More on Moral Hazard in US Alliances: Explaining Japan-Korea Tension (and Greece-Turkey?)

by on 2014-01-27 in Duck- 10 Comments


So this post is a bleg to those of you who know more about alliances than me. I am considering writing this up for an article, so I thought I would ‘crowd-source’ early comments on the basic argument. I also wonder if someone elsewhere has already suggested this idea in the vast alliance literature. So please let me know. The motivation is inductive – the deepening tension between Japan and Korea has suggested the addendum to alliance theory I am proposing here. But I wonder if others have said this before.

Put briefly, I don’t think entrapment or abandonment captures the US position between allies in dispute, like Japan and Korea, or Greece and Turkey (perhaps – I know that latter case less well). Instead, each seems to use the US alliance patron to: a) compete with each other, because b) the US alliance relieves external pressures (China and North Korea, and the USSR and chaos in the Balkans and Middle East, respectively) that would otherwise incentivize a rapprochement. These four states are not trying to ‘entrap’ the US so much as leverage it for an intra-alliance squabble, with the shared patron as referee. I’ve not read this theorized elsewhere, so here is an effort to do so.


The following is written in laymen’s terms, as it is the fuller, unedited version of something originally written for the Lowy Institute:

“The last six months have seen some of the worst Korean-Japanese tension in a long time. It should be painfully apparent now that Korea and Japan are not allies, despite, curiously, parallel alliances with the US. Specifically, it is an open question where South Korea’s allegiances would land were Japan and China to stumble into conflict in the coming years. While the US would likely arm-twist Seoul into siding with Tokyo, South Korea would probably tilt against Japan were its choice fully independent.

This strikes me as a theoretical tangle for international relations theory (IR), as I think we expect alliances to create a polarizing effect of two distinct camps. But contemporary Asia is not really shaping up like 1914, WWII, or the Cold War with their clearly defined blocs, and Japanese-South Korean tension is one of the big reasons why. This is a rich theoretical puzzle for any aspiring IR Asianists in need of a good dissertation topic.

I have been trying theoretically for awhile to explain Korea-Japan tension (here, here, here among others). Given their position right next to China, North Korea, and Russia, and shared liberal democratic values, they should be allies, but they aren’t. In fact, they are the opposite: Robert Gates’ memoir notes that former Korean President Roh Moo-Hyun labeled Japan as the greatest threat to Asian security, a notion that Gates said was a ‘little crazy’. So I’m increasingly using the idea of ‘moral hazard’ to explain their lack of alliance. Both are ‘insured’ by the US against more serious threats, so they can say all kinds of outlandish stuff about each other with no consequence.

Alliances are an old topic in IR. Going back to Stephen Walt, and Thomas Christensen and Jack Snyder and who knows how many other realists, a basic alliance concern has been either abandonment of the client or entrapment of the patron.

Abandonment is when a patron does not help an ally in a conflict; the ally then must capitulate to a threat or bandwagon with it. A recent Asia-Pacific example would be the Philippines’ conflict with China over the Scarborough Shoal. The US was simply unwilling to be chain-ganged into a potential conflict with China. The Philippines in turn has been forced to accept Chinese control and has sought a tighter relationship with Japan to compensate.

Conversely, the mechanism of the entrapment is recklessness by clients that chain-gangs patrons into conflicts against their wishes. Examples include: Austria-Hungary pulling Germany into World War I; North Korea pulling both the Soviet Union and Communist China into the Korean War; Cuba almost chain-ganging the USSR into the Cuban Missile Crisis; Abe’s Japan trying to pull the US today into its stand-off with China. In each case, a smaller, weaker state picks a fight with a larger opponent, confident that a larger ally, a patron, will back it up.

But even though the small ally picks a fight with an larger opponent, the patron also broadly shares the ‘enemy image’ of the opponent. That makes it easier for the patron to get sucked in. Some amount of the domestic level game in the patron’s politics has already been won by small state. It has allies in the patron’s foreign policy decision-making process who share its threat assessment and help entrapment happen. Israel and Iran is a really good example of that today. US neoconservatives and hawks have aligned with Benjamin Netanyahu against their own president (Obama).

