SURVEYING INTELLIGENCE CONSUMERS

• Collection: FOIA Collection
• Collection: Declassified Articles from Studies in Intelligence: The IC’s Journal for the Intelligence Professional
• Document Number (FOIA) /ESDN (CREST): 0000621334
• Document Page Count: 13
• Document Release Date: July 30, 2014
• Case Number: F-2013-02322
• Publication Date: September 1, 1989

TITLE:AUTHOR:VOLUME:Approved for Release: 2014/07/29 000621334Surveying Intelligence Consumers(b)(3)(c)33 ISSUE: Fall YEAR: 1989Approved for Release: 2014/07/29 000621334pproved for Release: 2014/07/29 000621334STUDIES ININTELLIGENCEA collection of articles on the historical, operational, doctrinal, and theoretical aspects of intelligence.All statements of fact, opinion or analysis expressed in Studies in Intelligence are those ofthe authors. They do not necessarily reflect official positions or views of the CentralIntelligence Agency or any other US Government entity, past or present. Nothing in thecontents should be construed as asserting or implying US Government endorsement of anarticle's factual statements and interpretations.Approved for Release: 2014/07/29 000621334Approved for Release: 2014/07/29 000621334Lessons for producers44FtEtEr11ALSURVEYING INTELLIGENCE CONSUMERS(b)(3)(c)The recently published Intelligence Consumer Survey, 1988 provides important, newinformation on the views of senior consumers toward intelligence. The 234 officials of theReagan administration who responded to the survey were broadly representative of consumersat the deputy assistant secretary and above levels at the National Security Council, State,Defense, and at other departments chiefly concerned with international economic issues.Respondents included Vernon Walters and Elliott Abrams at State, Donald Gregg and RobertOakley at the White House, Caspar Weinberger and Fred Ikle at Defense, and Alan Greenspanat the Federal Reserve. The attitudes of such officials toward intelligence can be compared withthose expressed by 133 former Carter administration officials of equivalent rank in a 1980survey also undertaken by the Intelligence Producers' Council.'There are at least four reasons to pay attention to the survey data.? The responses of the officials surveyed, both as a group and along agency lines, arecredible because they are consistent with much of our anecdotal or case-specificknowledge of senior consumer attitudes towards intelligence. As a result, the surveyprovides a much broader, yet still believable baseline of knowledge about what seniorconsumers think of intelligence than do the random set of anecdotes on which most ofus normally rely.? Analysis of the data indicates that officials from all agencies provide very similarresponses on most questions. Where differences do appear, the 1988 data can be brokenout into the agency groupings indicated in Table 1 for closer analysis.? The 1980 and 1988 results are very consistent with each other. This indicates that thesurveys have tapped policymaker concerns and satisfactions about intelligence thatpersist over time and political admininistrations.? The 1988 survey indicates areas where intelligence producers need to improve theservices provided to their principal consumers. If consumers were giving intelligenceproducers a letter grade, a "gentleman's B- would come to mind after perusing thesurvey data. This suggests room for improvement, and in some areas more seriousdeficiencies appear to exist.Table 1The 1980 and 1988 Survey Populations1980 Survey1988 SurveyNumber ofPercentNumber ofPercentAgency GroupingRespondentsof SampleRespondentsof SampleDept. of Defense4232%11348%Dept. of State3526%5825%White House1914%198%Economic Agencies3728%3515%Other Agencies0?94%Totals133100%234100%I See Intelligence Consumer Survey (IPC 82-10002, September 1982, Confidential).Approved for Release: 2014/07/29 000621334TIALApproved for Release: 2014/07/29 000621334This article focuses on what I believe are the four key findings of the 1988 survey and,where possible, compares them with data from the 1980 survey to indicate any trends overtime. The implications of all of these findings indicate that intelligence producers should makechanges in the way they do business, if they are to serve policymakers effectively in the 1990s.Key Finding #1: All Forms of Intelligence Are Equally UsefulBoth surveys asked consumers to indicate the usefulness of five basic kinds of information:nonintelligence materials, such as newspapers or academic research; basic intelligence, such asbiographic, order of battle or other factual data; current intelligence, such as the NationalIntelligence Daily (NID) or the State Department's Morning Summary; in-depth analyticstudies, such as CIA Intelligence Assessments or National Intelligence Estimates (NIEs); andraw or uninterpreted reporting, such as State cables or attache reports. Unfortunately, the twosurveys gave respondents a different range of possible answers, with the 1980 survey offeringmore negative choices and the 1.988 survey offering more positive choices on the issue of utility.This has naturally skewed each survey's results in the indicated direction. To minimize thisbias, the responses from each survey have been reorganized into three basic categories:useful," "not very useful," and "not used.