How to warn: ‘Outside-in warnings’ of Western governments about violent conflict and mass atrocities

Abstract

The literature on the warning-response gap in conflict prevention over-emphasizes political will as the crucial variable, whereas warning is not considered problematic. This article makes the case for distinguishing more clearly signs and indications from actual warnings. Furthermore the article argue that the quality of warnings matters for achieving at least partial persuasive success with decision-makers. The article identify key factors limiting or enhancing warning impact, focusing on source credibility, message content and communication mode. They argue that warning communicators need to take credibility problems more seriously, invest more time in identifying, understanding and building relationships with the most relevant recipients and tailor warnings accordingly in terms of content, timing and communication mode. If organizations lack the capacity to provide credible prescriptions on how to act, they should concentrate on high quality reporting to enhance rather than damage their credibility.

Keywords

advocacy NGO persuasion prevention strategic surprise violent conflict warning

Introduction

It has become almost conventional wisdom in conflict and genocide studies that Western governments fail to prevent violent conflict and mass atrocities in foreign countries because of a lack of political will to help foreign citizens, not because of problems with the availability, timeliness or quality of warnings. This argument is repeated in many so-called official ‘post-mortem’ reports and academic writing, as well as public discourse, for example, on the 20th anniversary of the Rwandan genocide (Global Centre for the Responsibility to Protect, 2014; United Nations, 1999: 25; Zartman, 2005). We are not the first to argue that such political will by outsiders to the conflict should not be seen as emerging from a fixed hierarchy of narrowly defined state interests, but rather as the outcome of learning, contestation, advocacy and persuasion (Jentleson, 2005: 24; Woocher, 2001). Warnings about impending conflict in other countries are conceived here as an essential part of a persuasion process about whether, when and how to act preventatively. The process is strongly influenced by the relationship between the sources and the recipients, which is why ‘inside-up warnings’ from the target government’s ‘own’ diplomats, advisors or analysts follow different formal and informal rules than ‘outside-in’ warnings from journalists, NGO staff or individual experts who can more openly and actively persuade, but may also lack information, access channels and credibility.¹ Scholars have studied NGO communication and advocacy efforts around conflict (Aday and Livingston, 2009; Bob, 2005; Fenton, 2009; Goodhand, 2006), but relevant discussions about warning performance can be found with very few exceptions (De Waal, 2015) only in intelligence studies (Grabo, 2010; Kam, 2010: 22–31) and research on disasters and emergencies (Bazerman and Watkins, 2008; Rosenthal et al., 2001). Particularly, the intelligence literature can improve our understanding of how senior decision-makers perceive warning intelligence. However, these insights are not directly transferable to ‘outside-in’ warning about intra-state conflict because this literature has typically focused on direct threats to vital strategic or security interests and is shaped by the widely accepted norm requiring intelligence officials to remain politically neutral.

We start by highlighting three essential content criteria for distinguishing warnings from indications in order to avoid distorted assessments of warning availability and quality. Furthermore, we highlight the importance of warning communication to ensure that decision-makers have actually received a warning message, let alone be persuaded or prompted into action by it. Our research has shown that there have been far fewer actual warnings than one might have expected and they were often late rather than early. Moreover, many of the warnings have often been deficient in terms of their specificity, clarity and actionability. Furthermore, many officials expressed low confidence in the potential of forecasting violent conflict, considered dedicated warning products as irrelevant and had doubts about the credibility of both internal and external sources of warning (Meyer et al., under review).

We then use this framework and research evidence to examine what non-governmental actors can do to enhance their influence on decision-makers (see the methodological Appendix). In particular, we argue that warning producers need to take credibility problems more seriously, invest more time in building relationships with the most relevant recipients and tailor warnings accordingly in terms of content and timing. We do not suggest that better warning alone will fix the warning response-gap first identified by George and Holl (1997). Successful warning depends on a host of factors, some of which are largely outside a warner’s reach, such as competition arising from similar or indeed quite different crises, organizational cultures and incentives against using particular instruments in a preventive way. At the same time, this article holds that better warning is an important part of bridging the gap.

This article draws on findings from a large research project on warning and conflict prevention, which pursued a mixed methods design with two main strands: firstly, a process-tracing of warnings and their impact on intra-governmental deliberations and public debates in the EU, Germany, France, the UK, and the US and across seven cases: Estonia 1991, Rwanda 1993–1994, Kosovo 1998–1999, Macedonia in 1999–2000, Darfur 2002–2004, Turkey/North-West Iraq 2003–2004 and Georgia 2007–2008. Secondly, we asked practitioners from national governments and International Organizations (IOs, such as the EU, UN and OSCE), news media and NGOs in semi-structured interviews about some of the factors helping or hindering the communication and reception of warnings and specific ones about their perception of different sources (see Appendix). In total, 120 respondents were interviewed in the period 2010–2012, plus three follow-up interviews with NGO representatives in 2015. The split was roughly 40 percent government officials, 35 percent journalists and 25 percent NGO and think-tank staff. Interviewees among the journalists were mostly foreign correspondents working for quality news media, as well as some capital-based foreign news editors, among government officials a mixture of desk-officers and heads of units, intelligence analysts, ambassadors and senior policy advisors and among NGO staff predominantly communication and advocacy directors as well as a smaller number of country/issue experts. The online survey covered 40 officials, roughly two thirds at desk-officer level and one third at head of unit level. Of these, 40 percent were working for national governments or national intelligence services and the rest for IOs, especially the EU, UN and OSCE. The findings and best practices were tested with eight representatives of relevant NGOs and government departments at a workshop on 7 November 2012.

