As much as I despise PowerPoint (and believe me, I do), I always think back to a presentation made by the late Captain Travis Patriquin which touches upon this very issue. Patriquin observed, in 2005, that Iraqi security forces were generally deployed far from their homes and knew little about their areas of operation. The cultural alienation the Iraqi Police and Army felt must pale in comparison to what Tajik police sergeants are facing in Pashtun territory, where ethnic groups do not even share a common language.
The police also said that establishing connections with residents had been difficult. Part of their problem, they said, was that many sergeants are Tajik, and do not speak Pashto, southern Afghanistan’s dominant language.
“Nobody can find a lot about the Taliban,” said Sgt. Akhmad Fahim from Mazar-i-Sharif.
Patriquin concluded that only local security forces operating on their own turf, such as the Sons of Iraq program, had the local knowledge to effectively hunt down insurgent groups like al Qaeda in Iraq.
So why not try the same thing in Afghanistan? Well, it's a bit trickier in Afghanistan. As a recent study noted, communities were home-grown security forces were successful--such as Anbar province in Iraq, or Arghandab in Afghanistan--are relatively ethnically homogeneous. The same study also suggests that these sorts of initiatives may prove a competitor to the local Afghan Police Force:
So we can't trust the police or concerned local citizens' groups? Tell me how this ends...
[P]rograms like the ANAP and AP3 at times in fact discourage recruits from joining the ANP by providing all the benefits and fewer of the problems. The command and control relationship between the ANP and these armed groups posed problems in all three of the programs examined here. ANP control was clearly spelled out in theory but realities on the ground are more competitive.