- 2 Table 2 summarizes consumer responses alongthese lines.Table 2Percent of Policymakers IndicatingUsefulness of Different Types of Information1981 Survey 1988 SurveyType ofInformation UsefulNot VeryNot VeryUseful Not Used Useful Useful Not UsedCurrent Intel 84% 8%In-depth Intel 72% 20%8% 78% 12% 10%8% 74% 16% 10%Non-intel Info 71%Basic Intel 75%Raw Reports 66%15%11%22%14%14%12%73%70%66%18%21%20%9%9%14%% Spending at LeastAn Hourly/Wkly67%39%63%38%50%The responses are notable both for the high proportion of senior officials who cite eachtype of information as useful and for the consistency of the figures across the two surveys.Two-thirds or more of our consumers find each type of information useful. The small samplesize in both surveys makes differences of less than 10 percent of negligible significance. Thismeans, for example, that nonintelligence information was about as useful as any form ofintelligence for Reagan administration officials. It also means that the comparative utility2 The following chart indicates how the responses available on the two surveys were reorganized into thecategories used in this study.New Category1981 Responses 1988 ResponsesUseful Very + Fairly Useful = Essential + Very Useful +Somewhat Useful/2Not Very Not Particularly Useful = Somewhat Useful/2 + Of No UseUseful + Not UsefulNot Used Did Not Receive + Have No Experience + No AnswerNo AnswerAnalysis of each survey using its original range of responses produces the same findings, but it does not allowcomparisons over time.Approved for Release: 2014/07/29 000621334Approved for Release: 2014/07/29 000621334--GeftrtDEFITrALadvantage held by current intelligence over all other information sources in 1980 virtuallydisappears in 1988; only raw reporting remains less useful than current intelligence for seniorofficials.These aggregate survey results hold up across three of the four agency groupings in the1988 survey. Consumers at Defense, State and the other, mostly economic agencies display nostrong preference for one form of information over another. However, nearly 90 percent of theofficials from the Reagan White House cited current and in-depth intelligence as useful, ahigher proportion than at the other agencies. NSC officials, with their broadranging and rapidlychanging focus of responsibilities, probably have a greater need for these finished forms ofintelligence. Departmental officials are somewhat more specialized and experienced in theirrespective areas.As the last column in Table 2 reveals, there are sharp differences in the amount of timesenior officials spend on various types of information. The survey results, however, allow us toestimate that senior officials spend between two and eight hours a week reviewing intelligenceof any form.There are apparent inconsistencies between how much time consumers say they spend onvarious types of information and how useful they say each type is. Fewer consumers spend atleast an hour a week on in-depth intelligence (39 percent) than on raw reports (50 percent) andnonintelligence sources (63 percent), even though in-depth studies are rated useful by at leastas many consumers. Current intelligence and nonintelligence sources receive significantly moretime than the other sources, but do not get significantly higher marks for their utility.This time-utility contradiction can be explained by two conditions confronting mostpolicymakers. First, incentives are high for senior officials to focus on immediate problems.They fear the consequences of not spending as much time as other players in the policy gamesearching for currently relevant material. The substantial amount of rapidly and widelydisseminated raw reporting that flows to senior officials puts an additional time burden on theconsumer's daily encounters with intelligence. Ironically, the surveys suggest that whenconsumers take the time to read more in-depth analyses, they find them equally useful. But thisdoes not reduce the policymaker's near obsession with keeping current.The second reason consumers spend as much time on nonintelligence