Conceptualizing warnings and their success

We have previously argued that warnings are a form of persuasion and should be understood as ‘a communicative act with the intention to raise a given recipient’s awareness about a potential threat to a valued good or interest in order to enhance her ability to take preventive or mitigating action’ (Meyer et al., 2010: 567). For a communicative act to qualify as a warning, the warning message needs to contain three core elements: a knowledge claim about future harm, a judgement about the importance of this harm and an assessment about what action could be taken to avoid or mitigate the risk. Not all of these assessments need to be highly explicit and specific to qualify as a warning, especially with regard to the third criterion of prescription. In addition to these message criteria, warning producers need to have at least tried to communicate these judgements to relevant recipients in a way that stood a chance of reaching them. Finally, warnings need to originate from a source with at least the basic competence to make such a knowledge claim, whether it is because of their training, experience or background, the process by which knowledge is generated, and/or their access to relevant up-to-date information about the factors driving conflict.

These minimalist criteria for identifying a warning are central to our critique of the prevailing consensus on early warnings. The first and most important problem is the frequent confusion of warnings with ‘warning signs’ or ‘indications’ which are used interchangeably by leading authors in the field of conflict prevention such as Zartman (2005: 5), who criticize decision-makers after the fact for taking a certain course of action despite ‘all the warning signs were there’. In contrast to what is often understood as warning, the mere reporting of an event does not qualify as such: when protests erupt in the run-up to or after elections in a country, such events may not be unusual for the country and society in question, and security services and institutions may have shown in the past the resilience to cope with them. Alternatively, they may be unusual in nature or frequency, or come at a time of reduced coping capacity, therefore indicating a rapidly rising risk of escalation and the outbreak of violence. Indications do not speak for themselves, but require interpretation and judgement about a situation and how it is likely to develop in the future: ‘warning is an intangible, an abstraction, a theory, a deduction, a perception, a belief. It is the product of reasoning or of logic, a hypothesis whose validity can be neither confirmed nor refuted until it is too late’, as Grabo (2010: 14) writes in her handbook on warning intelligence. The use of the term ‘warning signs’ disguises the potential hindsight bias because the wording implies that the signs should have been interpreted in a certain way (Fischhoff and Beyth, 1975). If the attribution of accountability is the aim, references to ‘warning signs’ may suffice to provoke public indignation, but are not a sufficient basis for attributing accountability for failure, nor learning the right lesson from ‘failure’.

Secondly, even when warnings have been available and communicated to decision-makers, we should not judge their performance solely in terms of whether decision-makers have taken action or not. Warning impact on the policy process is a graduated outcome involving several stages such as reception, acceptance, prioritization, mobilization to act, and ultimately the implementation phase (see Figure 1).

Figure 1.

Modelling outside-in conflict warning success.

Warnings may succeed at one stage, but not the following, and they may ‘fail’ relatively without anyone being necessarily ‘at fault’. Potentially justifiable reasons for why decision-makers discount or de-prioritize a warning could include, for instance, a lack of convincing evidence, too much background ‘noise’ or contradictory evidence from other sources (Wohlstetter, 1962), a weak warning track-record of the source, supposed flaws in the warning’s reasoning, or other, plausible alternative scenarios. After all, warnings can also be right for the wrong reasons (Jervis, 2010: 2–3). Furthermore, decision-makers may be persuaded by a given warning, but do not deem the consequences important enough to mobilize further resources given other pressing policy concerns or crises, whether domestic or international in origin. They may also accept the forecasting element of the warning, but do not accept the explicit – or implied – policy prescriptions as they may deem the risks or costs too high. Some of these diverging judgements by decision-makers may be perfectly legitimate as they are rooted in divergent interpretations of values and political priorities. Alternatively, warning recipients may have superior insights into the prospects for political mobilization, or the feasibility of using certain tools as compared with warning producers.

However, there may also be unjustifiable reasons for discounting ‘quality’ warnings. Some policymakers may hold ‘the strongest views on what the policy should be’ despite not being very knowledgeable, as the founder of US intelligence studies Sherman Kent wrote in 1968. Individual or policy predispositions, or ‘predilections’ to use Kent’s term, may prove insurmountable obstacles to persuasion attempts and inhibit policymakers from learning and adjusting their attitudes. They may have a wholly unjustified confidence in their knowledge of certain countries and conflicts or trust instead judgements of close political advisors who play a particularly strong role in US foreign policy according to the former Head of the US Agency for International Development, Andrew Natsios (22 October 2009). Even less justifiable than professionally deficient judgements are cases when decision-makers do not even try to learn from warnings or wilfully distort relevant knowledge claims because they are politically inconvenient, or because they reject any ethical and legal responsibility to protect people from impending humanitarian disasters or mass atrocity crimes.

This discussion shows already that warnings are unlikely to succeed completely and persuasion may fail, regardless of how good a warning is. Whilst it is true that ‘perfecting intelligence production does not necessarily lead to perfecting intelligence consumption’ (Betts, 1978: 61), we disagree with the widespread scepticism in the intelligence community (Betts, 1982, 2007; Jervis, 2010; May and Zelikow, 2006a) and indeed peace studies about the potential for better warnings to affect also the consumption-side and thus increase the probability of early action.