But this entrapment dynamic, although sometimes used to explain the Japan-Korea relationship, does not actually fit well. Japan-Korea is better modeled this way: one ally (A) picking a fight with another ally (B), sticking the patron with an unwanted adjudication dilemma. Moral hazard captures that both are insured by the same patron (P). Because P’s elites do not view B as an enemy, P will not get chain-ganged in no matter how much A raises a fuss. Greece and Turkey in NATO are an example too, alongside Korea and Japan. Greece and Turkey, like Japan and Korea, have both sought to use their alliance with the US against the other, sticking the US with an unwanted adjudication it simply will not make.

So here is a possible theoretical addition to the alliance literature, with a nice empirical focus on Asia. This model captures Japan and Korea by focusing on ally-ally bickering rather than ally-opponent bickering. This carves out a third theoretical position, moral hazard, distinct from the traditional ideas of entrapment or abandonment.

This moral hazard model for Asia also opens a rich domestic level discussion about the patron’s friend/enemy images and threat assessments. The patron’s internal debate on A and B – which is a better ally and why – would drive how the patron responds to A and B’s bickering. So in the Japan-Korea case, the US has positive, almost equal friendship assessments of Korea and Japan, so the outcome is US strategic ambiguity, effectively freezing the conflict in place. In other words, the US will not choose Tokyo or Seoul, much the great frustration of both, and the conflict rolls on. But one could also foresee a situation like Greece and Turkey, where Greece is probably preferred, because Turkey is drifting toward Islamism under Erdogan. In that case the outcome would be an ‘informal tilt’ toward one ally over the other. But in neither dyad (Korea-Japan; Greece-Turkey) does ‘entrapment’ capture what is going on in the triangular relationship.

So I think this adds to our empirical understanding of Asia by explaining one of the most awkward (non)alliance outcomes in the region – Japan vs Korea. It also tightens the alliance literature through concept refinement/disaggregation in a major area of future relevance – Asia. Too much of the Asian alliance literature focuses on hub-and-spoke management and confrontation with China. But my idea captures inter-allied relations, which does not get nearly so much attention. The idea of ‘patron adjudication’ also strikes me as new. I cannot say that I have seen that idea before, but it seems reasonable to imagine allies competing for patron tilts. Clearly Korea’s current grand strategy debate springs from the perception that the US is tilting toward Japan.”

Print article

10 CommentsAdd yours

  • Chris Darnton - 2014-01-27

    Alliances are full of
    disharmony, but although that basic point has been clear for decades, we’re
    still trying to explain the roots of intra-alliance conflict and its
    resolution. Moral hazard might be one reason why a patron is reluctant to take
    sides among its client allies and resolve their dispute, but there are several
    others (low-grade conflict among clients might keep them competing to stay in
    the patron’s good graces, prior mediation efforts might have brought little but
    headaches, foreign policy attention might be diverted elsewhere onto the crisis
    of the week, etc.), and it’s not clear which goal is most important. This is
    doubly true because moral hazard arguments are primarily about behavioral
    incentives facing unitary actors in repeated interactions, but different decisionmakers
    within a country (whether patron or client) might have different perceptions
    and preferences regarding particular threats and partners in the region, and so
    any signals of restraint or encouragement have to filter through two-level
    games. You’re probably already familiar with Tim Crawford’s Pivotal Deterrence and Tom Christensen’s
    Worse than a Monolith; I’ll also plug
    my forthcoming book Rivalry and Alliance
    Politics in Cold War Latin America (Johns Hopkins, 2014:,
    which looks particularly at the issue of rapprochement among allies but also at
    the vicissitudes of patron adjudication that you mention. The parallels between
    Latin America and East Asia are hardly exact—for instance, you point out that
    Korea and Japan are really better understood as having parallel alliances with
    the US rather than any sort of nominal collective security agreement with one
    another, while Latin America has the Rio Treaty and other institutions—but hopefully
    relevant. Good luck with the project; I’d be happy to discuss this off-site as