sources as onintelligence material is, in my view, the crucial importance of the domestic and bureaucraticpolitical context of American foreign policy. Awareness of what issues and events are receivingpublic or Congressional attention and of what one's competitors in the policy game are doingis at least as important to senior officials as knowledge of foreign developments.The initial message for the Intelligence Community is gratifying: a lrge majority of seniorconsumers find all forms of intelligence useful in one way or another. A future survey wouldhelp producers even more if it asked for consumer feedback on what is more or less useful indifferent types of intelligence, when each type of intelligence is most useful, and what sort ofpackaging or presentation makes written intelligence more or less useful.The implication for producers of current and in-depth finished intelligence is lessencouraging. Such material is cited as no more useful than are newspapers and raw reporting.The basic message here should not be missed: the information marketplace is highlycompetitive, with consumers picking out whatever is most relevant and available at the timeneeded. The surveys indicate that finished intelligence does not have a built-in comparativeadvantage in these terms.Approved for Release: 2014/07/29 000621334Approved for Release: 2014/07/29 000621334There are three lessons for finished intelligence producers to draw from this finding. First,producers must place a premium on conciseness and on a format that highlights what isimportant. The current policy peg also should be identified and emphasized at the outset.Additionally, intelligence producers need to consider more frequently what the compar-ative advantage of finished intelligence is in any given situation. Many times we cannot "beat"our competitors to the consumer, even if we want to. Do policymakers benefit and does thecredibility of intelligence increase if a story containing no unique intelligence is published justto get it on the record? Alternatively, if producers decide they have something of value to offerthe consumer that is not already available from other sources, the survey indicates that simplydropping the product in the consumer's mailbox is not a strategy for success in a competitivemarketplace. Time and energy have to be devoted to marketing finished intelligence.Finally, newspapers and raw reporting may be as useful as finished intelligence becausetheir style and approach to interpreting events are often more in line with the way policyofficials think about or approach events. These sources tend to be conversational in style,anecdotal in their treatment of events, and often focused on personalities. Finished intelligencetends toward a denser style, seeks to highlight generalizations about events, and 'focuses onabstract forces, thereby reducing its potential utility.Key Finding #2: Personal Staffs Play a Pivotal RoleBoth surveys confirmed that most senior officials rely on their personal staffs to screen orsummarize the intelligence they receive. Over 80 percent of the Carter consumers said thattheir staffs screened some or all of their intelligence; nearly 65 percent of the Reagan consumerssaid their staffs summarized or otherwise edited some or all of the intelligence they saw. Thedrop in the proportion of officials acknowledging such staff interventions presumably reflectsthe narrower question wording in the 1988 survey and not a drop in policymaker reliance ongatekeepers. Such a drop would run contrary to much of our anecdotal evidence.Evidence of the increasing role of the personal staffs is found in the answers provided tosurvey questions on the channels used to request intelligence. Table 3 indicates the percentageof consumers who indicated they requested intelligence either -daily or weekly- or -not at all"via five possible tasking channels: personal staff, intelligence liaison officers, directly to otherintelligence officers, via formal requirements, or via the directors or deputy directors of anintelligence agency.Table 3Percent of Senior Officials IndicatingUse of Various Channels to Request Intelligence1981 Survey1988 Sur'veyDaily orNotDaily orNotDaily/Weekly Use atDoDStateVVHseOtherChannelWeeklyAt AllWeeklyAt AllPersonal Staff55%18%65%14%66%65%83%59%Via Intel Liaison Off.65%15%40%25%41%24%44%55%Directly to an Intel Off.44%15%21%28%21%12%56%18%Via Formal RequirementsNANA19%44%23%7%22%20%Via Agency Director/DD16%31%3%61%3%2%22%0%% Citing Any Intel Channel75%15%55%25%55%45%70%55%Note: Table excludes consumers who indicated they used a tasking channel less than weekly.Approved for Release: 2014/07/29 000621334Approved for