There are at least two reasons to be more optimistic about the potential of ‘outside-in warnings’ about violent conflict within states: firstly, non-governmental actors increasingly have an information advantage as compared with governments, at least in cases of foreign countries that are not considered strategically important (interviews with International Crisis Group [ICG] Communication Director, 1 September 2011, and Jonathan Prentice, ICG Advocacy Lead, 4 June 2015). Western governments prioritize intelligence gathering vis-à-vis countries that either constitute a direct threat or are widely considered to be of strategic economic or political significance, criteria which exclude many of the most conflict-prone countries. Moreover, our interviewees repeatedly highlighted that many Western governments have made cuts over the years to their diplomatic presence in foreign countries and have neglected cultivating linguistic and social knowledge of countries and regions most at risk of instability. In contrast, conflict prevention and humanitarian NGOs such as ICG, Human Rights Watch (HRW) and Médicins sans Frontières (MSF) have become more numerous and expanded their capabilities, and have thus been able to launch and maintain field missions and offices in countries at risk of instability. Many NGOs have also developed in-house expert knowledge and built track-records for accurate reporting and good judgements according to a number of our interviews with officials from various organizations. As NGOs often maintain their distance and neutrality to government and political parties, they may have greater access to actors deemed politically controversial such as groups classified as terrorist, which state diplomats could not speak to in their official capacity. Even in those areas where NGOs and other non-governmental sources do not fill a significant knowledge gap left by state intelligence and foreign affairs reporting, products and indeed warnings generated by those actors have become an inextricable part of the ‘swirl of intelligence’ informing government as one former member of the US National Security Council called it (anonymous interview, 20 October 2009). The so-called ‘swirl’ comprises not just formalized systems of gathering Open Source Intelligence (Hobbs et al., 2014), but also informal channels through which knowledge claims and advocacy efforts of NGOs affect foreign policy.

Secondly, non-governmental actors have more freedom than officials in how they communicate their warnings and to whom. While there are considerable differences in the extent to which NGOs, and indeed journalists, can be explicit about their normative and political agendas, they generally need to worry much less about overstepping the mark to politics as intelligence agencies and officials in foreign affairs ministries do. This enables skilled and well-resourced ‘outside-in’ warners not just to be more explicit about the importance of a given warning and what should be done, but also to employ a far wider range of persuasive means such as appeals to public sentiment or decision-makers’ worldviews and values. ‘Outside-in’ warners also have more discretion in deciding whom they wish to address as compared to an intelligence analyst, who, first and foremost, reports to his or her line manager or according to established procedures. They may decide to approach one or many individual decision-makers at different levels directly, employ intermediaries, or decide to ‘go public’ in various ways by targeting audiences of the greatest relevance to decision-makers. Yet, this relative freedom comes at a price: many outside-in warners may struggle to get their messages ‘into the system’ in the first place. Thus, when discussing best practices of warning, one key issue is whether sources are actually recognized as such and whether they are in a position to get their messages heard.

‘How to warn’ from outside the foreign policy system

Perceptions of sources

Source credibility is the most important precondition for a warning message to make an impact according to our research, so sources should be concerned about how potential recipients see them (Meyer et al., under review). As decision-makers and those working for them in governments usually have to deal with a large amount of information, their interactions are inevitably routinized in order to cope with their workload. Changing the perceptions of officials and creating a response from hierarchically organized foreign affairs bureaucracies is a considerable challenge, not least due to frequent information overload and potentially bureaucratic inertia. Even warners who are confident about their unique expertise and superior evidence will not succeed in getting past bureaucratic and software filters designed to allow access only to certain individuals or organizations.

Above all, however, interviewees from government and IOs emphasized that sources of warning will not get a full hearing from decision-makers if they are not deemed trustworthy to start with (e.g. mid-level EU official, 14 April 2011). Thus, potential ‘warners’ need to establish themselves in the ‘realm of credible sources’ in the first place. We found that warning sources differ not only with regard to their perceived credibility but also to their capacity to argue their case convincingly and communicate effectively. Source capacities have a bearing on the degree of impact and, depending on individual skills, they may also facilitate gaining access. In general, however, source credibility decides about access, both physically and, most importantly, in a cognitive way. Trusted civil servants or political advisors, for example, are usually in a good position to get their message heard at the highest levels of foreign policy decision-makers. If the person is also a skilled communicator, the likelihood increases that he or she will also be able to influence a decision-maker’s appraisal of the situation. However, the key to gaining access and having an impact is ‘trust’. Physical face-to-face access facilitates that a message reaches recipients, but source credibility is at the core of successful warning.

Credibility is difficult to deconstruct into its constituent parts. In addition to the above-mentioned aspects, ideological proximity often plays a strong role. When recipients suspect sources are advancing hidden agendas, this is often closely connected to ideological distance between source and recipient. It is worth remembering that decision-makers are acutely aware that they are commonly the aim of persuasion attempts, which is why they will be rather cautious if there are no well-established relationships with a certain ‘warner’. Some warners were perceived as being prone to exaggeration to attract greater attention from policy-makers, media and potential donors, for instance, by mislabelling intercommunal fighting as genocide. While a stated commitment to a certain cause may not be problematic per se (see below), the credibility can suffer if decision-makers gain the impression that strong ideational beliefs pollute good conflict analysis or the appropriate use of legal concepts under international law. Sources are also more likely to be branded as either ‘do-gooders’ or ‘zealots’ by diplomats, military staff and politicians if they advance prescriptions on how to act that are deemed ill-informed as to their political or practical feasibility.

It would be unrealistic to suggest that NGOs should sacrifice their principles or agendas for the sake of increasing their impact on policy as they may very well alienate their other constituencies. However, they can still act on three fronts: firstly, if an organization – or individual – has a clear commitment to a specific goal, there is no point in hiding it. Many decision-makers and politicians in particular are confident that they can ‘compensate’ for biases from individuals and organizations with a known agenda or standpoint, but a source’s credibility can take grave and lasting damage if the recipient senses a ‘hidden agenda’. Secondly, sources are well advised to ensure through appropriate funding models, rules and training that their advocacy and fundraising activities do not unduly influence their research and decision-making about when, how and about what to warn. Warnings are by their very nature inconvenient to some political actors, including some of the NGO funders, and this may result in self-censorship. NGOs can also increase their credibility by admitting and correcting errors and by learning even from those who may appear as natural opponents. For instance, even though HRW’s main tactic is naming and shaming human rights abusers, they are also aware that understanding the abusers’ motives will enhance the credibility of their analysis and help to better target advocacy (interviews with HRW Communications Director Emma Daly, 14 and 16 July 2015). Thirdly, if a source perceives itself too ideologically tainted in the eye of the target recipient to successfully warn itself, it may try the indirect approach by seeking to convince those organizations and actors who will not suffer from the same ideological distance to the recipient.