  • Tim Crawford - 2014-01-27

    As Chris Darnton noted, see my book, Pivotal Deterrence, chaps. 5-6, which deals quite a bit with what you usefully describe as an “adjudication dilemma” in Greco-Turkish and India-Pakistan conflicts in the Cold War. Also, on Greece-Turkey, see Ron Kreb’s article in IO, “Perverse Instutionalism” (1999). –Tim Crawford

  • Robert Kelly - 2014-01-27

    Thank for all these suggestions, including the reading. I was thinking of moral hazard for the allies along the lines of the traditional insurance dilemma. Teenagers are guaranteed against the consequences of their reckless driving by their parents’ insurance, likely increasing teens’ risky driving. That was certainly the case when I first got my license.

    Similarly, both allies – Korea/Japam, Greece/Turkey – are insured against the consequences of their outlandish claims and behaviors about the other. As a result, there is no incentive to stop, to speak responsibly, to seek rapprochement. Like the parent guaranteeing the teen, so the US guarantees the infighting allies, against both each other, and external threats. So why back down? Why not just say anything you want? You face no costs, because the US is insuring your security. That strikes me as a good explanation of Korea and Japan at the moment.

    And yes, I would be happy to discuss this in an email chain. Thanks for offering:

  • Tim Crawford - 2014-01-27

    Also, as you probably know, Victor Cha’s Alignment Despite Antagonism, addresses some of the issues between Japan and South Korea that you highlight, esp. the tendency for closer US commitments to underwrite greater Japanese-South Korean tensions.

  • Robert Kelly - 2014-01-27

    Thank you. I think I should have emphasized more that I think ‘moral hazard’ is a good model, because it captures the irresponsibility or lack of restraint in the allies fostered by the patron. As I say below to Chris Darnton, as long as the US is around to guarantee your country, why no say anything? You don’t risk entrapping the patron if you are criticizing another ally, so the patron can’t tell you to shut, because you might start a war. So the patron is stuck with two allies who carry no costs from ripping into each other, so there’s no incentive to stop. I think this captures Korea and Japan today pretty well.

  • Robert Kelly - 2014-01-27

    Yes, thank you, but I don’t think he used the idea of moral hazard. That is what I was hoping was theoretically new – or at least hadn’t been ripped off from economics by someone else before me…

  • Tim Crawford - 2014-01-27

    I think you are onto something in your characterization of the moral hazard problem. I would expect however that the patron can exert some leverage over the others by holding out the prospect that it may lean in favor of one or the other’s claims in the intra-alliance conflict, especially if neither of them have a realistic exit option. (In the case at hand, can Tokyo or Seoul really opt for support from China if they strain their ties to the US?)

  • LFC - 2014-01-28

    “Conversely, the mechanism of the entrapment is recklessness by clients
    that chain-gangs patrons into conflicts against their wishes.”

    On this, you might want to see Dominic Tierney, “Does Chain-Ganging Cause the Outbreak of War?” ISQ 55:2 (June 2011) followed by a brief reply in the same issue by Christensen & Snyder. I don’t remember the Tierney article that well as I’m sure I more skimmed than read it, but he argues that WW1 is not an instance of chain-ganging. Putting that particular article to one side, many historians would disagree that A-H pulled Germany into WW1 “against its wishes” as you say in the post. I’m not even sure the most “pro-German” of the recent books on the outbreak of WW1, Clark’s The Sleepwalkers, goes that far. This is all somewhat removed from your main point in the post, but I’m just saying you shdn’t assume that the chain-ganging-caused-or-helped-cause-WW1 line is universally accepted — it isn’t.

  • Jake A. Douglas - 2014-01-28

    Over at the Cato Institute, Justin Logan’s recent study “China, the United States, and the Pivot to Asia” has a page-long discussion of “Moral Hazard in U.S. Commitments.” On a quick second reading, I don’t see mention of intra-alliance rivalry, though I swear I recall reading it from him.

  • Tim Crawford - 2014-01-28

    The connection(s) between Moral Hazard and Alliances are thoroughly unpacked in Brett Benson, Constructing International Security: Alliances, Deterrence, and Moral Hazard (2012).

Leave a Comment

Please complete required fields