Release: 2014/07/29 000621334The change between the two surveys in the proportion of policymakers requestingintelligence via each channel is striking. The proportion of consumers who used any of thedirect channels to the Intelligence Community on a daily or weekly basis declined by about 30percent. The proportion who indicated that they did not use a direct channel rose in everyinstance. In contrast, the percentage of consumers asking their personal staffs to obtain neededinformation rose between the two surveys. The bright spot in the data is the equal significanceof the intelligence liaison officer and personal staff channels at the other, mostly economicagencies. The minimal use made of intelligence agency directors, except by the White House,is not surprising; most consumers in the survey would not have frequent access to this high-levelchannel. The limited use made of formal requirements systems confirms longstanding concernsthat consumers are largely ignorant of how to use such systems and, when knowledgeable, findthem confusing, rigid, and often too slow to meet their needs.The evidence of a decline in the frequency of direct tasking of the IntelligenceCommunity is troubling. Personal staffs may usually come to Community components to fulfillthe needs of their principals, but they may also be filling those needs themselves by taskingcomponents in the policy agency itself. Interviews with policy officials or other research areneeded to confirm exactly what has happened and why.Whatever the case, the prominent role of senior staffs has two significant implications forintelligence producers. First, intelligence production managers should be in frequent personalcontact with the personal staffs of the senior consumers that they are trying to serve. In manycases, such staffers should be seen as the firstline, or even primary, consumers. Infrequentcontact with senior staffs increases the risk of not being in on the action when intelligence isneeded or of providing intelligence that is not focused on the actual needs of the policymaker.The challenge is often seen by producers as one of trying to get past these gatekeepers. The realchallenge is providing gatekeepers with the appropriate assistance at the appropriate time inserving their principals.Second, the continuing importance of intelligence liaison officer channels at the Depart-ments of Treasury, Commerce, and Energy appears to provide a model worth building onelsewhere. Every military commander of a unit of battalion or squadron size or larger benefitsfrom a full-time intelligence officer looking out for his needs. It would seem useful to extendthe same service to every assistant secretary or higher-level official in our key foreignpolicymaking agencies. Such liaison officers would help particular producers to stay close toconsumer needs and to protect consumers from being swamped with calls by many producingunits trying to market their services or be of help. Rotational service as a liaison officer wouldalso increase any intelligence officer's awareness of how his or her self-initiated intelligence canbetter serve consumers upon return to a production component.Key Finding #3: Oral Briefings PreferredThe 1988 survey asked consumers about their preferences for various mediums forreceiving intelligence. Between 50 and 60 percent said that they liked getting intelligence viaevery medium?written, oral, or electronic. This is consistent with the Carter-era survey,which found that 60 percent of consumers liked to get intelligence in a combination of writtenand oral forms. As Table 4 indicates, however, the 1988 survey revealed some differences in thepropOrtion of senior officials who had stronger likes or dislikes for the various forms of delivery.Approved for Release: 2014/07/29 000621334Approved for Release: 2014/07/29 000621334Table 4Percent of Policymakers Indicating StrongLikes or Dislikes for Different Intelligence Mediums% Who % Who Have Used Medium &MediumHaven'tUsedLike VeryMuchDislike VeryMuch/Dislike% Who Like Very Much atDoD State White H. OthersOral Intel Briefings8%37%5%37% 30% 41% 43%Formal Intel Pubs4%28%6%24% 35% 47% 23%Written IntelDone for You11%27%7%23% 24% 41% 34%To Who Have Not Used atDoD State White H. OthersVideotape Intel65%17%20%63% 82% 56% 55%Intel on Computer Terminal81%16%26%80% 89% 72% 82%The strong preference for oral briefings is not surprising. Oral presentations, when donewell, are more focused on a particular policymaker's concerns, and there is a chance forfollowup discussion. In some ways, it is surprising that the number of consumers preferring oralbriefings is not higher and that only consumers at Defense and the