Fortunately, ideological compatibility is just one factor influencing source perceptions. Deliberations of the authoritativeness of sources are also part of how ‘warners’ are perceived, and if a source is usually better informed than others, ideological distance is likely to matter less. Authoritativeness is best understood as a combination of quality of information and analysis, as well as the relevance of this analysis to decision-makers. Unless there are strong personal ties to recipients, the track record of sources with regard to timeliness, accuracy and relevance matters greatly for being considered authoritative. Furthermore, the relative performance of a ‘warner’ in comparison to other sources is expected to play a role as well (Brante, 2012: 262). If certain actors can regularly offer ‘exclusive’ information or relevant advice, this is likely to contribute to a positive track record. According to Brante, to build a ‘solid track record’, ‘producers must be significantly more right than they are wrong and must not have ‘“cried wolf” too often in the past’. The latter task is far from straightforward as any attempt to avoid false positive warnings altogether will bias an organization towards under-warning. Sources should think carefully about where they can offer specialist knowledge, and how to acquire an expert reputation over time. One approach is to develop expertise for a certain country, region or issue through hiring, retaining and supporting staff, as well as tools for organizational knowledge management and analysis such as rigorous cross-checking and ‘when in doubt, leave out’ rules. If an organization or individual has attained an expert status over a prolonged time in a ‘non-warning’ role, it would facilitate their credibility when they communicate a warning. Communicating a strongly worded warning about the relatively rare threat of violent conflict without such a good track record in a non-warning role raises the risk of damaging one’s reputation for a long time. Furthermore, authoritativeness is frequently boosted by what is perceived as ‘first-hand experience’, especially if a source has recently been to the country in question (interview with former USAID Director Andrew Natsios). Somewhat surprisingly, recipients tend not to distinguish between whether a source actually visited the area of conflict or only a country’s capital. Similarly, the duration of the stay does not seem to matter greatly either (Meyer and Otto, 2012).

The extent to which ‘warners’ have established relationships with recipients is closely connected to this issue. The term ‘relationship’ should be understood broadly in this context. Although personal acquaintance and interaction is a clear advantage, ‘established relationships’ need to be understood more abstractly. Recipients need to have heard about the source, its activities, operations and ideological commitments before. Once a warning is communicated in a given situation, recipients will not have to start from scratch in forming a view about the source. Above all, if there is an established relationship between a source of warning and a recipient, one can expect that there will be a certain degree of trust in the interactions, which in turn is likely to increase receptivity.

The term ‘relationship’ also implies that we are dealing with a two-way process instead of information travelling one way to recipients only. For warnings to be relevant, ‘warners’ need to have an understanding of what kind of information is valued by consumers, what their concerns are, what objectives they have and which constraints they face (Betts, 2003: 61). Within the intelligence community, this approach has become known as the so-called ‘Gates model’, named after Robert Gates, the director of Central Intelligence under President George HW Bush. It represents a departure from the so-called ‘Kent model’ which emphasizes that analysts should keep their distance from policymakers as objectivity should never be compromised by including political deliberations in reporting and analysis. For sources of warning from outside government the problem may be less that they let ‘policymakers believe what they want’ but that their perspectives may be too divergent. Thus, their challenge is to maintain accuracy and truthfulness in their analysis and still establish some common ground in terms of decision-makers’ needs.

The message

Our research shows that the content of the warning message – whether delivered orally or in writing – does matter, even if it is less crucial than source credibility, which should be viewed as a precondition for message acceptance. The quality of the manifest warning content can be enhanced through maximizing clarity and relevance, the two key factors by which recipients will judge the quality of the message. Clarity can be understood as the degree of specificity and certainty of the judgements comprised in the warning. Being specific about the direct consequences of violent conflict is of utmost importance to warning of intra-state conflicts in other countries as the nature, degree and spread of the harm may differ greatly. Certainty can be expressed in various ways, but is a key marker of how confident a warning source is about a given threat. One key indicator of confidence is to actually use the word ‘warning’, especially since our research found that this kind of explicitness is the exception rather than the rule in public communication on conflict. If a source intends to warn of an impending harm, the term ‘warning’ or ‘warn’ should be used explicitly and sufficiently prominently in the text, such as in executive summaries (Kam, 2010: 25). This is all the more the case because as Betts (1980) points out, ‘decision makers are used to living in an environment of some warning’ (p. 561, emphasis in original), so lowered attention thresholds have to be factored into how the message is presented. They also typically lack the time or inclination to read through long reports and thus easily miss warnings hidden in the text (Betts, 2003: 62). At the very least, outside-in warning messages conveyed on the basis of longer, research intensive reports, should present the most important supporting evidence underpinning the warning judgement prominently in order to make sure that key information is not buried in the long text. Policymakers frequently assume the role of a ‘last analyst’, i.e. they do not necessarily trust the analysis of others, but prefer to reach their own conclusions (Kuhns, 2003: 95). Recipients may disagree with the conclusions, but they cannot claim ignorance – neither of the ‘facts’ nor the resulting warning – as a defence if key evidence was impossible to overlook.

