Charli linked to a great round-up of theories circulating that propose to answer the rather interesting question of why countries that sign the Convention Against Torture seem to have a greater likelihood of committing torture.

One working paper in particular , by James Hollyer and Peter Rosendorff of NYU, has caused quite a stir. They propose that dictators use the signing of such a treaty as a costly signal to domestic opposition groups that they fully intend to continue torturing those that oppose their regime. How does this work?

We argue that authoritarian states ratify human rights treaties explicitly because they do not intend to comply. And it is important to those signatories that all observers understand that they have no intention of complying at the time of accession. The logic, while counterintuitive, is straightforward: an elite facing threats from a domestic opposition can mitigate these threats by engaging in torture. If there is any additional cost to the elite of signing and then being found to torture, the act of signing the agreement signals to the opposition the strength of the elite’s commitment to remaining in power. Accession is a signal to the opposition of the very high value the elite places on holding onto power and its willingness to use torture if necessary. On observing the government’s accession, the opposition – now better informed about the value the elite places on holding power – will rationally reduce its anti-regime activities. The government continues to torture, but will torture less. On the other hand a regime that doesn’t sign shows itself to be vulnerable to the added costs associated with the use of torture. Thus, the opposition will increase e orts to remove the regime on seeing that the government does not sign.

It’s an intriguing theory, and the authors do a good job of linking their claims to empirical evidence. However, I am skeptical regarding their proposed mechanisms and whether they actually obtain:

1) If it is well known that the conversion rate, if you will, for bringing accused torturers to justice in connection with the CAT is quite low, then why would a domestic audience see this as a credible signal? Any dictator could sign the treaty regardless of whether their type was of a moderate or an extreme–if the signal could easily be sent by either type it can’t differentiate. Additionally, the more credible signal is one that actually demonstrates the will and, more importantly, the capability of the sender. In this case, if a dictator’s real aim and desire is to signal that they will do whatever it takes to stay in power why not just make an example of revolutionaries, rebels, etc? Show that you have the will and capability to do whatever it takes to stay in power (thinking here of Barbara Walter’s work on why some states negotiate with separatist groups and others choose violent repression).

2) Given that, I am more inclined to see it as a low-cost, public relations move to placate domestic and international critics. By signing the CAT a dictator can point to his/her efforts to play by the same rules as other governments and to treat their citizens humanly. The next time they are getting reamed out at some summit or UN meeting they can say “yes, but, we did sign the CAT”. They don’t really need to send a credible signal with the move, just create a useful tool in their public relations arsenal.

3) Additionally, violent, repressive dictators are less likely to fall from power. Therefore, they are less likely to be placed in a position where they could be prosecuted for their actions under the CAT. Combine this fact with the potential usefulness of signing the CAT from a public relations standpoint and you have another potential explanation aside from signaling. Moreover, it would also explain why ‘moderate’ or ‘competitive’ dictatorships are less likely to sign the CAT–precisely because their are more vulnerable to losing power and could therefore be brought to justice under the CAT. Given that the public relations gains are modest compared to the potential costs of actually being prosecuted, moderate dictators would be less likely to sign.

Just my initial thoughts. In general, the data is quite intriguing as is Hollyer and Rosendorff’s theory–certainly a puzzle well worth exploring.