration's Office of Information Technology. With the PRB there--and now in CIA's new Office of Information Management (OIM), created on 1 October 1997, where the PRB staff has been designated the Publications Review Division (PRD)--went the much larger component handling matters related to the Freedom of Information Act (FOIA) as well as the recently developed automatic declassification program that is working on records 25 years old and older under Executive Order 12958. The Historical Review Group, previously based in the Center for the Study of Intelligence, joined OIM at the beginning of 1998.
The FOIA effort, the automatic records declassification program, and the historical review program are engaged in reviewing, declassifying, and releasing official Agency records--which the PRB does not do, although its functions are related to what is and what is not deemed classified. Marked improvement in clarifying and coordinating that relationship should result from the close, day-to-day working relationships being forged between prepublication review and the declassification and release activities in the Office of Information Management.
Unlike the declassification and release components, the PRB does not review or even possess Agency information apart from the records it keeps on reviews. The written material submitted to it is the private property of an author; it is copyrighted, proprietary information. The Agency can neither classify nor declassify it because it does not belong to the Agency. We simply have the right to review it. The PRD files manuscripts in a separate, vaulted area and makes them available only to those directly involved in the reviewing process. We are fortunate that the courts have endorsed that process, which validates it, but which also obligates us to adhere to it and to hold closely the material under review.
CIA regulations explain that the review requirement applies to "all writings and scripts or outlines of oral presentations intended for nonofficial publication, including works of fiction, which contain any mention of the CIA, intelligence data or intelligence activities, or material on any subject about which the author has had access to classified information in the course of his or her employment."(8)
So, if a former CIA analyst, for example, served for a time on a Balkan Task Force and, after retiring, wrote an article on US policy toward Bosnia, the analyst would be obligated to submit the piece for review because of the analyst's access to classified information on that subject area, even if the article did not specifically mention CIA or intelligence. On the other hand, a manuscript about the growing of azaleas or the inadequacy of public transportation, or a murder mystery or romance novel unrelated to intelligence, would not need to be submitted for review.
The PRB is made up of senior officers representing each of CIA's four directorates--Administration, Intelligence, Operations, and Science and Technology--plus the pertinent offices responsible for cover and for personnel security, along with a legal adviser from the OGC. The idea is that each Board member should be able to make focused, high-level policy decisions on behalf of his or her directorate and, in turn, help make the Board decision on behalf of the Agency within the 30-day time limit prescribed by the courts.
The Chair of the Board is empowered to make review decisions on its behalf; indeed, the entire Board rarely convenes. Inasmuch as a number of reviews are in process at any one time, including lengthy manuscripts, the usual procedure is for the PRB to manage the process, conducting research for precedents and consulting only those Board members whose equities are directly involved. We keep the full Board apprised on a biweekly basis of actions taken and the status of reviews in progress.
Standards Are Different
CIA's prepublication review proceeds on two tracks--one for current and one for former employees and contractors. The tougher one is for those still on the job. Current employees and contractors have to submit material proposed for nonofficial publication through their supervisory chain of command to their Deputy Director or the head of their independent office, although they may ask the PRB Chair for guidance as to whether any Agency review is necessary.
The reason for component review is that, for current employees and contractors, the Agency applies a stricter standard: it may also deny permission to publish statements or opinions that could impair the author's performance of duties, interfere with the authorized functions of the Agency, or have an adverse impact on US foreign relations. For example, an analyst providing support to members of an arms control delegation probably should not expect to publish personal views that could undermine the ability of the team to negotiate. The Agency component in which the officer serves is best suited to make that kind of determination. For all component-approved material, the PRD then serves as the Agency's office of record.
The Board may take on the review of a current employee's manuscript, however, at the request of an official in the employee's chain of command--usually if the proposed publication has Agency-wide implications or involves the equities of more than the employee's directorate. This was the case when a disaffected employee proposed publication of a book accusing the Agency of complicity in covering up evidence of Gulf War syndrome; the head of the employee's directorate requested that the Board take over the review, which it did.
When replying to an author, the PRB's response either will identify words or passages for deletion or revision on the one hand or, on the other, will provide a nihil obstat, granting permission to publish by posing no objection. And no objection is just that: posing "no objection" to publication does not constitute official release, confirm accuracy, or endorse an author's views. We ask authors to help make this clear by publishing a disclaimer that spells it out, and most will do so.
