The Surprising Decline of Mediation in armed conflicts

Remy Maduit | Authors published

The Surprising Decline of International Mediation in Armed Conflicts

Magnus Lundgren is a Senior Lecturer and Associate Professor at the Department of Political Science
of the University of Gothenburg, Sweden. 
Isak Svensson is a Professor at the Department of Peace and Conflict Research at Uppsala University, Sweden.

Over the last decades, states, international organizations, and other actors have mobilized to build a stronger infrastructure to support conflict mediation. Paradoxically, this mobilization has not translated into more mediation. The number of armed conflicts in which a third party intervenes to facilitate a negotiated solution has stagnated and, in recent years, has gone into decline. This presents a puzzle: why are so few conflicts mediated when the capacity to mediate is greater than ever before?

In this study, we seek to place this puzzle on the research agenda and to make a preliminary assessment of the most promising candidate’s explanations. We conceptualize mediation occurrence as a function of supply and demand-side conditions and investigate whether empirical shifts in these conditions can account for the increasing shortfall in mediation. [i] We employ data from the Uppsala Conflict Data Programme (UCDP) covering the period between 1989 and 2013 and carry out a descriptive comparative analysis supported by statistical modeling (presented in the online appendix). [ii]

We find that a mismatch between supply and demand can explain the puzzle in the international mediation ‘market’. Although there are more mediators available, the changing nature of the armed conflict, specifically the increasing involvement of Islamist actors, has made mediation applicable to fewer cases. In parallel with the rise of Islamist armed conflicts, key norms and international practices of conflict management have changed, privileging stability and counterterrorism over conflict resolution. We identify the increasing use of the prescription of armed groups through terror listing as a key mechanism linking the changing nature of conflicts to the decline in mediation. We show empirically that Islamist armed conflicts, which are less likely to receive mediation, make up a larger proportion of the population of armed conflicts, and that conflicts that involve terror-listed actors are considerably less likely to receive mediation. Our analysis provides statistical corroboration of arguments about shifts in international norms [iii] and case studies arguing that terrorist proscription creates obstacles to peace negotiations. [iv]

Although we cannot determine conclusively whether it is the changing nature of conflicts or the changing norms of conflict management that best explain the decline in mediation, our initial analysis suggests that it is primarily the latter; terror-like violence has not increased, and before terror-listing became commonplace, mediation rates of conflicts involving Islamists did not differ significantly from other conflicts.

In the following, we explain the puzzle of mediation occurrence and then briefly discuss the main rival explanations. We find that the decline in mediation cannot be attributed to the increased fragmentation, protractedness, or internationalization of the civil conflict, nor that it is likely to result from the uneven regional distribution of mediation capabilities. We also consider and debunk the possibility that the observed trends are artifacts of the data.

A puzzling relationship

Mediation is on the decline. Figure 1 exhibits the yearly count of conflicts and mediated conflicts since the end of the Cold War. [v] Both follow declining trends and the proportion of conflicts that receive mediation is decreasing (see also Figure 2). [vi] In the last 15 years, two-thirds of armed conflicts did not receive mediation. Although some scholars have made passing references to declining mediation trends, the pattern has largely gone unnoticed in the literature.[vii] Rather, the conventional interpretation is that mediation is on the increase, reaching a historically high rate during the post-Cold War years. [viii] For example, according to Gowan and Stedman (2018: 175), a ‘dramatic rise in mediation in civil wars’ characterizes the post-Cold War period. This interpretation is not incorrect, but it emerged from data with a limited temporal scope. Our more expansive data suggest that the 1990s may have been an outlier decade, with historically high rates of mediation, rather than a ‘new normal’.

Based on what we know from previous research, we would have expected such increases in technical mediation capacity to lead to more mediation. For example, Gowan and Stedman (2018: 172) portray a development of mediation and mediation capacity that go hand in hand, as a sign that conflicts are treated in a standard way: ‘The ascendancy of mediation as a frequent, almost automatic international response to civil wars over the last thirty years, as well as the development of institutions and technical capacity in peacemaking and peace-building, suggests that a new international regime for treating civil wars took hold in the 1990s.’ We show that the two trends go in opposite directions, suggesting a puzzling relationship. Although the international system has added mediation capabilities, mediation incidence has not followed suit.

