Moral hazard, intervention and internal war: A conceptual analysis

Abstract

Intervention may cause as well as calm internal wars. One way it may cause them is captured by the concept of moral hazard, which suggests that domestic groups which would not otherwise resort to political violence may be encouraged to do so by the prospect of outside support. In this piece we unpack and examine the descriptive and causal logic of that concept of moral hazard. First, we explore the links from our concept of moral hazard to more general social science concepts—perverse incentives, negative precedents and unintended consequences. Second, we focus on three key propositions embodied in that concept which explain how intervention may cause internal wars, and indicate empirical patterns which must obtain if the explanations are to be valid. These three are: 1) that the rebels' resort to political violence is induced by incentives created by the intervenor's actions—and not by a change in their underlying motivation; 2) that the result is harmful to the intervenor's goals and interests; and 3) that the intervenor did not intend to induce that result. Next, we distinguish between moral hazard as a remote versus proximate cause of conflict and between using it to explain a single conflict and to explain multiple conflicts. We then map out the four main explanatory contexts implied by these categories and discuss issues attending each of them. Finally, we conclude with key points to consider when advancing or opposing moral hazard explanations of internal war.

Acknowledgements

Earlier versions of this paper were presented at the Department of Political Science IR Workshop at Boston College, November 2004, the Annual Meeting of the American Political Science Association, September 2004 and the Georgetown University Workshop on Intervention, October 2003. Thanks for helpful feedback to: Orly Mishan, Jon Culp, Alexandre Provencher-Gravel, David Deese, Arman Grigorian, Don Hafner, Dennis Hale, Ron Hassner, Alan Kuperman, Robert Rauchhaus, Jack Snyder, Leslie Vinjamuri, Harrison Wagner, and Jon Western.

Notes

1. As a causal mechanism, moral hazard may produce the onset of war, or prolong it. This piece will focus on the former role, although many of the points apply as well to the latter. See Crawford (2004).

2. A fuller discussion of normative issues appears in Crawford (2004). For an exhaustive cross-disciplinary study of the normative dimensions of the concept, see Baker (1996).

3. The close to Viner's passage is telling: “it was possible to picture [this] as a clash between the humane and hard-hearted”. Similarly, Kuperman notes that “a policy of not intervening in cases of intentionally provoked genocide is open to criticism as being hard-hearted” (2003, p. 68). Although the poor laws debate obviously evoked the insurance–moral hazard connection, the term ‘moral hazard’ did not come into use until 30 years later, specifically in relation to private insurance rather than government welfare policies. See Baker (1996, p. 248, n. 44).

4. James D. Fearon (forthcoming) reinforces this traditional critique of self-determination. A melding of the negative precedent problems of appeasement and partition appears in relation to government responses to secessionist movements. Why do some governments absorb tremendous costs to crush breakaway rebellions on largely worthless territory? The most compelling answer is that they do so in order to avoid encouraging secessionist movements waiting elsewhere in the wings. This is the thrust of Walter (2003).

5. Nevertheless, Rauchhaus (2005) does not agree that the concept of moral hazard fits well with the empirical contours of the political problem.

6. It is essential to posit such revisionist motivations, because the prospect of protection against loss may have a very different effect on status quo seekers—that is, it may embolden them to take risks to preserve peace.

7. As I put it elsewhere, moral hazard “arises when you encourage others to do things against your interest with the prospect that you will bear the costs of their behavior” (Crawford, 2001, p. 504).

8. India's ambassador to the UN declared that India's intervention to “rescue the people of East Bengal” was motivated by “absolutely nothing but the purest of motives and the purest of intentions”. Quoted in Franck and Rodley (1973, p. 276).

9. This calls to mind the principle of ‘double effect’ in just war theory. See Walzer (2000, p. 153). Thanks to Ron Hassner for pointing this out. On using the principle to think more broadly about the unintended consequences of foreign policy, see Jokela (2005).

10. Note, however, that Grigorian (2005) and Rauchhaus (2005) may disagree with this point; for them a moral hazard problem cannot exist if the intervenor is able to observe the perverse consequences of its policy, and is also able to alter the policy, but persists in it anyway.

11. According to the second edition of the Oxford English Dictionary, ‘unwitting’ unambiguously denotes “having no knowledge or cognizance of a particular fact, thing, etc.”. Very close in meaning is ‘inadvertence’, which denotes “inattentive, negligent, heedless” persons, or actions “characterized by want of attention or taking notice; hence unintentional”.

12. This is my extrapolation from Kuperman's larger argument.

13. According to Hugh Thomas (2001, pp. 328–329.), “apparently Franco did [not] know of” Mussolini's support for the failed rightist plot of 1934, but General Emelio Mola Vidal, the instigator of the part of the 1936 uprising that began in Spain proper, certainly did. Franco's half of the revolt started in Morocco.

