In the National Interest (online), Amitai Etzioni argues that, in Afghanistan, the United States should avoid building out the central state, and instead co-opt militia leaders, including the Taliban.

Amen. Americans, even those writing counterinsurgency manuals, conflate counterinsurgency and state-building.

Both presidential candidates’ plans for Afghanistan share this failing. They both support a surge of troops and effort in Afghanistan based explicitly on the idea that our objective should be to build the Afghan state to win the loyalty of the people in the insurgent, meaning Pashtun, areas.

A better plan is a rough replication of what we did in Iraq’s Anbar province. There, we paid off the main body of insurgents and allowed them to rule in their area, provided that they turned against Al Qaeda in Iraq. Don’t believe the myth of the surge. This tactic, which pre-dated General David Petraeus’ assumption of command and had nothing to do with higher troop levels, was the main cause of the pacification of Iraq’s Sunni regions.

You could call this counterinsurgency strategy “appeasement” or “state-breaking,” as opposed to state-building. Having bought peace by dividing authority, there is no obvious way to put Humpty Dumpty back together — the dilemma we now face in Iraq. But if (a big if in Iraq) the division of power is remotely stable, that is not necessarily a problem, at least from a US perspective. You prioritize; sacrificing cohesive central authority for counterterrorism and rough stability.

This strategy stems from the idea that the trouble is the central state itself. You limit its sphere and leave the insurgency, essentially an alternative state, to its. Doing so is only possible where the insurgency has limited geographic orbit and ambitions — a common condition in divided societies with weak governments. Saddam Hussein himself employed this tactic late in his rule, as Austin Long explains in Survival.

In Afghanistan, where power has always been decentralized, the state-breaking strategy has more obvious merit than Iraq. Extending central governance is to undertake a struggle of indeterminate length, which is likely to fail at tremendous expense, while feeding the insurgency by antagonizing the Pashtun population. As Nir Rosen’s informative Rolling Stone article points out, the Afghan state reaches many Afghans not through the provision of services but via predatory national police. Our effort in Afghanistan, with its limited ambitions and reliance on local powers, has always had an element of this strategy, rhetoric to the contrary notwithstanding. But Washington’s embrace of the idea that we have neglected state-building in Afghanistan in favor of Iraq threatens to change this.

Our enemy in Afghanistan is really three insurgencies, and even the main body of the Taliban is really a loose-knit group of militias. Summit-style meetings with purported Taliban leaders will not do the trick. Deals with Taliban commanders will be incremental.

Sooner or later, the United States needs to leave Afghanistan. The idea that we can only do so once it is a centralized, peaceful country that collects taxes and provides services throughout its territory is a recipe for staying forever. We invaded Afghanistan to deny anti-American terrorists haven and deter anyone from offering it. We can maintain those conditions without a strong central goverment, and therefore without a perpetual occupation, if we do something like what Etzioni suggests.