Why so Few Contemporary Armed Conflicts are Mediated—Existing Explanations

How can we account for this puzzling pattern? For mediation to occur, a willing mediator and belligerents who are ready to engage with the mediator need to exist. It is, therefore, useful to think of mediation occurrence as a function of supply-side-side factors about the incentives and capabilities of prospective mediators and demand-side factors of the incentives of the belligerents. Either or both may contain explanations for the observed shift in mediation.

On the demand side, research has pointed to factors such as the dispute and the power balance between incumbents and insurgents. [ix] A significant part of this discussion builds on Zartman’s notion [x] of ‘ripeness’, showing that belligerents’ demands for external mediation will increase when they find themselves in a ‘mutually hurting stalemate’, a costly and ineffective military situation. On the supply side, research has suggested that mediation occurrence is a function of third-party interest, geographic proximity, and mediator capability. [xi]

From this perspective, the temporal shift in mediation incidence must stem from temporal shifts in supply and demand factors in the population of ongoing conflicts. On the demand side, if the population of conflicts is adding cases less likely to develop into ‘ripe’ situations, it could help to explain why international mediation has been less frequently applied. On the supply side, if the population of conflicts is becoming dominated by cases that external mediators find less interesting, or are incapable of reaching, it could feasibly explain why overall mediation rates are decreasing.

If the nature of conflicts is changing, how can it help to explain declining mediation rates? Previous research has identified several ways in which armed conflicts have changed in contemporary times. Some point to fragmentation, leading to a proliferation of veto players that can create obstacles for mediators. [xii] Others point to the increasing intractability and extended duration of contemporary conflicts. [xiii] The third line of thought emphasizes increasing internationalization as the key change in the landscape of armed conflicts. [xiv]

Our analysis casts doubt on these factors as plausible explanations for the mediation puzzle. Theoretically, it is unclear why trends of fragmentation, intractability, and internationalization would decrease the space for mediation. The opposite should be true. Fragmented conflicts would be more difficult to end through peace agreements, but a higher number of conflictual relations should increase, not decrease, the need and opportunity for mediators. Duration should, theoretically, be associated with a higher chance of a negotiated settlement, and because the length of conflict increases its costs, it should reward primary and third parties to come to the negotiating table. [xv] Also, although previous research has shown that conflicts with outside interventions last longer, it is conceivable that internationalization would increase the likelihood of mediation. [xvi] For example, Greig [xvii] shows that involvement by one or more of the great powers predicts mediation, and Lundgren and Svensson [xviii] show that mediation between belligerents that receives external support is more likely to facilitate agreement.

Our empirical analysis reinforces the theoretical suspicion that these trends cannot account for the decline in mediation. First, we examined the possibility that the decline is explained by the increasing fragmentation of civil conflict. Our data show that conflict fragmentation, defined by dyad count per conflict, has not changed significantly during the post-Cold War era, undermining this as an explanation of a shift with a clear temporal dimension (Figure A1). Second, we explored whether the decline of mediation was because of an increase in protracted conflicts. We found that although the average duration (time since the outbreak) of conflicts increased from 9.7 years in 1989 to 14.1 years in 2013, there is no evidence that protracted conflicts receive less mediation (Figure A2). Third, our analysis suggests that internationalization does not provide a convincing independent explanation. When controlling for conflict characteristics, internationalized conflicts do not have a lower chance of receiving mediation than other conflicts (Table A1 and Figure A3).

Why so Few Contemporary Armed Conflicts are Mediated—Radicalization and Proscription

Instead, we suggest another explanation. A subset of conflicts drives the decline in mediation, Islamist conflicts, which are extremely under-mediated and on the rise. [xix] The increase is quite dramatic. Although only 8% of conflicts were classified as Islamist in 1989, the proportion increased to 56% by 2013. The increasing prevalence of Islamist armed conflicts has been noted [xx], but not linked to mediation occurrence. Our data shows conflicts fought over Islamist incompatibilities exhibit lower (and decreasing) mediation probabilities. Figure 3 demonstrates how the number of Islamist conflicts increases and the number of mediated Islamist conflicts decreases, leading to a widening mediation gap for this type of conflict. Only 14% of conflict years with Islamist claims received mediation after 2001, compared with 37% of conflict years in which neither side had made such claims. This difference is statistically significant, even when controlling for type of conflict, intensity, rebel strength, region, and period-specific fixed effects (Table A1). Given that Islamist conflicts represent an expanding share of all conflicts, their low mediation rates may help explain part of the puzzle.

