Cost Accounting Approach.
An analyst employing this strategy focuses on the cost incurred by the adversary to conduct a deception. If the cost is higher than any potential gain, the deception hypothesis is rejected. Most frequently considered is the cost of providing valid and valuable information to establish and maintain the credibility of a deception channel. This approach judges a source on the basis of it’s production. This is the strategy generally used by the Directorate of Operations (DO) reports officer, as distinct from the DO operations officer, who is more likely to focus on anomalies or inconsistencies in the circumstances under which the source obtained the information or passed it on to us.
The cost accounting approach assumes an accurate understanding of how much an adversary is willing to sacrifice to achieve his goals. This is difficult because, in principle, there always can be a goal that is important enough to justify any cost. Such analysis, therefore, may be correct in routine cases that follow the established pattern, but wrong in the rare, exceptionally high-stakes situation when it counts the most. Traditional theory holds that deception is most likely when the stakes are highest; for example, when an adversary would be willing to pay the greatest price to achieve his goals.
The three previous approaches, motive, anomalies and inconsistencies, and litmus test, are best used to show the presence of deception. One cannot do that with the cost accounting strategy. This strategy can be used only to prove the absence of deception, not its presence. This is because high cost may (or may not) imply the absence of deception, but low cost does not necessarily imply its presence. Cost accounting was the primary strategy employed by those who believed Nosenko was a bona fide defector. Briefly, it was argued that information provided by Nosenko was so valuable to CIA and so damaging to the USSR that it was inconceivable the KGB would willingly give it away. The KGB would never deliberately place one of its own officers into the hands of its principal adversary.
Bruce Solie, the Office of Security officer who handled Nosenko's rehabilitation, believed a source should be judged on the quality of information received. He felt that too much attention had been paid to challenging Nosenko’s story, and not enough to finding out everything he could tell us about KGB operations. Accordingly, Solie's debriefing concentrated on obtaining more information. Since the value of Nosenko's information created a presumption of bona fides, Solie's approach to anomalies and inconsistencies was to seek some plausible explanation and to recognize that, given the nature of the case, some anomalies may never be explained. Subsequently, the FBI reported that a minimum of nine new counterintelligence cases were developed as a result of Solie's reexamination, and that new information of considerable importance was developed on old cases; additional detailed information was also obtained on KGB operations in other countries, and on KGB organization, modus operandi, and personnel. Altogether, the information pointed to a valuable, bona fide source.
Nosenko provided identification of, or leads to, some 238 Americans and about 200 foreign nationals in whom the KGB had displayed varying degrees of interest, and against whom they had enjoyed varying degrees of success. He provided information on about 2,000 KGB staff officers and 300 Soviet national agents or contacts of the KGB. His information on the methods and scope of Second Chief Directorate operations against foreign diplomats and journalists in Moscow and visitors to the Soviet Union filled a large gap in our knowledge and had an enormous impact on the raising of CIA's consciousness of these operations; the result was important improvements in the physical security of U.S. installations and the personal security of U.S. officials and advisors to the USSR. The Soviet Union suffered additional costs through the adverse publicity and deterrent effect of Nosenko's defection and the arrest of several agents he identified.
The master plot theorists also assessed the value of Nosenko's information, not as the overriding criterion for evaluating his bona fides, but certainly as a relevant consideration. They came to the opposite conclusion. They argued that KGB operations reported by Nosenko had been previously reported by Golitsyn or other sources, involved agents who had lost their access, who the KGB had reason to believe had been compromised, who were unimportant, or where the information could not be confirmed. At times, this argument was made with a stacked deck, and Nosenko always came out the loser. If Nosenko reported on a KGB operation previously reported by an- other defector, this was not counted as a plus for Nosenko, because the KGB already knew the operation had been compromised. If he reported an operation not previously known, the information was judged doubtful because it could not be confirmed.The argument that Nosenko's information was unimportant was always difficult to sustain, as it kept being contradicted by reality. For a time, various rationalizations were possible, but eventually Solie's supplemental debriefing of Nosenko, a more objective interpretation of his information, and successful investigation of the leads developed a clear record to Nosenko's credit. For example, Bagley argued that Nosenko's information on microphones in the American Embassy in Moscow was a KGB giveaway, as Golitsyn had already reported on this subject. Golitsyn's report, however, was only a general statement that the Embassy was bugged. Nosenko identified which offices, where in these offices the mikes were located, and what conversations had been monitored. It was only Nosenko's more detailed, first-hand information that persuaded the State Department to incur the significant cost and disruption involved in tearing out walls to find and neutralize the installation. They found 52 mikes, 47 of them still active, covering many of the most interesting offices in the Embassy.
