Acerca del autor

Thomas Fingar is the Oksenberg/Rohlen Distinguished Fellow in the Freeman Spogli Institute for International Studies at Stanford University. From May 2005 through December 2008 he served as the first Deputy Director of National Intelligence for Analysis and, concurrently, as Chairman of the National Intelligence Council. He served previously as Assistant Secretary of the State Department's Bureau of Intelligence and Research, Principal Deputy Assistant Secretary (2001-2003), Deputy Assistant Secretary for Analysis (1994-2000), Director of the Office of Analysis for East Asia and the Pacific (1989-1994), and Chief of the China Division (1986-1989).

Fragmento. © Reproducción autorizada. Todos los derechos reservados.

Reducing Uncertainty

Stanford University Press

Contents

Acknowledgments...........................................................................ix1 Reducing Uncertainty....................................................................12 Myths, Fears, and Expectations..........................................................193 Spies Collect Data, Analysts Provide Insight............................................334 Using Intelligence to Anticipate Opportunities and Shape the Future.....................505 Estimative Analysis: What It Is, What It Isn't, and How to Read It......................676 A Tale of Two Estimates.................................................................897 Epilogue: Lessons and Challenges........................................................126Notes.....................................................................................141Index.....................................................................................171

Chapter One

the U.S. government spends billions of dollars every year to reduce uncertainty. The National Weather Service spends more than $1 billion a year to forecast precipitation amounts, track storms, and predict the weather. The Centers for Disease Control spend more than $6 billion to detect and investigate health problems in the United States and abroad. The Departments of Agriculture and Energy track and predict production of crops and various types of energy. Virtually every agency of the federal government monitors and forecasts a wide range of developments because farmers, manufacturers, state governments, travelers, and citizens in every walk of life want information that will enable them to make better-informed decisions about what to grow, whether to invest, and where to travel. In other words, we spend a lot of money to anticipate problems, identify opportunities, and avoid mistakes.

A substantial portion of what we spend to reduce uncertainty—almost $50 billion a year—goes to the U.S. Intelligence Community (IC). The need for this amount of money is justified through a process that emphasizes threats to our nation, our interests, and our people. For example, the classified and unclassified versions of the Annual Threat Assessment submitted to Congress by the director of national intelligence devote far more attention to problems and perils than to opportunities to shape events. This emphasis is understandable, but it is also unfortunate because it obscures one of the most important functions of the Intelligence Community and causes both analysts and agencies to devote too little attention to potential opportunities to move developments in a more favorable direction.

Intelligence Community work to reduce uncertainty differs from that of other U.S. government (USG) agencies in a number of important ways. The most obvious, of course, is that it has access to clandestinely acquired and other classified information. Indeed, the IC exists, in part, to ferret out secrets and to collect information that cannot be obtained by scholars, journalists, bankers, diplomats, or other "collectors." The use of classified information does not automatically make analyses produced by the Intelligence Community better than those produced using only unclassified information, but reducing uncertainty about a large number of traditional national security issues can be more precise and often more reliable if IC collectors have managed to intercept, inveigle, purchase, or steal information that others wish to keep hidden.

Another way in which IC efforts to reduce uncertainty differ from those of other USG agencies is that most IC products remain classified for a long time. The reasons they do include the need to protect sources and methods (that is, how we obtained the information), the value attached to objectivity and clarity (it can be difficult to obtain a hard-nosed assessment of a foreign political leader or military capability if the analysis is intended for public dissemination), and the fact that many IC assessments are used to determine policies and negotiating positions of the U.S. government (telling others what U.S. officials are considering before they have made a decision on the issue is generally thought to be unhelpful). Despite recurring assertions to the contrary, IC analytic products are not classified for long periods of time to prevent taxpayers from knowing what they get for $50 billion a year. Classification decisions are made when assessments are produced, almost always on the basis of the classification level of information used in the report. I cannot imagine a situation in which an analyst or manager would say, in effect, "This analysis is really bad. Let's disseminate it to decision makers but give it a very restrictive security classification so the public will not see how bad it is." In my experience, most intelligence assessments are not very sexy or exciting, and few present theoretical breakthroughs of any kind. But that does not mean that they are poorly crafted or of little value. As noted in Chapter 5, hundreds of National Intelligence Estimates have been declassified and released. Most of them stand up pretty well, especially if judged against the criteria of "Were they useful to decision makers at the time they were produced?" and "Did they help to reduce uncertainty—even if only by reaffirming what officials thought they understood to be the case—about issues being deliberated in the USG?"