When a CIA officer writes about a sensitive intelligence topic that already has been covered elsewhere--for example, in a newspaper article--it tends to confirm the previous publication and lend credence to the story. It gives the information more weight, and thereby might make it more damaging. But it does not amount to an official acknowledgment. People often say, "Well, it might as well be!" Nevertheless, allowing something to be published unofficially, by a private citizen (even a recently resigned CIA officer), does not mean it has been declassified. Deciding we will not go to court over a passage in a book does not mean we have released the subject of that passage. The courts have made clear that statements made by a former employee, regardless of rank, do not constitute official release or acknowledgment.(9) Because this is the case, the Agency may, in response to a Freedom of Information suit, refuse to declassify and officially release records on a subject which, paradoxically, former employees have written about without objection.
The unofficial nature of an individual's writing is the distinction that justifies and explains this apparent double standard. The fact that a reader is likely to infer authenticity and thus attach greater weight to the writings of a former CIA officer can be grounds for denying the author the freedom to write unofficially about official activities of the Agency and his or her involvement with them. But as long as the author complies with the secrecy agreement by seeking review and discloses no classified information that would cause demonstrable damage to national security, the courts hold nonofficial publication to be a fundamental exercise of First Amendment freedom.
Drawing the line about what really is sensitive is not that hard. What gets hard is having to give up on information you would rather not see published, but which you have to conclude you could not successfully litigate. If the information is in the public domain, it is tough to judge that additional damage would be so compelling that it is worth going to court and taking a chance on an adverse decision and guaranteeing free publicity for the very subject we would rather not see in print. It is a calculated risk, and sometimes "you have to know when to fold'em."
The Board's interpretation of damage is not absolute and unchanging. Notwithstanding a firm commitment to fairness and evenhandedness and with every intention of applying standards uniformly, definitions just do not stand still. Definitions of "damage" and of "national security" are neither absolute nor constant. For one thing, an assessment of the likely damage from a disclosure as it pertains to a single country may change drastically from one period of time to another, depending on changing issues in our foreign relations and changing operational equities relating to another intelligence service. A good relationship five years ago may be exceedingly strained today over an issue that had not arisen then. Thus, an author may have written something in the past, with no objection, that would not be allowed today.
In addition, the perception of threat clearly has changed in a palpable way. The end of the Cold War has brought a recognition that national security is less threatened in the wake of that global conflict between two superpower alliances involving, as it did, a pervasive, worldwide espionage war in which the ultimate stakes were a nuclear holocaust. There are still threats, however, as well as new sensitivities, to which the Board continues to apply key principles for protecting the Agency's ability to function effectively. But in the aftermath of the Cold War era, the same dangers do not always apply to the same degree when looking back into that historical context. Coupled with this is another reality: a commitment to greater openness declared by the president, Congressional leaders, and a succession of DCIs. So it is not surprising that some former CIA officers have concluded that "now it can be told," and they are writing more candidly about subjects no longer seen, in many instances, as damaging.
"The Clarridge Precedent"
A notable example--one that, indeed, has effectively set a new standard--is a book entitled A Spy for All Seasons, published early in 1997 by a colorful, former senior CIA officer, Duane C. ("Dewey") Clarridge. Whatever its sales outside Washington, the book seems to have been snapped up by former Agency officers interested in writing their own books, and they use it as a ready reference and guide! Clarridge sought to break new ground, and did, by writing the story of his operational career, assignment by assignment. He played by the rules. He submitted his manuscript for review, and even came to CIA Headquarters to answer questions from reviewers who were stumped by some of the operations he recounted, only to discover that, for security reasons, he had invented composite agents and operational scenarios, as he notes in the foreword of his book.
The PRB--and especially the reviewers to which it deferred in the Directorate of Operations (DO), where Clarridge spent his career--reviewed the book carefully and thoroughly. It was a complicated review, involving several components and more than three decades of people, places, and operational activity. Clarridge was pushing the envelope. He told where he served and when, and he said a lot about what he did, but he did not reveal cover or sources. In reviewing the book, the Board relaxed a restriction previously applied during the Cold War, in favor of allowing former officers to say where they served, so long as that fact alone is not damaging to national security--as it could be in some locales--and to describe in general terms what they did, so long as they do not reveal sources, cover arrangements, sensitive liaison relationships, or covert facilities.