In our research we also found that warnings of intra-state conflict, which expressed their assessments confidently and did not hedge their judgements, fared relatively better in terms of being noticed and prioritized in comparison to more cautious assessments. This was not only the case with regard to policymakers, but also in relation to attracting media attention (Otto, 2012). However, warnings which were specific as to the who, what, to whom, when and why and which expressed their judgements with a high degree of certainty were fewer than one could expect from much of the literature on the warning-response gap and almost always communicated immediately before a conflict escalated or after the outbreak of violence. Actual early warnings have been comparatively rare, and the large majority of them have been vague and not very specific. So-called ‘post-escalation warnings’, i.e. those that highlighted a further deterioration of the situation or aimed at clarifying what was actually happening after conflict had broken out, performed better with regard to certainty and specificity (Meyer et al., under review).

This is not an unusual finding for the strategic surprise literature as ‘only rarely has a strategic warning been issued in time’, according to Kam (2010: 30). For warnings to be considered ‘timely’ they would need to allow decision-makers sufficient lead-time for the use of those instruments that can realistically prevent violence to escalate. Yet, political mobilization to act is influenced by the certainty and specificity expressed in the warning message. ‘Warners’ thus face a complicated trade-off between warning early with less information and often highly hedged, vague and ambiguous language, or warning too late but being more certain and specific. Scholars of intelligence and foreign policy are highly aware of the difficulties of judging enemy intentions (Handel, 1984: 239–241), but warning of violent conflict in foreign countries adds an additional layer of problems due to the need to assess not just the immediate threat, such as the severity of violence affecting foreign populations, but also its nature, spread and the multiple and indirect consequences to government interests that may flow from it over time. However, rather than managing this dilemma by being vague, warners should detail different scenarios, specify the factors underlying each of them and, if possible, judge their relative likelihood (Smith, 1969). If neither the scope conditions nor the likelihood are explicated in warnings, providing different scenarios will not solve the issue. It may even be highly counter-productive because recipients might get the impression that warning producers seek primarily to hedge against all eventualities.

Related to the previous aspect, we found that most of the publicly communicated warnings we identified in our research were not very nuanced in the use of probabilistic language, veering between the extremes of over-confident prediction that something ‘will happen’ or highly vague cautious language stating that ‘the situation may deteriorate’. More generally, expressions of probability, such as ‘likely’, ‘probable’, ‘a serious possibility’ or ‘almost certain’ have limited utility and may mean very different things to different people. The problem of divergent interpretations of probability-related terms is well understood in intelligence studies (Kent, 1964), but hardly appreciated by NGO staff, journalists or indeed the general audience. There is no agreement within the intelligence community as to whether probabilities should be spelt out in exact numbers or odds but, for the sake of clarity, ‘outside-in’ warners should consider using either numbers or odds directly in their messages or assign them to frequently used expressions of probabilities. Such a step will force warners to probe their assumptions and judgements more intensively and minimize misunderstandings stemming from diverging interpretations. If one arrives at the conclusion that the outbreak of conflict is ‘likely’ and the decision is made to communicate this judgement, one should be ready to say what ‘likely’ actually means, as for instance, with guidance on expressions of uncertainty contained in a report of the Intergovernmental Panel on Climate Change (IPCC, 2007).

A more tangible approach to increasing the impact of warnings on the level of message content concerns the way messages are ‘framed’ in order to make them relevant to recipients. This does not mean that knowledge should be distorted or manipulated. As Betts (2007: 98) notes, ‘the challenge remains to make intelligence relevant without making it dishonest.’ When top decision-makers’ policy predispositions are deeply entrenched, producers may have to moderate their tone in order to get their opinions heard. For instance, when the crisis of President Ferdinand Marcos’s regime in the Philippines deepened from 1984 onwards, the US intelligence community deliberately balanced its reporting on Marcos to the US administration by mapping a range of different scenarios, even though there was an evolving consensus that Marcos would not be able to hold on to power and the US government should re-consider its previous support for the regime (Kline and Worthen, 2006: 153).

Our research indicates that the warnings, which had the highest impact, usually addressed the specific concerns of individual recipients, for instance, when George W Bush was warned how the escalation of violence in Darfur could affect his re-election campaign. This ties in closely with the above-mentioned ‘Gates model’. Furthermore, it cannot be expected that a publicly communicated warning will resonate equally with a variety of different national and sectoral ‘audiences’. Thus, ‘warners’ should carefully think about how they can tailor their message so that they relate to the ‘needs’ of the most important recipients. The basis for being in a position to address these issues is a thorough understanding of what matters to policymakers and what they already know – knowledge that one typically acquires through building a relationship with them. A potentially helpful approach is to ask a set of questions that approach the problem from the perspective of policymakers such as: What interests are at stake and how does the conflict relate to existing policy priorities? What are the costs of getting involved/not getting involved? What worldviews does he or she hold and what individual experience do recipients have with issues related to the problem?

If decision-makers operate on the basis of strong beliefs about a certain issue, May and Zelikow (2006b: 211) recommend an approach consisting of three steps, which is also illustrative of how to get a better understanding of recipients’ priorities: firstly, ‘identify the presumptions in the minds of decision-makers’, for example by following or tracing what high-level government sources stated in the media. Secondly, ‘assess the strength of a given presumption in the mind of a given decision-maker’ and ‘ask what new facts could either strengthen or weaken a key presumption’. In the case of a warning running counter to deeply entrenched beliefs, the question would be what information could lead the recipient to change his or her attitudes, and how can problematic assumptions, which are likely to limit receptivity, be addressed. Thirdly, ‘start a search for facts that bolster or weaken an important assumption.’ Related to this approach, Brante (2012: 347) suggests that ‘warners’ should ‘“anchor” warnings in the knowledge about consumers’ predispositions’ and ‘customize warnings in order to better answer the “so what”-question and push the right “hot buttons”.’ Another technique May and Zelikow (2006b: 215) suggest is called ‘placing’, which starts from seeing individual recipients ‘in their own histories and thinking about experiences from which they may have carried away lessons’. The warning source would then deploy a very specific and targeted way of analogical reasoning by using, for instance, historical analogies to underscore the severity of a situation and illustrate potential scenarios (‘another Bosnia’, ‘another Rwanda’, etc). However, the degree to which such cases resonate may well differ among individual decision-makers voters and party-members, so warners need to be able to modulate.

