A decade or so ago Oona Hathaway and others noted that states that ratified the Convention Against Torture (CAT) appeared to be more likely to practice torture than states that did not sign this treaty. Ever since, scholars have been trying to make sense of this, especially since ratifying the CAT is somewhat costly (compared to other global human rights treaties).
A first set of explanations suggested that dictators signed the treaty to gain reputation or win aid concessions from Western Liberal democracies. Beth Simmons and others pointed out, however, that it wouldn’t make much sense for Western states to reward such obviously insincere behavior and that there is no evidence that authoritarian states actually get goodies for signing human rights treaties. This suggests that we should look inside autocracies. For example, my colleague Jim Vreeland’s article in International Organization argues that more pluralistic authocracies are more likely to exhibit pressures for signing treaties but also have more dissent, which provides incentives for torture.
James Hollyer and Peter Rosendorff have a new working paper that takes this quite a few steps further. They argue that dictatorships use signing the CAT as a costly signal to domestic opposition groups that they will continue to employ repressive tactics to stay in power (paper available from PEIO). Truly “bad ass” dictators are willing to pay the cost that CAT brings with it. So, these regimes sign precisely because they have no intention to comply and because everyone knows this. The model has an interesting set of empirical implications, such as that worse torturers are more likely to sign, that the CAT does help to reduce torture in those states but that the CAT also helps these dictators to stay in power longer.
I am oscillating between admiration for the sheer ingenuity of the theory and disbelief at the sheer implausibility of it all. Wouldn’t bad ass dictators have better technologies for committing to their bad-assness than signing treaties? Hollyer and Rosendorf do offer a number of anecdotes to support their mechanisms and the empirical implications they get out of the model are really interesting and largely consistent with the evidence. They are also upfront that this is not necessarily the only mechanism that explains CAT ratification. In any case, a really interesting and novel contribution to this debate that is well worth a read.
Update The Economist has picked up on this.