Nosenko's supporters and detractors largely ignored each other's arguments, as they approached the analysis of Nosenko's bona fides from entirely different paths and used different criteria for judging the truth. In the context of the master plot, the value of information provided by Nosenko measured only the magnitude of the deception to come. To believers in Nosenko, on the other hand, the list of espionage cases the FBI had developed from his information and the many leads to KGB agents in other countries were ample proof of his bona fides. As the KGB would not willingly sacrifice this information, there must be some other explanation for the conflicts in his story. Ironically, one of the strongest points in favor of Nosenko was not made in any of the studies of his bona fides. This is because of the tacit agreement by Nosenko's detractors and supporters alike that he should be considered on his own merits, without reference to the master plot as a whole. This was an artificial limitation, as almost everyone on both sides of the issue recognized that, if Nosenko were bona fide, the master plot theory would collapse. Conversely, if the master plot theory could be disproved, the case against Nosenko would be difficult to sustain.
The point is that Nosenko's information was only a small part of the total information on Soviet intelligence being received at that time from multiple sources believed to be part of the master plot. If all the valuable information from all the supposedly controlled sources had been collected in one massive compendium to show the incredible magnitude of what was allegedly being sacrificed, it would have made a most powerful argument against the master plot theory and, therefore, against the argument that Nosenko was under control.
To the few who were informed of all these other cases and saw events from the master plot perspective, it began to appear as though the entire KGB might have become a sacrificial lamb on the altar of deception and diversion. The diehards (some would call them paranoids) rationalized this by concluding that the deception must be even bigger and more important than previously believed; as the KGB sacrifices mounted, the level and importance of the assumed penetration went higher and higher. Others, however, who initially had accepted the inherent plausibility of the master plot, began to have doubts. There had to be some limit on how much the KGB would give away.
A former Chief of the CI Group in SB Division told me, for example, that he first began to doubt the master plot theory when he learned about a senior East European military officer who had been supplying CIA with very important Warsaw Pact documents for many years. Giving away counterintelligence leads to divert opposition security services was one thing, but the wholesale sacrifice of military secrets was quite another. Continuation of this operation seemed to belie the existence of a well-placed KGB penetration of CIA, which was an essential part of the master plot theory.
In 1976, John Hart visited Tennant Bagley, who had retired four years earlier, to inquire whether, after so much time, he still rejected Nosenko's bona fides. Bagley asked if there had been any new evidence confirming Nosenko's story, to which Hart replied there had not. I believe that Hart was wrong; there was much additional evidence to refute the master plot theory. During the intervening years, old sources continued and new counterintelligence sources were developed that produced a continuous flow of very valuable information, thereby greatly increasing the total cost to the Soviets if the master plan theory were valid. The most knowledgeable new sources would have failed the master plot litmus test, for they confirmed genuine KGB distress at Nosenko's defection, reported that the KGB would be unwilling to use a KGB staff officer as a double agent or false defector, and failed to report any massive counterintelligence deception program or penetration of CIA at the level that had been feared.
In summary, this cost accounting strategy led to an accurate judgment of Nosenko's bona fides. Our discussion does, however, illustrate the principal weakness of this approach, namely, how to determine what the adversary might regard as an acceptable cost. Analysts on both sides of the issue recognized that a source's production is a useful criterion for judging bona fides, but they had radically different views on how much information and the kinds of information the KGB would be willing to sacrifice in the interest of deception.
Predictive Test Approach.
This approach is not normally used in intelligence analysis, but it could and should be used under select circumstances. It might also be called the scientific approach, for it addresses a problem the way a scientist does. The scientist develops a tentative hypothesis to explain the phenomenon under study, then devises tests to prove or disprove the hypothesis. The same approach can be used in deception analysis. A tentative hypothesis about deception can often be tested by using the presumption of deception to make a series of predictions. If the deception theory generates predictions that are borne out by subsequent experience, this suggests that the hypothesis is true. If expectations are contradicted by experience, this suggests that the hypothesis may be wrong.