Communication modes

Warnings have to be communicated to recipients, as ‘warning which exists only in the mind of the analyst is useless’ (Grabo, 2010: 25). ‘Outside-in warners’ lack the communication channels available to government officials and intelligence agencies such as regular briefing documents, reports and indeed specialized risk assessment and warning products. However, this lack of formalized access does not mean that NGOs and other warning sources such as journalists cannot cultivate and skilfully use other channels and communication modes to ensure their messages reach the intended recipient and are heard. ‘Outside-in warners’ can choose in principle between four such modes: inter-personal direct, inter-personal indirect, public-mediatized, and public-direct. This is not necessarily an either-or choice, but could involve skilfully combining or alternating communication modes. However, some organizations will lack the resources, networks or expertise to use all of these communication modes equally well.

Our research clearly shows that for reaching and persuading top-decision-makers, direct, inter-personal communication provides the highest chance of warnings to be understood, given attention and accepted (Meyer et al., under review). However, this presupposes direct access, which is partly a function of source credibility and established personal relationships as discussed earlier, partly determined by an organization’s wider influence and resources in a given policy-field. Some of the largest and more reputable NGOs manage to gain direct access to senior decision-makers quite easily even if this is far from a guarantee that their analysis let alone their recommendations will be accepted.

When access to top decision-makers is impossible for a less resourceful or well-connected organization or individual, an alternative is to work laterally with more easily accessible officials at working level, such as country desk-officers, to get the warning into the system. However, given that such sources may be several levels of hierarchy removed from the ultimate target recipient, warners may need to increase the frequency of the message and seek to convince officials in different parts of the foreign affairs machinery such as defence, development, foreign ministries or embassies at the local level. Communicating a warning via one or many intermediaries within government, especially through trusted advisors, also offers a solution for scenarios in which sources are aware that recipients do not generally trust their track-record, or see them as politically biased as discussed under source credibility. How to reach expert communities at working levels will largely depend on resources available for research, advocacy and outreach. Large INGOs with a strong field presence such as Oxfam or MSF have such resources, but also comparatively smaller NGOs specializing in peace and Human Rights advocacy. ICG, for instance, has established a number of products, including conflict warnings that are regularly distributed to and widely read by expert subscribers across the world, many of them within government ministries, as confirmed by many of our interviewees and the responses to our expert survey question to name ‘the THREE most useful and trustworthy sources of warning about intra-state conflict (e.g. a specific named NGO, newspaper, early warning system, academic research unit, other government or international organization)’. For organizations and individuals with less reach, credibility and breadth of expertise, the importance of developing relationships with relevant officials cannot be overstated.

‘Outside-in warners’ can also aim to boost the message impact by forming coalitions with other warning organizations in communicating their messages through public reports, which can be expected to become part of Open Source Intelligence (OSINT) systems. Multiple ‘warners’ not only increase message frequency, but they can also reinforce each other’s messages, provided cross-source consistency is high. Once sources have identified which other players are active on an issue or country, they can explore how to join forces to increase their impact with decision-makers.² This approach can be further boosted by improving the timing of such communication and advocacy activities to match the receptivity of their target audiences – for instance, avoiding major party conventions or summits or exploiting windows of opportunity after changes in key personnel or during periods of policy review. Increasing warning frequency, modulating their timing to recipients’ attention spans and joining forces can compensate for lack of access as long as the escalation of violence on the ground is not imminent and preventive action not highly urgent.

If the escalation of a conflict is imminent and senior decision-makers’ engagement is particularly important, alternative communication modes are needed that offer the potential to quickly shift attention and attitudes about a potential conflict among a large number of people. NGO interviewees highlighted that there is often no substitute to getting into wide circulation or opinion-leading papers to break through with new or urgent messages. This can be through ‘outside-in warners’ submitting an op-ed piece themselves or through influencing commentators and bloggers to write a piece about a specific topic. Even if warners do not achieve or wish to mobilize political pressure for specific kinds of action, they may seek to influence the media reporting. An in-depth report about a country or region published in a leading news medium can contribute to an information environment in which warnings are more likely to be understood and prioritized. Furthermore, news coverage can provide a ‘peg’ for civil servants or members of an administration who support a warning, to bring the issue on the agenda. The downside of going through the news media is that key aspects of the warning message may be adapted, simplified and indeed distorted to make the message compatible with journalistic notions of newsworthiness and increase its appeal to audiences.