The occurrence of Islamist armed conflicts is closely associated—as a cause, but also to some extent as a result—with the ‘War on Terror’ (WOT), the wide-sweeping military campaign launched in response to the 9/11 attacks, and the concomitant prevailing norm of not ‘talking to terrorists’. As Howard and Stark (2018) have suggested, shifting norms of conflict management have led we labeled to a situation in which we saw more conflicts through a securitized prism and in which armed actors as terrorists. Terror-listing decreases both the demand for mediation and the supply of mediators. Previous research shows that prescription via terror listing can place legal restrictions on contact with conflict actors. [xxi] It also strengthens ‘hawks’ over ‘doves’ in intra-insurgency competition, decreases perceptions that a negotiated ‘way out’ exists and creates reputational problems for third parties reluctant to be seen as legitimizing terror groups. Indeed, the military framing of the problem of terrorism implicit when referring to the WOT places high political costs on mediation in certain conflicts: Many mediators would think twice before engaging with actors deemed illegitimate by the UN or powerful countries, leading to a ‘chilling effect’ that reduces mediation.

To examine if our data support this interpretation, we identified conflicts with links to terrorism and assessed their mediation status. [xxii] As seen in Figure 4, the proportion of conflicts in which at least one actor is designated as a terrorist organization by the UN has soared. At the start of our observation period, no such conflicts existed; in the end, two-fifths of all conflicts involve such actors. Remarkably, of the 95 conflict years involving terror-listed actors, only 2 received mediation, suggesting that terror-listing has a considerable ‘chilling effect’ on international mediation. [xxiii]

Importantly, our analysis shows that this change is primarily driven by shifts in the practice and norms of conflict management rather than changes in the conflicts themselves. Two patterns in the data support this conclusion. First, the number of actors using one-sided violence against civilians—a reasonable proxy for the use of terrorist tactics—is relatively constant over time (Figure 4). Second, before 2001, Islamist conflicts were not less likely to receive mediation than other conflicts (Table A1, Figure 3). It is the international community’s response to conflicts that exhibit one-sided violence or involve Islamist armed actors that have changed, leading to a diminished role for mediators in these cases.

Do We Know that Mediation is in Decline?

Before drawing this article to a close, we want to briefly discuss the robustness of the main finding we have presented: the decline of mediated conflicts. Three concerns need to be addressed here.

First, can the observed trend be an artifact of the data? If observing mediation has become more difficult, for example, similarly, we would observe a declining aggregate trend. We do not think this is a convincing explanation. If a temporal reporting bias exists, it is more likely to affect earlier periods. Indeed, it would be expected that the risk of missing data is lower in the post-2000 period than in earlier years, which is central to our argument, both because globalization has made remote locations more accessible and because social media makes it more difficult for informal or secret mediation initiatives to escape notice. It is, therefore, unlikely that the observed trend results from a temporal reporting bias. Analysis of alternative data, sourced from the International Crisis Behaviour dataset and reported in the online appendix, supports this conclusion. [xxiv] We also replicated the major trend, aggregating mediation episodes at the dyad level rather than at the conflict level (Figures A4 and A5).

Second, there may be a geographic mismatch in the distribution of mediators and mediation capabilities, with some regions receiving fewer than others. Because the European conflicts had an above-average mediation rate, their disappearance from the universe of conflict cases would imply a decrease in the average mediation rate. Similarly, Asia, which has low mediation rates, has not experienced a dramatic reduction in the number of conflicts, with the implication that its pool of under-mediated conflicts remains longer in the population of conflicts, reducing the overall mediation rate. However, even though discrepancies exist about regional mediator resources (again, Europe has the most and Asia the least), it does not explain why global actors such as the UN would mediate less in certain regions, likewise; it does not tell us why regional mediators would not be willing to mediate in other regions (for example; the EU has been willing to serve as a mediator for Asian conflicts).

Third, it should be recognized that our investigation of the mediation puzzle has been confined to high-level mediation (Track 1). We have not studied mediation at lower levels (Track 2 or Track 3). Substitution effects, by which informal and local mediators fill the gap left when formal mediators disengage, cannot be excluded. Unofficial mediators may have particular possibilities of engaging with Islamist actors, for example, by more easily overcoming the obstacles of terror listing.