The principal value of testing views in this manner is that it makes it more difficult to rationalize contradictory evidence. There is a human tendency to interpret new information in ways that do not require us to change our minds. By using a hypothesis about deception to make explicit predictions, one specifies the circumstances under which one could be proved wrong. The predictions are either confirmed or they are not. The latter outcome points to possible flaws in one's understanding of the subject and provokes thought as to where reasoning went astray. If one has not made explicit predictions, one may not concede or even recognize that the reasoning was ill founded.
The weakness of this strategy is that it presupposes some form of ongoing activity, so that one can make short-term predictions and see if they come true. Another weakness is that it requires more self-conscious introspection and willingness to question one's own assumptions than most intelligence analysts are comfortable with, and therefore is seldom used.
This strategy was not used explicitly in any of the analyses of the Nosenko case. This form of reasoning did, however, implicitly affect the thinking of some persons involved in the case. The master plot theory did lead to certain expectations, or implied predictions, and when they were not borne out in practice this did engender doubts about the theory. The most obvious expectation was that it might be possible to break Nosenko. Of course, one could never be certain Nosenko would break even if he were under KGB control, but the interrogators had important psychological advantages. They had total control over Nosenko, were "certain" he was a KGB-dispatched agent, and even "knew" the purpose of the KGB operations.Yet Nosenko did not confess to anything more than a few self-serving lies. Similar advantages existed in the CIA interrogation of Loginov in South Africa; again, the interrogation was unsuccessful.
The master plot theory assumed the existence of a well-placed KGB penetration of CIA. It predicted that KGB agents, such as Nosenko, were either being directed with guidance from the penetration, or were being used to build up or protect the penetration. This limited the suspects to a finite number of CIA officers. It was estimated that astute counterintelligence analysis and security investigation would identify the penetration, but no such penetration was discovered despite extensive investigation. Again, the theory was contradicted by events. This might logically have led the proponents of the theory to question their assumptions, but it did not. Additional expectations concerned the KGB operations on which Nosenko reported. One would expect the KGB to try to achieve its deception goals with the minimum necessary cost to itself. This led to the view that Nosenko's revelations about KGB operations were probably less important than they appeared to be. Specifically, if was anticipated that investigation of Nosenko's leads to seemingly well-placed KGB agents would show that these agents had recently lost their access or were, for some other reason, expendable to the KGB. Further, it was expected that investigation of tantalizing leads to agents who could not be fully identified by Nosenko would lead to dead ends, so the KGB would really have sacrificed nothing at all. In some cases, these expectations were accurate, but in a significant number of cases investigations did identify valuable KGB agents.
These beliefs and expectations were, in effect, logical deductions or predictions based on the master plot hypothesis.If they had been set forth explicitly as predictions, and recognized as valid if partial and imperfect tests of the hypothesis, it would have been far more difficult to ignore the implications of their being contradicted by events. Unfortunately, the advocates of the master plot theory were seeking to prove it, not to test it. They regarded the master plot as a fact, not as a hypothesis to be subjected to critical examination, so developments that seemed to contradict this view were ignored, rationalized, or misinterpreted.
III: Conclusions
Heretofore, I have tried to present an objective account of the diverse arguments used, or in some cases not used, by those on both sides of the Nosenko issue. What follows is personal opinion. The reader is cautioned that opinions on this case are as numerous and varied as the many CIA and FBI officers who were personally involved in one or another phase of it. The opinions expressed here are certainly not the final word.
I will start by making my personal bias clear. I became a believer in the master plot theory in 1965 when first exposed to the reasoning described above under the motive approach. Although initially a believer, I never put much stock in Bagley's Thousand Pager, as I had learned from experience to be skeptical of conclusions based on the anomalies and inconsistencies approach to counterintelligence analysis. My first doubts arose when, one by one, various expectations failed to materialize, which is the reasoning described above under the predictive test approach. Subsequently, for reasons discussed under the cost accounting approach-the high volume of significant intelligence being received through multiple sources-I rejected the master plot theory and concluded that Nosenko was not acting under KGB control. This conclusion was recently reinforced when, in doing research for this study, I learned how the many anomalies and contradictions were eventually explained.