The public-direct channel cannot just be used to target expert policy-communities who subscribe for professional reasons to their communications, but could be aimed at larger lay audiences. Some organizations such as HRW have 2 million followers on Twitter and can make use of this audience to raise attention (interviews with Emma Daly, Communication Director, 14 and 16 July 2015). NGOs are also increasingly producing videos of different lengths that can be effectively distributed via social media and, if they go viral, can create a huge amount of attention with relatively little investment. The relatively largest and most advocacy-focused NGOs have become genuinely multi-channel organizations, which will look very carefully at which channel or platform is most likely to resonate with the target audience. This can be highly variable depending on whether the target is the UN or the EU, the US or European governments, foreign governments in conflict regions or citizens. If there is sufficient time, organizations can try to mobilize grass-roots for preventive or mitigating foreign policy action among their members, communities and constituencies that are of particular importance to senior decision-makers. We know from our interviews that diaspora groups in certain Western countries such as the UK, especially the Kurds, Tamils or Palestinians, are often very active in mobilizing their members for public demonstrations and letter-writing to MPs in order to raise awareness of and prompt action against impending or current conflict. In the case of the 2003 violence in the Western Sudanese region of Darfur, we found that the George W Bush administration became more engaged in the conflict at least due in part to the mobilization of Christian groups in support of fellow Christians in the South of Sudan.

Conclusion

This article set out to provide advice for boosting the impact of ‘outside-in’ warnings of violent conflict targeted at Western governments. Whilst acknowledging the multifaceted and significant obstacles to prompting effective preventive action through warnings, we have argued for a more nuanced measurement of warning success and suggested a clearer definition of warnings as conveying three specific judgements with the intention to persuade. Furthermore, we challenged some of the fatalism in the literature about the immutability of political will to heed warnings and argued that current warning practices can be substantially improved in terms of source credibility, message content and communication mode, making at least partial success more likely.

Which of these practices an organization should adopt when and in what combination will depend on a whole range of factors, including its resources and reputation, agenda competition from other crises, the specific conflict case and which governments are to be targeted. We stressed that much of the success of warning will depend on the relationship that a given warning source has managed to cultivate with relevant officials and decision-makers well in advance. The most skilfully worded and evidenced warning message will not succeed if the warning source is not considered credible. While there may be practical, as well as normative limits to building such relationships, for many NGOs – given their values and mission statements – there can be no substitute to anchoring the warning process in reliable knowledge of the warning recipients: their interests, worldviews, available instruments and more. Sources need this knowledge to better tailor the content, timing and mode of their warning messages, but need to do so without meddling with the facts or losing their ability to think outside established policy paradigms and conventional wisdoms.

Warnings that do not convey a clear judgement and which fail to at least specify what is likely to happen when and where, and which are merely ‘thrown at the system’ have little value. They may be downright harmful as they can contribute to the creation of noise, which risks drowning out more relevant signals or actual warnings. If uncertainty makes the formulation of clear warning judgements impossible, sources should rather limit themselves to reporting facts. As mentioned above, decision-makers are heavily inclined to do their own analysis and they generally value factual reports. Thus, sound reporting is anything but futile. At the very least, it can help sources to build a positive track record that will reinforce their credibility over time.

The credibility of outside-in warners may also benefit from giving credit for preventive action taken and lives saved where it is due. Whereas there is no shortage of analysis dealing with missed opportunities, wrong steps taken at the wrong time and the consequences of inaction, success stories are hardly communicated by NGOs and the media, reflecting the negativity bias of much foreign affairs coverage. One reason for this is the difficulty of convincing a sceptical public that prevention actually worked. Yet, it is not impossible to identify at least partially successful cases and show that certain measures taken by governments have saved lives as in Sierra Leone in 2000 or Macedonia in 2001. If successes of preventive action were highlighted more often by outside actors disinterested in political credit claiming, this could help to change the calculus of decision-makers over time that reacting to warning can also be an opportunity.

Footnotes

Appendix: Sample questions used in semi-structured interviews

Acknowledgements

The authors would like to thank the anonymous reviewers and editors of the journal for their helpful comments. We are grateful to the many interviewees who gave generously of their time as well as a group of practitioners for their feedback during a special workshop on Warning and Conflict Prevention. The authors would like to thank John Brante and Chiara de Franco for their contribution to some of the underlying research and thinking. The usual disclaimer applies.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research was made by possible by a European Research Council grant (FORESIGHT, No 202022) and follow-up funding from the Policy Institute at King’s for research and the workshop.

Notes

Author biographies

Christoph O Meyer is Professor of European & International Politics at King’s College London and Vice-Dean Research in the Faculty of Social Sciences and Public Policy. He is also Co-PI on an EU-funded project on media and violent conflict (). He has been working on security and defence policy, including early warning for conflict prevention, public communication and media coverage as well as economic governance. He has contributed to debates about the European public sphere and political integration, European strategic culture and questions of forecasting, preventive policy and risk management.

Florian Otto leads the country risk analysis on Europe and Central Asia of Verisk Maplecroft, a leading international risk analytics, research and strategic forecasting company. In this capacity, Florian regularly comments on current affairs and his analysis has appeared in various trade publications. Before joining Verisk Maplecroft, Florian completed his PhD in War Studies, King’s College London, and conducted extensive research on warning of intra-state conflict and the changing nature of international media coverage. He retains a strong interest international security, communicating risks and forecasting.

Address: Verisk Maplecroft, 40 Gracechurch Street, London EC3V 0BT, UK.

Email: florian.otto@maplecroft.com

References

Aday

Livingston

(2009) NGOs as intelligence agencies: The empowerment of transnational advocacy networks and the media by commercial remote sensing in the case of the Iranian nuclear program. Geoforum 40: 514–522.

Bazerman

Watkins

(2008) Predictable Surprises: The Disasters You Should Have Seen Coming, and How to Prevent Them. Boston, MA: Harvard Business School Publishing.

Betts

(1978) Analysis, war and decision: Why intelligence failures are inevitable. World Politics 31: 61–89.

Betts

(1980) Surprise despite warning: Why sudden attacks succeed. Political Science Quarterly 95: 551–572.

Betts

(1982) Surprise Attack, Washington, DC: The Brookings Institution.