Reducing uncertainty sometimes involves acquiring information, overtly and covertly, that is thought to be useful to understanding developments or intentions. Sometimes it involves research and analysis to produce what Carl Ford refers to as "new knowledge," that is, better understanding and new insights derived from thinking hard about information we possess but have not considered or combined in the way that led to the new assessment. At still other times it involves efforts to substantiate or disconfirm a hunch articulated by a customer or fellow analyst or to refute statements made in a meeting or the media. Reducing uncertainty, in other words, can take many forms and involve many types of analysis, but it almost always strives to enhance understanding of what is known, what remains unknown, what is happening, where events seem to be headed, what is driving them, and what might alter the trajectory of developments.

Contrary to fictional depictions and popular misconceptions fueled by political grandstanding and media caricatures, the intelligence enterprise exists to do more than steal secrets and "connect the dots." Ferreting out information that adversaries wish to hide and discovering (and disrupting) terrorist plots and other threats to our nation and our interests are important missions and, disparaging characterizations notwithstanding, we perform them very well—most of the time. We never have batted 1.000 and never will, but getting it mostly right, most of the time, in time to shape, prevent, or prepare for developments with the potential to affect our nation, our citizens, and our interests is and will remain our most important goal and criterion of success. Unlike highly paid ballplayers, we will never be satisfied with a .300 batting average. The security of our nation requires that we come as close to 1.000 as is humanly possible, and every IC analyst worthy of the title is determined to do so, both individually and collectively.

For reasons of patriotism, professionalism, and personal pride, Intelligence Community analysts aspire to meet exceptionally high performance standards. High standards, and high expectations, are intrinsic to the profession because how well or badly analysts perform can have real-world consequences. As I have told thousands of new recruits, they must always be cognizant of the facts that U.S. government officials will be influenced as well as informed by what they say and write and that the efficacy of U.S. policies and actions will be determined, in part, by the analysis they produce. For those who serve in the U.S. Intelligence Community, the often-derisive phrase "good enough for government work" has a very different meaning than it does in conventional usage. In the Intelligence Community, "good enough for government work" means not simply that it must be as accurate as possible but also that it specifies clearly what is and is not known about the issue, the quantity and quality of available information, what assumptions have been used to bridge intelligence gaps, what alternatives have been considered, and how much confidence analysts have in the evidence and their judgments. It also requires addressing the right questions at the right time and ensuring that information and insights are delivered to all the right people.

The first list of requirements summarized in the previous paragraph involves tradecraft issues. In most respects, the requisites for good intelligence analysis are identical to the requirements for good academic analysis and good analysis of all other kinds, and IC analysts can—and must—rely heavily on the analytic methods they learned in graduate school. There are differences between academe and the world of intelligence (for example, deliberate efforts to hide information and to deceive or mislead foreign governments are much more common in the work of the Intelligence Community than they are in academic research), but the differences should tip the balance in the direction of enforcing even higher standards for IC analysis than for peer-reviewed academic papers. In other words, I have no sympathy for arguments that one must give IC analysts a break because they operate under conditions and constraints that generally are not found in university or think tank research settings. To the contrary, standards of performance for the Intelligence Community can be no lower and arguably must be higher than those in academe for the obvious reason that the potential impacts of IC analysis are far more consequential. An academic who does bad or sloppy work will be chastised by peers and perhaps denied tenure, but faulty intelligence analysis has the potential to redirect U.S. foreign or security policies, discredit or endorse the positions of foreign leaders or governments, raise doubts about the loyalty of citizens or corporate actors, or cause the United States to undertake unwarranted or counterproductive military actions.

