Research Projects

This article develops a new theory of how a strong target regime can (partially) deter a sender from imposing sanctions over major issues. We develop a formal theoretic framework to show that, following a sender's sanction threat, a strong target regime can signal its strength by deliberately weakening itself through always sharing power with a domestic opposition, while a weak target regime mixes between sharing and not sharing power with the opposition. The strategy offers the sender a better assessment of a target's strength, enabling it to adjust its sanction policy accordingly and avoid the costs of a potentially ineffective sanction over a strong target regime. As a result, a strong target regime will be sanctioned with a lower probability than a weak one.

These days many governments employ scoring auctions---an auction format where bidders submit a pair of price and quality. Its goal is to achieve a high-quality procurement, even at the expense of a higher price. In this paper, we theoretically confirm this trade-off. We compare scoring auctions to more traditional auctions where the auctioneer first chooses a quality optimally, and then bidders submit prices. Our main result identifies conditions for which scoring auctions yield higher quality but at a higher price.

This paper studies the conduct of war. We introduce the idea that one of the adversaries has a single potentially decisive strategic or tactical option (one-time advantage) and they must decide whether and when to use it. We illustrate in a war of attrition setting with one-sided incomplete information that countries do not necessarily use their one-time advantage as soon as possible, and show the existence of equilibria where strong countries always withhold to signal strength. We differentiate between three cases and find that weak countries typically use the one-time advantage in the first period. The paper also demonstrates that in the presence of signaling incentives in military strategy, the battlefield might no longer serve as a reliable indicator of relative strength, which has implications for war duration. To illustrate the main logic of our model, we analyze Germany’s decision to employ unrestricted submarine warfare during World War I.

  This paper develops a context-specific game-theoretic model of reputation building in separatist conflicts and examines its central elimination mechanism empirically. In a two-period model in which a government may face two challengers sequentially, we identify political removal following military defeat as a novel mechanism that can reverse the government’s intertemporal incentives. When the conditional risk of elimination is sufficiently high, the normal government accommodates rather than investing in reputation for toughness, even under conditions that would otherwise sustain reputation-building conflicts. We examine this mechanism using movement-year data from the Self-Determination Movements (SDM) dataset and an original index of incumbent elimination vulnerability constructed from prior coup exposure, irregular leader exits, and leader tenure. The empirical analysis indicates that elimination vulnerability is more strongly associated with accommodation in denser multi-group environments. Although the theoretical model recovers Walter-style (2006a) comparative statics under particular conditions, the empirical illustration does not recover a broad, unconditional reputation-building pattern in the SDM data. Overall, the results show how the political consequences of military defeat can shape governments’ incentives to build a reputation for being tough.

 This article develops a formal model showing how an opponent can use limited military operations (LMOs) to infer a patron state's resolve to defend its client in an all-out conflict, and under what conditions such probing can paradoxically increase the risk of war. Specifically, LMOs can generate an outcome in which a highly resolved patron perfectly distinguishes itself by offering strong support in response to the LMO, while moderately resolved and low-resolved patrons fail to do so. By reallocating uncertainty toward these latter two types, LMOs can sharpen beliefs in a way that increases the risk of war relative to a world without LMOs. The model thereby identifies a counterintuitive source of escalation: wars are most likely to occur not when a patron offers strong support to a client resisting an LMO, but when it offers moderate support.