Conclusions

This study has established a puzzling contradiction in contemporary mediation. Although the preparedness of international mediators has increased, the proportion of conflicts that receive mediation is stagnant or decreasing. We examined candidate supply and demand-side determinants for the observed pattern. We found that the observed shortfall is unlikely to stem from fragmentation, protractedness, or internationalization. Scrutinizing the robustness of the findings, we found that neither regionalization of conflict mediation capacities nor reporting bias provide satisfactory explanations for the observed decline.

Instead, we have suggested that the increasing prevalence of Islamist armed conflicts is the most prominent explanation behind this contradictory trend. Over the last two decades, conflicts in which at least one side made self-proclaimed Islamist demands have become both more common and increasingly unlikely to experience mediation. We identified terrorist prescriptions as a related mechanism; conflicts that involve terror-listed actors are considerably less likely to receive mediation. Because the international community has become increasingly willing to terror-list actors, this has placed an increasing number of conflicts beyond the reach of international mediators. It linked the observed pattern to a discontinuity in mediation rates around the turn of the millennium. Our explanation linking these two phenomena is the WOT, the campaign against terror organizations, frequently with Islamist ideologies, that was started in the wake of the 9/11 attacks and which paved the wave for a normative shift in the international community’s approach to conflict management.

Our results point to further important implications for the study of mediation. First, it suggests that the increase in mediation experienced in the first decade after the Cold War may be a historical anomaly, rather than the start of an era with consistently high mediation rates. [xxv]

Second, although previous research has characterized the mediation field as a ‘crowded stage’, our analyses suggest that there is still plenty of room for engagement. [xxvi] Despite the increasing mediation capabilities around the world, about two-thirds of all conflicts experience no international mediation in a year.

Finally, from a policy perspective, it is contradictory that western countries, the key architects behind the mobilization of international mediation capabilities, are also the most ardent pursuers of anti-terror sanctions. Whereas rhetoric and funding have increasingly supported diplomatic solutions via mediation, the growing use of another instrument, terror listing, which effectively places many conflicts beyond the reach of mediation diminished the range of conflicts to which such tools can be applied. If the practice of mediation is to remain relevant for managing and resolving armed conflicts in the future, there is a need to find ways of broadening involvement.

[i] Mediation is defined as a non-violent third-party intervention seeking to resolve or manage conflict. We focus here on mediators accepted by both sides and involving high-level representatives. We do not include Track 2 or Track 3 efforts, because no data sources suitable for comprehensive comparison exist.

[ii] Armed conflict is defined as a contested incompatibility resulting in at least 25 battle-related deaths. See UCDP. Although data on armed conflict are available until the end of 2018, mediation data are no longer available after 2013. Once further mediation data become available, mediation rates in the period after 2013 should be assessed.

[iii] Howard, LM, Stark, A (2018) How civil wars end: The international system, norms, and the role of external actors. International Security 42(3): 127–171.

[iv] Haspeslagh, S (2013) ‘Listing terrorists’: The impact of proscription on third-party efforts to engage armed groups in peace processes—a practitioner’s perspective. Critical Studies on Terrorism 6(1): 189–208.
Toros, H (2008) ‘We don’t negotiate with terrorists!’: Legitimacy and complexity in terrorist conflicts. Security Dialogue 39 (4): 407–426.

[v] We count a conflict as mediated if there is at least one mediation attempt in the year of observation. We replicate the pattern on dyad-level data in Figures A4 and A5.

[vi] Time series analysis of dyad-level data suggests the downturn is unlikely to result from random chance. In a model that controls for conflict characteristics and geographic region, the coefficient of a period fixed effect for the post-2001 period is negative and significant at the 99% level. See Table A1.

[vii] Beber, B (2010) ‘The (Non-)Efficacy of Multi-Party Mediation in Wars Since 1990’. Unpublished manuscript.
Svensson, I, Wallensteen, P (2016) When do we mediate? Global trends in armed conflict and peace mediation. In: HD Centre | Norwegian Ministry for Foreign Affairs (eds) The 2016 Oslo Forum Network of Mediators. Oslo: Norwegian Ministry for Foreign Affairs, 40–47.