economic agencies displayedsuch a relative preference in the survey. This probably reflects the situational element that ismissing from the survey question: sometimes oral briefings are preferred; sometimes writtenintelligence is wanted.In contrast to these marginal differences, the response to intelligence via videotape orcomputer terminal is much less positive. Our conclusions have to be tempered by the fact thattwo-thirds or more of our senior consumers have yet to sample these mediums. Of those whohave, however, fewer like them very much, and more disliked the electronic mediums than foreither written or oral intelligence.The heavy investment of time by at least some producers of finished intelligence in thepreparation of formal current and in-depth intelligence publications appears to be out of linewith the marginal preferences of consumers. For example, Table 5 provides evidence on howproduction branches in CIA's Directorate of Intelligence (DI) allocate their production effortand on the frequency with which various product types have been requested.The pattern is clear. Over 60 percent of the DI's effort is put into preparing the NID andthe President's Daily Brief (PDB), hardcover publications, and its periodicals. These productsare the least requested by consumers. In contrast, oral briefings and shorter written products arefrequently requested, but they receive the least amount of the DI's effoA. While a continuinginvestment in basic research by intelligence producers is necessary to make subsequent oralbriefings or talking points both credible and useful, the evidence indicates that we couldprovide better service to key consumers by making changes in two areas. At the margins, moreeffort in preparing, marketing, and presenting intelligence orally and less effort in preparingand reviewing NID items and hardcovers would respond to consumer preferences on how todeliver intelligence. Also, better mechanisms are needed to determine whether a policymakerwants a written report of some kind of whether he or she really wants to pose a few questionsto an intelligence expert. More personal contact with policymaker staffs and more extensive useof intelligence liaison officers would improve the ability of producers to provide the right typeof product at the right time.16 ?cETgEA proved for Release: 2014/07/29 000621334Approved for Release: 2014/07/29 000621334Table 5How Do Production Branches in the CIA'S Directorate ofIntelligence Allocate Their Production Effort?Branch Average inBranch Average inBranchEstimated To of Product TypeType ofFive RegionalFour FunctionalAverage ForDone at RequestProductionOfficesOfficesDirectorateOf a ConsumerOral Briefings10%10%.10%80%+Typescripts,Talking Points,Short Memos18%12%15%60%+PeriodicalPublications15%15%15%3% ?NID/PDB25%10%18%10%Hardcovers25%35%30%20%Bios, Maps,Databases,Other7%18%12%35%Total100%100%100%Note: Based on data collected from 95 DI branch chiefs and experienced analysts who took the Supervision ofAnalysis, Workshop on-Reviewing Analytical Papers, or Reaching Policymakers courses in 1988-89. Excludestime spent on travel, training, administrative, or housekeeping matters.Producers need to find out more about the reasons for consumer resistance to the electronicdelivery of intelligence. Is it mostly reflective of a generational gap between those who gainedtheir professional success before the information revolution and those who began theirprofessional lives in a multimedia world? Or do the new mediums fail to provide somethingthat is valued in the traditional formats: a computer screen cannot be taken to a meeting orhighlighted by magic marker; a videotape lacks the authoritative credibility of an expertbriefing, and it cannot answer your followup questions on the spot. The expense of newelectronic delivery systems will be justified only if they are used.Key Finding #4: A Need for Better Policy-Support IntelligenceThe 1988 survey also recognized the importance of discovering consumer perceptions ofthe quality of intelligence. Even unsatisfactory or low-quality support may be useful, or at leastbetter than no support at all: Table 6 provides the responses of consumers to the question, -Howsatisfied have you been with the performance of the Intelligence Community in providing youwith information that keeps you generally informed about world affairs, that keeps youinformed about your specific areas of responsibility, or that addresses policy objectives of thecurrent administration?" While the meaning of the last category could have been made clearer,other evidence in the survey indicates that it at least partially taps consumer perceptions of howwell the Community