I've just returned from the second part of a major irregular warfare wargame. Interestingly, the most pressing finding in this game (and others I've participated in) is that the most significant shortfall when the United States provides counterinsurgency support or undertakes SSTR is our inability to provide advice and support to the national police in a partner state. Everyone agrees that the key to stabilization and counterinsurgency are some kind of constabulary or gendarmeries that are more robust than local police, but something less than the military. But because we ourselves don't have such an organization (particularly a deployable one), we are ill-equipped to provide advice and support to partners who might want to develop one.

Several solutions are normally discussed for this problem. All have shortcomings:

1. Use existing military resources, particularly military police, to provide such advice and support. The problem is that American MPs are not trained, educated, and organized to be a national police.
2. Task the State Department to develop such a capability. The problem with this is that they have nothing like the requisite expertise or institutional interest.
3. Create a new organization within the Department of Defense to do this. I kind of like this one, but it would be expensive.
4. "Outsource" it, either to contractors or other nations that do have such a national police capability (France, Italy, Canada, etc). The problem with this, of course, is that contracting it out does not have a very good track record, and other states may not have the capability or the will to undertake the job at a necessary level.
5. Task the Department of Homeland Security to develop an expeditionary law enforcement capability. The problem here is that, say, pulling American border guards out of Arizona to go help stabilize another state and train its national police would be politically difficult, maybe impossible.

Ultimately, it's a big problem for which all solutions are flawed.