[viii] DeRouen, K, Bercovitch, J, Pospieszna, P (2011) Introducing the civil wars mediation (CWM) dataset. Journal of Peace Research 48(5): 663–672.
Greig, JM, Diehl, PF (2012) International Mediation. Cambridge: Polity Press.

[ix] Clayton, G, Gleditsch, KS (2014) Will we see helping hands? Predicting civil war mediation and likely success. Conflict Management and Peace Science 31(1): 265–284. Greig, JM, Regan, PM (2008) When do they say yes? An analysis of the willingness to offer and accept mediation in civil wars. International Studies Quarterly 52(4): 759–781.
[ix] Clayton, G, Gleditsch, KS (2014) Will we see helping hands? Predicting civil war mediation and likely success. Conflict Management and Peace Science 31(1): 265–284. Greig, JM, Regan, PM (2008) When do they say yes? An analysis of the willingness to offer and accept mediation in civil wars. International Studies Quarterly 52(4): 759–781.

[x] Zartman, IW (2001) The timing of peace initiatives: Hurting stalemates and ripe moments. The Global Review of Ethnopolitics 1(1): 8–18. Greig, JM, Regan, PM (2008) When do they say yes? An analysis of the willingness to offer and accept mediation in civil wars. International Studies Quarterly 52(4): 759–781.

[xi] When do they say yes? An analysis of the willingness to offer and accept mediation in civil wars. International Studies Quarterly 52(4): 759–781.  

[xii] Cunningham, DE (2006) Veto players and civil war duration. American Journal of Political Science 50(4): 875–892.

[xiii] Fearon, JD (2017) Civil war and the current international system. Dædalus 146(4): 18–32.

[xiv] Melander, E, Pettersson, T, Themnér, L (2016) Organized violence, 1989–2015. Journal of Peace Research 53(5): 727–742.

[xv] Mason, DT, Fett, PJ (1996) How civil wars end: A rational choice approach. Journal of Conflict Resolution 40(4): 546–568.

[xvi] Regan, PM (2002) Third-party interventions and the duration of intrastate conflicts. Journal of Conflict Resolution 46(1): 55–73.

[xvii] Greig, JM (2005) Stepping into the fray: When do mediators mediate? American Journal of Political Science 49(2): 249–266.

[xviii] Lundgren, M, Svensson, I (2014) Leanings and dealings: Exploring bias and trade leverage in civil war mediation by international organizations. International Negotiation 19(2): 315–342.

[xix] In line with previous research, we define an Islamist conflict as a conflict in which at least one actor plans explicit Islamist political aspirations at the onset (Svensson and Nilsson, 2017).

Svensson, I, Nilsson, D (2018) Disputes over the divine: Introducing the religion and armed conflict (RELAC) data, 1975–2015. Journal of Conflict Resolution 62(5): 1127–1148.

[xx] Gleditsch, NP, Rudolfsen, I (2016) Are Muslim countries more prone to violence? Research & Politics 3(2): 1–9.

[xxi] Haspeslagh, S (2013) ‘Listing terrorists’: The impact of proscription on third-party efforts to engage armed groups in peace processes—a practitioner’s perspective. Critical Studies on Terrorism 6(1): 189–208.

Toros, H (2008) ‘We don’t negotiate with terrorists!’: Legitimacy and complexity in terrorist conflicts. Security Dialogue 39(4): 407–426.

[xxii] We identify terror-linked armed groups based on the UN Security Council’s Consolidated Sanctions List, available at www.un.org/sc/suborg/en/sanctions/un-sc-consolidated-list.

[xxiii] Statistical modeling supports the finding; see model 5 in the online appendix.

[xxiv] Brecher, M, Wilkenfeld, J (2000) A Study of Crisis. Ann Arbor: University of Michigan Press.

[xxv] DeRouen, K, Bercovitch, J, Pospieszna, P (2011) Introducing the civil wars mediation (CWM) dataset. Journal of Peace Research 48(5): 663–672.
Greig, JM, Diehl, PF (2012) International Mediation. Cambridge: Polity Press.

[xxvi] Crocker, CA, Hampson, FO, Aall, P (2001) A crowded stage: Liabilities and benefits of multiparty mediation. International Studies Perspectives 2(1): 51–67.

.