The second set of requirements previously noted—right questions, right time, and right people—is more germane to intelligence analysis than it is to academic research but has both similarities and analogs for certain types of think tank research and for consulting firms producing analysis for particular clients. There are important differences, but the IC, think tanks, and consultants all produce "targeted" assessments timed to inform—or influence—particular decisions. All three produce other types of analysis as well, but my focus here is on products that have been requested by a particular customer or are produced at the initiative of an analyst or manager who knows the issue and the target audience well enough to judge that examining a particular question or set of questions at a particular point in time would be helpful to the targeted decision maker.

The line between analysis produced to inform and analysis produced to influence can be very vague and may exist mainly in the eye of the beholder, and many argue that intelligence analysts must stay far back from that line lest they be guilty or suspected of policy advocacy. I certainly agree that intelligence analysts must not be advocates for policy and that they must be—and be seen to be—as objective as possible. That, as well as having access to classified information that is not available to most analysts outside the Intelligence Community, is what distinguishes them from the many other individuals and organizations pushing information and ideas to U.S. policy makers.

Intelligence analysis is exacting work, but the intellectual and psychic rewards can be substantial. The challenge of unraveling a mystery, solving a geopolitical puzzle, or discovering previously unknown or unappreciated dimensions of situations with the potential to affect the security of our nation and the efficacy of its policies requires rigor, dedication, and flashes of inspiration. For an analyst, the process itself is enjoyable; producing an assessment that enhances understanding and assists those we support is very gratifying. Nevertheless, most of the time analytic achievements earn mainly—or only—psychic rewards and commendation from peers and immediate supervisors. Pats on the head or "attaboys" from those we support are much rarer than they should be. That is unfortunate but probably inevitable because the nature of intelligence support makes it an iterative process in which continuous interchange between decision makers and analysts clarifies issues, integrates new information into preexisting intellectual frameworks, and produces more incremental increases in understanding than "eureka" moments of discovery. The analyst plays a critical role in this process by providing new information and new insights, but the payoff comes when the decision maker takes ownership of the new idea. Having made it their own, customers seldom want to share credit unless they still have only low confidence in the idea or insight. Intelligence analysts must remember that the goal is not to make themselves smarter but to make those whom they support more knowledgeable about the issues they are working on in the hope that policy makers will make better decisions. In my experience, customers are more likely to show their appreciation for analytic assistance by asking more questions and granting greater access to deliberative meetings than they are to give analysts credit for providing an insight that triggered or facilitated new policy recommendations.

Solving analytic puzzles is more difficult than providing valuable input to those we support. Much of the time, we can be useful without being brilliant. As Carl Ford regularly reminded our analysts when he was assistant secretary of state for intelligence and research, on any matter of importance there are probably dozens of things that policy makers think they should know about the issue. Of the dozens of things they think it important to understand, they think they know only a few and actually understand fewer than that. This creates an enormous opportunity for analysts to provide valuable input. If they can generate new information and/or new insight on just one or a few of the things policy makers would like to know, they have increased understanding and reduced uncertainty. They may also have given the customer they support a bureaucratic advantage because if he or she understands the issue better than others working the problem, or at least is able to convince them that he or she understands it better, the likelihood of gaining deference or support for that customer's preferred course of action goes up.

Policy makers appreciate it when analysts give them information and insight that they can use to advantage as well as to deepen their own understanding, and they often reward utility with greater confidence and better access. That is natural and helps analysts to become even better informed about the intelligence and analytic needs of those they support. Such a relationship should not-must not-lead to situations in which analysts provide new information and insight only to their primary customer. Analysts need the confidence and trust of those they support if they are to do their jobs effectively, but an important part of their jobs is to ensure that the insights they produce are provided to "everyone" working the issue and, an even more difficult challenge, to all who might find them useful. This means sharing their insights with analyst counterparts who support other policy makers, publishing them in widely disseminated publications, posting them on appropriate websites, and writing them at a level of classification that makes them easily accessible to all who might find them useful.

(Continues...)

Excerpted from Reducing Uncertaintyby Thomas Fingar Copyright © 2011 by Board of Trustees of the Leland Stanford Junior University. Excerpted by permission of Stanford University Press. All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.