does in preparing intelligence that directly supports decisionmaking, asopposed to intelligence that provides background facts or analysis. The first two informationcategories in this question clearly fall into this latter area.Approved for Release: 2014/07/29 000621334Approved for Release: 2014/07/29 000621334Satisfaction RatingTable 6Policymaker SatisfactionWith Different Types of IntelligenceType of IntelligenceIntelligence Intelligence on Intelligenceon Official's Areas To SupportWorld Affairs of Responsibility Decision makingVery Satisfied 24% 28% 9%Satisfied 69% 61% 64%Dissatisfied 6% 10% 21%Very Dissatisfied 1% 1% 6%% Indicating SuchIntelligence Is:EssentialVery Useful8% 71% 15%52% 24% 37%Over 90 percent of senior officials indicated satisfaction with intelligence that essentiallyprovided background on world affairs, although only 8 percent said that such intelligence wasessential. Small pockets of dissatisfaction existed only among officials at State and Defense.These responses validate the high proportion of consumers citing current intelligence publica-tions as useful: 65 percent of all consumers said so about the widely disseminated NID; 75percent of DoD consumers said DIA's Defense Intelligence Summary was useful, and 71percent of State consumers said the same about the State Morning Summary.Eigthy-nine percent of the policymakers expressed satisfaction with intelligence receivedon their areas of concern. This is gratifying, in light of the 71 percent of consumers whoindicated that such intelligence was essential to their jobs. At the margins, dissatisfaction isslightly higher with this type of intelligence in comparison with background on world affairs.Twenty-three officials, including some at the White House and the economic agencies, weredissatisfied with intelligence on their areas of concern. This is 11 percent of the total.These responses are consistent with the significant proportion of consumers who citevarious in-depth intelligence publications as useful. For example, 58 percent of all consumerssaid they found CIA Intelligence Assessments and Research Papers useful; 67 percent at Statesaid INR's Intelligence Reports were useful; and 46 percent of DoD consumers cited DIAAppraisals as useful.Consumer satisfaction drops significantly concerning intelligence meant to supportdecisionmaking and policy implementation. While about one-fourth of all consumers werevery satisfied with the first two categories of intelligence, less than 10 percent said the samehere. The rise of dissatisfaction to 27 percent of the total is of greater concern. Results varysomewhat when they are broken down along agency lines. Only 7 percent of officials at themostly economic agencies were dissatisfied with such policy-support intelligence, but 17percent of the officials surveyed at the White House, 30 percent at Defense, and 38 percent atState were dissatisfied.These responses on the 1988 survey are consistent with comments by senior Carteradministration officials in the earlier survey. Over 40 percent of the Carter officials said thatthey received an insufficient quantity of -intelligence analyses specifically prepared to supportthe development of policy options or operational planning.- Over half of the Carter survey18Approved for Release: 2014/07/29 C00621334Approved for Release: 2014/07/29 000621334respondents said that the quality of the policy-support intelligence actually received was only-fair- or -poor.- The consistency of the data across the two surveys indicates a deep-rooted andcontinuing area of deficiency in such Intelligence Community work.Neither survey asked directly for the reasons why consumers were dissatisfied withpolicy-support intelligence. Some consumer dissatisfaction probably reflects unhappiness withthe judgments provided by intelligence. Answers to three questions on the 1988 survey providesome clues about other sources of dissatisfaction. They also suggest that while 73 percent ofconsumers said that they were satisfied with intelligence to support decisionmaking, many ofthese respondents almost certainly were providing at best a soft endorsement of suchintelligence work.Policymaker staffs frequently are forced to modify or repackage intelligence to make itfocus on their principals' concerns. Listed below are the percentage of policymakers in the 1988survey giving various reasons for the editing done by their staff before they receive intelligencematerial. Many cited more than one reason. The responses indicate that producers of finishedintelligence are too often unaware of what policy officials need.Reason For Staff Editing Percent Citing ReasonTo reduce the length 66%To remove