Betts

(2003) Politicization of intelligence: Costs and benefits. In: Betts

Mahnken

(eds) Paradoxes of Strategic Intelligence: Essays in Honor of Michael I. Handel. London: Frank Cass.

Betts

(2007) Enemies of Intelligence: Knowledge and Power in American National Security, New York: Columbia University Press.

Bob

(2005) The Marketing of Rebellion: Insurgents, Media, and International Activism. Cambridge: Cambridge Ubiversity Press.

Bock

(2015) Firmer footing for a policy of early intervention: Conflict early warning and early response comes of age. Journal of Information Technology & Politics 12: 103–111.

10.

Brante

(2012) A European Strategic Surprise: Analysis and Warning in the EU before the Russo-Georgian Five-Day War of 2008. London: King’s College London, Department of European and International Studies.

11.

De Waal

(2015) Advocacy in Conflict: Critical Perspectives on Transnational Activism. London: Zed Books.

12.

Fenton

(2009) New Media, Old News: Journalism and Democracy in the Digital Age. Los Angeles: Sage.

13.

Fischhoff

Beyth

(1975) ‘I knew it would happen’: Remembered probabilities of once–future things. Organizational Behavior and Human Performance 13: 1–16.

14.

George

Holl

(1997) The Warning-Response Problem and Missed Opportunities in Preventive Diplomacy: A Report to the Carnegie Commission on Preventing Deadly Conflict. Washington, DC: Carnegie Corporation.

15.

Global Centre for the Responsibility to Protect (2014) Rwanda 20 Campaign. Available at: http://www.globalr2p.org/our_work/rwanda_20 (accessed 30 March 2015).

16.

Goodhand

(2006) Aiding Peace? The Role of NGOs in Armed Conflict. Boulder, CO: Lynne Rienner.

17.

Grabo

(2010) Handbook of Warning Intelligence: Assessing the Threat to National Security, Lanham, MD: Scarecrow Press.

18.

Handel

(1984) Intelligence and the problem of strategic surprise. Journal of Strategic Studies 7: 239–241.

19.

Hobbs

Moran

Salisbury

(2014) Open Source Intelligence in the Twenty-First Century: New Approaches and Opportunities. Basingstoke: Palgrave.

20.

IPCC (2007) Climate Change 2007: Synthesis Report. An Assessment by the Intergovernmental Panel on Climate Change.

21.

Jentleson

(2005) Coercive prevention: Normative, political, and policy dilemmas. Peaceworks. Washington, DC: United States Institute for Peace.

22.

Jervis

(2010) Why Intelligence Fails: Lessons from the Iranian Revolution and the Iraq War. Ithaca, NY: Cornell University Press.

23.

Kam

(2010) Surprise Attack: The Victim’s Perspective. Cambridge, MA: Harvard University Press.

24.

Kent

(1964) Words of estimative probability. Studies in Intelligence, Fall.

25.

Kent

(1968) Estimates and influence. Studies of Intelligence, Summer. Available at: https://www.cia.gov/library/center-for-the-study-of-intelligence/csi-publications/books-and-monographs/sherman-kent-and-the-board-of-national-estimates-collected-essays/4estimates.html (accessed 16 June 2013).

26.

Kline

Worthen

(2006) The Philippines, 1983–1986: Arranging a divorce. In: May

Zelikow

(eds) Dealing With Dictators: Dilemmas of U.S. Diplomacy and Intelligence Analysis, 1945–1990. Cambridge, MA: MIT Press, 137–165.

27.

Kuhns

(2003) Intelligence failures: Forecasting and the lessons of epistemology. In: Betts

Mahnken

(eds) Paradoxes of Strategic Intelligence: Essays in Honor of Michael I. Handel. London: Frank Cass.

28.

May

Zelikow

(2006a) Introduction: Seven tenets. In: May

Zelikow

(eds) Dealing with Dictators: Dilemmas of U.S. Diplomacy and Intelligence Analysis, 1945–1990. Cambridge, MA: MIT Press, 4–13.

29.

May

Zelikow

(2006b) Learning from the cases. In: May

Zelikow

(eds) Dealing with Dictators: Dilemmas of U.S. Diplomacy and Intelligence Analysis, 1945–1990. Cambridge, MA: MIT Press.

30.

Meyer

Otto

(2012) Missing the story? Changes in foreign news reporting and their implications for conflict prevention. Media, War & Conflict 5: 205–221.

31.

Meyer

. (2010) Re-casting the warning-response-problem: Persuasion and preventive policy. International Studies Review 12: 556–578.

32.

Meyer

. (under review) Heeding Warnings about War: Persuasion in Conflict Prevention. Manuscript under revision for re-submission to Cambridge University Press

33.

Otto

(2012) Breaking the news: When warnings of intra-state conflict make news. PhD thesis, accepted by King’s College London, Department of War Studies.

34.

Rosenthal

Boin

Comfort

(2001) Managing Crises: Threats, Dilemmas, Opportunities. Springfield, IL: Charles C Thomas.

35.

Smith

(1969) On the accuracy of national intelligence estimates. Studies in Intelligence 13.

36.

United Nations (1999) Report of the Independent Inquiry into the Actions of the United Nations during the 1994 Genocide in Rwanda. New York: United Nations.

37.

Wohlstetter

(1962) Pearl Harbour: Warning and Decision. Stanford, CA: Stanford University Press.

38.

Woocher

(2001) Deconstructing ‘political will’: Explaining the failure to prevent deadly conflict and mass atrocities. Journal of Public and International Affairs 21: 179–206.

39.

Zartman

(2005) Cowardly Lions: Missed Opportunities to Prevent Deadly Conflict and State Collapse, Boulder, CO: Lynne Rienner.