The reviewers now may allow an author to say he or she was the CIA chief in a certain place, but not the chief of station. This is not a distinction without a difference. A station is a covert facility, widely held to be housed in US embassies and connoting a sizable and continuing CIA presence condoned by the host country. It suggests a liaison connection and operations directed from the embassy in conjunction with the official US diplomatic presence. Being described as the CIA chief in a country, however, could mean heading up a small presence, perhaps short term, and perhaps from an office building. Damage presumably is averted by avoiding reference to a precise facility or cover arrangement and by discussing generic espionage activities without all the particulars. Sometimes it is with such fine lines that damage boundaries are drawn.
Allowing former officers to recount where and when they served, rather than limiting them to broad generalizations such as "assignments in Latin America and in Western Europe during the 1970s," is an important gain for an author seeking to convey meaning and understanding through a first-person account. Although introduced with the review of the Clarridge manuscript, what is now dubbed "the Clarridge precedent" subsequently has been applied in all reviews--and even re-reviews that authors requested after the Clarridge book came out. Each manuscript necessarily is judged on a case-by-case-basis, but in most instances--though not all--exact locations and the nature of assignments now can be revealed. Thus, in a review following Clarridge's, an author was allowed to recount in considerable detail his role in what was then the Congo, but in another part of the same manuscript he could refer only to an assignment in "an Asian capital."
Some CIA retirees were aghast when Clarridge's book came out, contending that he was allowed to say anything and everything. That was hardly the case, as he doubtless would testify. The Board sent Clarridge a 19-page, single-spaced letter calling for deletions from his original manuscript, each of which he made. Nevertheless, his review--while making the life of reviewers more difficult by requiring them to weigh, rather than summarily rule out, exact references to time and place--represents a significant gain for authors to come. It is progress that is worth an inconsistency with previous reviews. And it preserves that which is truly sensitive. Protection of cover, for example, did not prevent the telling of a good story. The deletions and revisions such principles require are not unreasonable. (At the offices of the American Civil Liberties Union, while discussing with an ACLU attorney a PRB review brought to it after a lost appeal, it was interesting to see a copy of A Spy for All Seasons at the ready. The conversation made clear that the book had been perused line by line. In the event, the ACLU decided it would not litigate, once it was apparent that the review conformed to that which was given to the Clarridge book.)
Publication of A Spy for All Seasons probably will encourage others to write similar books, but there already was a tendency among operations officers to publish to an unprecedented extent. Before Clarridge's book, for example, a classified edition of Studies in Intelligence included the unclassified article by a former case officer and ex-hostage in Iran that is reprinted in this unclassified edition. (See page 1-45.)
Books Are on the Rise
What we are seeing now in prepublication review is a trend toward more books as a percentage of what we review (more than 18,000 manuscript pages a year for the second consecutive year) and more books by former operations officers about operational activity, which make for more complicated and time-consuming reviews because of the need to check operational files over an extended period. (Current employees who publish unofficially tend to be mainly from the Directorate of Intelligence; former employees who publish are primarily from the DO.) And we are getting more positive treatments of CIA. There was a time when word that a former officer was writing a book made insiders cringe at the prospect of a personal vendetta exposing cover, liaison arrangements, sources, and methods. Careers could be damaged and the Agency's reputation almost certainly tarnished by charges that would not be rebutted. Today, former Agency authors almost invariably assert that their motivations are the good of CIA and a sense of history.
To be sure, what is good for the Agency often is in the eye of the beholder (or the author). For example, some observers saw in Evan Thomas's The Very Best Men a positive portrayal of dashing figures in the Agency's early history, while some reviewers decried its subjects as the very worst men. Clarridge believes A Spy for All Seasons is a "good news" story, but not everyone feels he did the Agency a favor. There still are some at CIA--especially in the DO, where there understandably is a keen awareness that one's working life rests on protecting sources, cover arrangements, and Agency capabilities in particular overseas locations--who believe it is highly inappropriate, if not dead wrong, for those who served in a secret capacity in a secret organization to go public in retirement, no matter how well intentioned the tales they tell. Yet many inside the Agency, probably most of the general public, and virtually all scholars and journalists see greater openness as a positive development. They welcome the candor that helps dispel the notion that everything CIA has done in secret, especially in its overseas operations, has been sinister, scandalous, or stupid.