irrelevant material 62%To integrate other intelligence information 50%To identify or emphasize support for current policy concerns 50%To improve the clarity of presentation 47%Another question on the 1988 survey indicates that senior officials sometimes either do notrequest intelligence support before acting or request it and fail to receive it. Less than half ofour senior consumers said that they always sought intelligence before acting, even thoughnearly all of them said that they believed intelligence might have helped. About 40 percent saidthat they failed to do so because time pressures precluded making a request. This also means,however, that intelligence producers have not been close enough to senior officials on someoccasions to know that intelligence would have been helpful. The same is true for the 20percent who said that they did not request intelligence because they thought it was eitherunobtainable or would not be useful. Roughly 30 percent of all consumers said that theysometimes requested intelligence support and either never got it or got it too late to be used.This answer was given most frequently by White House officials; over half of them said so,twice the proportion at other agencies.An open-ended question at the end of the 1988 survey provides striong evidence that manysenior officials want more intelligence that is tailored to their particular needs. When asked towrite in -ways that intelligence producers can improve the usefulness and Quality of theirsupport,- nearly half of the Reagan officials wrote -be more responsive to specific needs,""respond more quickly,- -make tasking easier,- -establish more personal contact,- or -developa rifle rather than a shotgun approach to providing intelligence.- This collection of responsesfar overshadowed the roughly 15 percent who said that they thought the substantive contentor Quality of intelligence needed improvement or the 5 percent who wrote in -I'm satisfied."The strength and breadth across agency lines of the evidence that better intelligence tosupport policymaking is needed indicates the one aspect of our work that is deficient in morethan a marginal way. This judgment based on the survey data is supported by a growing bodyCONFIDENTIAL 19Approved for Release: 2014/07/29 000621334Approved for Release: 2014/07/29 000621334of case-study material and by recent public criticisms of the performance of the IntelligenceCommunity by members of the President's Foreign Intelligence Advisory Board and a numberof former Reagan administration officials.3The judgment that producers are deficient in providing policy-support intelligence mustbe tempered by the evidence in the surveys and in our personal experiences that widevariations exist across cases. Some consumers have gotten just what they needed to makedecisionmaking easier or more informed. The senior officials in the economic agencies appearto stand out in the survey data in this regard. Within every producing agency, some productionunits are more or less dedicated to the publication of basic intelligence. They should not be heldto an unfair performance standard that may indicate a need for substantially improvedpolicy-support work by others. Nonetheless, the basic judgment still stands: the policy-supportwork of intelligence producers is more problematic than their other types of production.How can the problem be fixed? Part of it is caused by policymakers themselves. ShermanKent was among the first intelligence producers to complain publicly about the failure of policyofficials to give intelligence analysts questions to address and about the tendency' of policyofficials to ask questions only at the last minute, thereby guaranteeing a hasty intelligenceresponse to often tough questions. Kent also complained about the failure of the policy agenciesto keep intelligence producers informed about their changing agendas and policies. Unfortu-nately, policymaker culpability for a lot of what is irrelevant or unfocused in intelligence wasdiagnosed in 1948, and it has since received serious comment a number of times. This suggeststhat policymakers are not about to make great strides in helping us better serve their needs inthe 1990s. The ball is in the intelligence producers' court.My own experience, reinforced by the pattern discerned in much of the survey data,indicates that shifts are necessary in how intelligence producers make three tradeoffs that havealways confronted them. First, intelligence analysts have to work harder to increase theirexpertise and knowledge about the US policy process and agenda and less on developing it onforeign lands. This means spending more time studying the US foreign policy