I remain firmly opposed to the view that the master plot was an irresponsible, paranoid fantasy. Given the information available at the time, as described under the motive approach, it would have been irresponsible not to have seriously considered this possibility. The mistake was not in pursuing the master plot theory, but in getting so locked into a position that one was unable to question basic assumptions or to note the gradual accumulation of contrary evidence. This type of analytical error is not uncommon, and all of us are susceptible to it. It can lead one to overlook deception as well as to perceive deception when it isn't there, as happened in the Nosenko case.
Gordon Stewart was correct in criticizing Bagley's analysis for assuming guilt and then interpreting every discrepancy as evidence of this guilt. After serious doubts began to be raised about this analysis, the Soviet Bloc Division should have assigned an officer in a devil's advocate role to do an analysis that assumed innocence, then examined how the evidence could be interpreted as being consistent with this view. By failing to give a fair shake to the opposing view, SB Division lost control of the case to another component that approached the analysis from a totally different perspective. This lesson should be taken to heart by any component involved in a serious controversy over deception .
Most important for this study is not what we learn about the Nosenko case, but what can be learned about deception analysis in general. There seem to be five quite separate and distinct paths to reaching judgments about deception. Which path one chooses is strongly influenced by one's past experience and the patterns of thought associated with one's functional responsibilities or organizational affiliation. The path one take determines, in turn, the evidence one seeks and, in large measure, the conclusion one reaches"
Generally, analysts in the Nosenko case gave greatest weight to the islands of evidence they were most familiar and most comfortable with, and this in turn determined which approach they took. Some counterintelligence personnel were very familiar with and influenced by the concern about penetration of CIA. Speaking for myself, for example, I had been the Headquarters desk officer on the Goleniewsld case at the time we received both of his reports, the one on the penetration of CIA and the one that led to identification of Felfe as a KGB penetration of West German intelligence. I had also seen first hand how much damage a well-placed penetration, such as Goleniewsld, could do to an opposition service. This personal experience made the master plot seem very real and very plausible. Other counterintelligence specialists, more directly involved in the Nosenko case than I, had personal experience with the seeming contradictions and extraordinary coincidences in his story, so this is what most influenced them.
One's professional experience and organizational affiliation play important roles in determining the strategy one employs to analyze deception. Counterintelligence officers, DO reports officers, and DI intelligence analysts look at deception through different conceptual lenses. The strategy employed, in turn, largely determines what evidence one seeks and the conclusion one reaches. Examination of Soviet motives led Bagley to review the entire history of CIA and KGB counterintelligence operations. Anomalies and inconsistencies perceived by counterintelligence personnel in the initial debriefing of Nosenko led to hostile interrogation to develop still more anomalies and inconsistencies. Solie's cost accounting approach led to friendly interrogation designed to produce as much information as possible, while playing down the anomalies. The litmus test approach focused attention on how other sources' information related to the Nosenko case, while overlooking the total value of these sources' information.
The Nosenko case vividly illustrates the weakness inherent in each of the five strategies for analyzing deception. Bagley identified a plausible motive for Soviet deception and supported it with voluminous circumstantial evidence, yet Nosenko was not under Soviet control. There was an enormous number of anomalies and inconsistencies in Nosenko's story, yet they were all produced by sloppy translation and inadequate debriefing, the unique aspects of Nosenko’s background and personality, genuine accident and coincidence, and the circumstances of his handling; they were not truly indicative of hostile control.
Comparing the value of information received from a source against some criterion of the cost the adversary would be willing to incur also leads to strong inferences about deception, but only if the putative threshold of acceptable cost is correct; in the Nosenko analysis, that threshold was itself a major point of disagreement. Comparing suspect information against some objective criterion of truth can lead to strong inferences about deception, but only If the criterion does indeed represent the truth; in the Nosenko case, it did not. Making test predictions can prompt reconsideration of one's views if they fail the test of experience, but, typically, this was not done in the Nosenko case; analysts were too busy trying to prove they were right, rather than testing their assumptions.
If there is a single lesson to be learned from this, it is that all five approaches are useful for complete analysis. Exclusive reliance on anyone strategy is dangerous. The cost accounting approach led to a correct conclusion in the Nosenko case, but it, too, has inherent vulnerabilities, and there is no guarantee it will be correct in all future cases. This places a heavy burden on the deception analyst. To do a thorough analysis of all the possibilities, one must examine the diverse bodies of evidence relevant to each of these strategies. This will often require investment of substantial time to gain understanding of previously unfamiliar fields.