Although a lot of what we review will never see the light of day, much of it becomes a valuable contribution to intelligence literature. Examples include Dino Brugioni's Eyeball to Eyeball: The Inside Story of the Cuban Missile Crisis; John Waller's The Unseen War in Europe; former DCI Robert Gates's From the Shadows; Clarridge's autobiographical volume; Elizabeth McIntosh's Sisterhood of Spies; and a forthcoming book by James H. Critchfield tentatively entitled Germany: From Enemy to Ally. David Murphy's Battleground Berlin was widely and favorably reviewed on both sides of the Atlantic and touted by The New York Times Book Review as one of the notable nonfiction books of 1997. Former DCI Richard Helms--assisted by William Hood, a former operations officer and published author of Mole and several spy novels--is working on a retrospective about Helms's time at the Agency. If anything, such interesting writing by Agency authors will enhance CIA's reputation and credibility. They help educate, inform, dispel misconceptions, and improve understanding.
Publication Review Board Chairman John Hollister Hedley (left) and David Murphy displaying the German- and English-language editions of Battleground Berlin, which was coauthored by Mr. Murphy.
Facing a New Era
Where are we headed with prepublication review in an era of greater openness? Not down a slippery slope to a diminished ability to function effectively as a secret organization. On the contrary, what we are seeing is a new era based on two indisputable facts: the Cold War is over, and this is a free country. CIA was properly cut a lot of slack during the height of the Cold War, and to an extent it still is. But ours is a robust democracy in which people want and deserve to know more about an organization, even a secret one, that exists to serve them. We have to respond to that interest even as we are responsible to our statutory obligations to keep certain sensitive matters secret.
The important thing is for us to be reasonable and professional about what we protect. It does not take a genius to know what information requires a hard look: for example, in an age of terrorism and for privacy act considerations, we have to protect identities not already in the public domain. Also taboo--because they impact adversely on our ability to conduct our business, most of it necessarily in secret--are cover arrangements, liaison relationships, covert facilities, and unique collection and analytic capabilities. These constitute the sources and methods that truly need protection. For the most part, they can easily be avoided without keeping an author from telling a story or restricting an author's opinion on a variety of intelligence subjects.
In prepublication reviews, we have to show we know the difference between what truly is sensitive and what is not. We do not earn respect just by saying "no," but neither do we earn respect just by giving away information. Our unique role is to judge whether a denial of disclosure would stand up in court--whether we could make a compelling case in a court of law that specific damage to US national security would result. We can have it both ways: we can protect that which needs to be protected, while being forthcoming about intelligence activities in a way that can help educate, inform, enlighten, and even entertain the general public. That is the cost of doing business in this free society we help to preserve; trying to have it both ways is a challenge that comes with the territory.
John Hollister Hedley is Chairman of CIA's Publications Review Board.
One appeal-of 614 reviews from the fall of 1995 through the fall of 1997, when this article was written-was dropped. It became irrelevant when CIA no longer asserted classification for the Gulf War documents in question. In another appeal, the Board's decision was upheld; in a third, the Board's decision was partially upheld and partially reversed in the author's favor. Appeals of PRB determinations go to the Agency's Executive Director via the General Counsel, who assesses the legal sufficiency of the Board's action. If not satisfied by this administrative process, the author's recourse is to litigate.
CIA Secrecy Agreement, Form 368.
The US Supreme Court decision can be found in US v. Snepp, 444 U.S. 509, N.3 (1980).
The 30-day time constraint was set forth by the circuit court decision in US v. Marchetti, 466 F2d 1309, 1317 (4th Cir. 1972). It was reiterated in US v. Snepp, 595 F2d. 934 (4th Cir. 1979), and it has been adopted as the standard by the Department of Justice.
US v. Snepp, 444 U.S. 509, N.3 (1980). Again, the secrecy agreement is an enforceable contract requiring the author to submit material for review; failure to comply is a violation of the contract, regardless of whether any deletions ultimately would be required.
McGehee v. Casey, 718 F.2d 1141 (D.C. Cir. 1983).
Initially issued 27 September 1979 as Headquarters Regulation 6-2, revised 14 March 1995 and currently in effect as Agency Regulation (AR) 6-2.
Afshar v. Dept. of State, 702 F.2d 1130-35 (D.C. Cir. 1983).