process, theshifting legislative and budgetary agendas that set the tempo for much decisionmaking, and theinterests and personalities of consumers. Analysts should try to obtain such knowledge firsthandby talking with officials in policy agencies, by attending significant Congressional hearings asobservers, and by participating in public conferences on foreign affairs issues. Infrequent-parish calls- or feverish running about are not the cure. Time spent out of the office has to beguided by specific purposes and be subjected to some division of labor. Although we will stillneed to devote the bulk of our time to building and maintaining the intellectual capital thatmakes our work credible, we should be spending close to one-third of our time on stayingattuned to the policy process.Second, the consumer surveys show that our self-initiated work is useful and appreciated,but the deficiencies in the area of policy-support work boil down to what intelligence producersare not doing. To do more of such work and to get better at the support we provide, some ofthe effort now devoted to getting out scheduled research has to be shifted to the area wheresenior officials are telling us they need more help.Nearly all the case studies developed by the Program in Intelligence Assessment and Policy at Harvard'sKennedy School of Government illustrate variations of the problem of intelligence - policy disconnect, for example.Recent public criticisms of the deficiencies of intelligence producers in providing policy-support intelligence can befound in Roy Godson, ed. Intelligence Requirements for the 1990s (Lexington Books, 1989) and in articles by RobertGates and Anne Armstrong in The Washington Quarterly (Winter, 1989).20 --GervirreApproved for Release: 2014/07/29 000621334Approved for Release: 2014/07/29 000621334Third, intelligence production managers need to spend more time and effort out of theoffice staying in touch with consumers at their appropriate level and less time and effort in theoffice reviewing paper products and attending to internal bureaucratic concerns. Adjustmentsin the first two tradeoff areas will not occur, unless managers display their own commitment tomeeting more of the particular needs of consumers. Equally important, as managers or theirliaison officers learn what is needed, the message has to be relayed quickly and clearly to theproduction units for action so that relevant products are available when needed.Some Final ThoughtsIf senior officials from at least the last two administrations have been asking for more andbetter intelligence tailored to their decisionmaking and policy implementation needs, why hasthe Intelligence Community been slow to respond? Reforms in these directions will not be easyto make because they run counter to how the intelligence culture operates and because suchchanges entail real risks.? It would be harder for intelligence producers to avoid being co-opted into thepolicymaking process itself.? All intelligence producers would be likely to have less freedom of activity, as thecustomer's agenda controlled increasing amounts of our time.? All DI managers would be less aware of intelligence production activity, and thus lessable to control it.? Consumers' judgments of the Quality of our work would become more important andwould loom much larger in career development.Nonetheless, the risks seem to be worth taking. The margin for policymaking error for theUS Government is declining as the relative power and influence of the US over internationalaffairs declines. When high-quality policy-support intelligence is unavailable, the result will bemore damaging in the future than it has been in the past. The US will increasingly lack theluxury of being able to make repeated or serious mistakes and still protect its interests.Moreover, only the Intelligence Community has the capability and mission to serve all thepolicymaker's information needs, including professionally tailored support of the decisionmak-ing and policy implementation process. Providing useful but often not essential backgroundanalysis and facts is not where the challenge lies for intelligence producers in the next decade.At the level of policy implementation, actionable intelligence and other kinds of support foroperational activity are the growth industry. At the level of policy formulation, -opportunitiesand vulnerabilities- analysis and evaluation of the likely foreign impact of, or responses to, USpolicy options are the areas for major improvements in intelligence; This is the essence of whatconsumers tried to tell us in the survey.This article is classified CO IAL.21Approved for Release: 2014/07/29 000621334