14. This reasoning suggests the work of scholars who emphasize the necessary role of ‘opportunity structures’ in the decisions of aggrieved and mobilized groups to rebel. See McAdam (1982) and Tarrow (1994).

15. There is a challenge to the deductive logic in this line of argument, which starts with the observation that it assumes away the problem of misperceptions in real-time decision making. Thus, as Kuperman (personal communication) has pointed out, “one could argue the opposite—that distant causation is less likely to produce unintended consequences because policy makers previously will have had time to see the initial unanticipated consequences of the causal variable and to make mid-course corrections to avoid their repetition”. The disagreement thus reduces to two larger disagreements about the prevalence of misperceptions. The first pertains to the relative proportion of misperceptions to accurate perceptions in general: how often do policy makers make choices based on inaccurate understandings of their situation? The second pertains to whether misperceptions are likely to be stronger and more prevalent in the near term or over the long run. The key work arguing for the high frequency and tenacity of misperceptions is Jervis (1976).

16. The ‘scope’ condition—the issue or behaviour influenced—has been made explicit throughout: that is, the propensity of the induced actors to rebel violently. The most thorough treatment of the need to specify scope, domain and other boundary conditions in power analysis is Baldwin (1989).

17. Kuperman notes that, “this was not the first time…that the US [gave] false hope to Iraq's Kurds only to abandon them to Saddam's mercy” (2003a, p. 60, n. 22). The previous incident, which elicited Henry Kissinger's famous remark that “covert action should not be confused with missionary work”, occurred in 1975 (Issacson, 1992, p. 564). In 1972, as a favour to the Shah of Iran, the USA began secretly arming the Kurds. Three years later, when Iran made amends with Iraq, the USA abandoned the Kurds, who were then demolished by Baghdad. The fact that in this case they had been summarily abandoned suggests, however, that it was not a precedent which encouraged them to rebel in 1991.

18. In a radio interview with Marvin Kalb (America Abroad Media, 2003), former Director of Central Intelligence Robert Gates, who served on the deputies committee which handled Iraq policy was asked: were “the Shiites in the South and the Kurds in the North…encouraged…to rise up against Saddam Hussein? Did you ever discuss coming to the immediate military help of these two groups?” Gates responded: “The President's encouragement had been an expression of hope that the Iraqi people would take action to depose Saddam Hussein. I don't recall that the Kurds or the Shia were specifically incited to rise up themselves. I think that in the back of our minds our hope had never been, or our belief had never been, that that the Kurds or the Shia could get rid of Saddam. It was the Generals that we anticipated would get rid of him. And frankly the uprisings gave Saddam the pretext for trying to hold the military together while he put them down.”

19. The International Crisis Group (2003: pp. i, 14) reported that the anti-Taylor factions in the fighting of 2003 had received at least “tacit backing” from the US government in their efforts to depose him. Yet, during the same period, Taylor was calling for US intervention to stop the fighting and broker a transfer of power under government elections. For example, on 8 July, in a CNN interview, Taylor called for US intervention to break the siege of Monrovia by rebel forces: “Some of us feel that you owe it to this country, because we've been by you and we continue to stand by you” (CNN.com Transcripts, 2003). In 1990 Taylor's foe Samuel Doe similarly appealed to Washington for support: “We implore you to come help your stepchildren who are in danger of losing their lives and their freedom” (quoted in Pham, 2004).

20. I have not found evidence of rebel decision making to support this hypothesis, so it must be considered valid only as an illustration of a potential explanation based on remote and singular moral hazard.

21. The USA did make explicit statements suggesting that it might intervene in the summer of 2003, but these came much too late to explain the rebels' decisions to fight in the first place.

22. Thanks to Kuperman for calling attention to the potential for structural change to trigger proximate and plural consequences.

23. “For many years”, notes the International Crisis Group, “observers of the Balkans had expected trouble to break out in Macedonia before Kosovo erupted rather than afterwards” (2001a, p. 9). See also International Crisis Group (2001b, p. 2) and Daalder (2000).

24. However, if one were to argue that a direct effect of the humanitarian intervention regime is to cause governments (anticipating its emboldening effect on subordinate groups) pre-emptively to inflict genocide on potential challengers, you would have a full-fledged instance of thick moral hazard. This causal mechanism is noted in Wagner (2005)—although he does not consider it a moral hazard phenomenon—and discussed at length in Bloom (1999).

25. On the idea that “US power creates its own foreign policy energy”, see Posen (2004, p. 8).