This lengthy and well-written piece by Dexter Filkins in The New Yorker is must reading.

Filkins attempts to give an overall assessment of Afghanistan’s future as American forces shrink and the country must increasingly rely upon the Afghan National Army for its security.   By all means, read the entire article, but the gist of Filkins’ assessment can be succinctly summarized as, “bleak.”

Filkins explores the question of whether Afghanistan is destined to return to a state of civil war as American combat troops leave.   While he makes no firm conclusions, the answer is an inescapable “yes.”   While American policymakers and military leaders boldly talk up the prospects for turning over security responsibilities to the ANA while hoping to garner a power-sharing deal with the Taliban, it is clear from the article that a debacle of enormous proportions is looming in 2014 (if not before) when American force levels are expected to drop to just fifteen thousand from an expected sixty-eight thousand after the September 2012 draw-down.

Some Western and Afghan experts say that fifteen thousand American troops would not be enough to secure Afghanistan, particularly when it comes to the use of airpower. The Afghan Air Force is far less advanced than the Soviet-trained force was at a similar moment. American officers told me that air strikes—bombs and rockets—are usually restricted to units in which Americans direct the fire. A force of fifteen thousand Americans would probably not be large enough to spread trainers and air controllers throughout the Afghan Army (and not throughout the police, who are at tiny checkpoints scattered around the country). “If they go below thirty thousand, it will be difficult for them to do any serious mentoring, and without the mentors they won’t call in airpower,” Giustozzi, the Italian researcher, said.

American officers have another concern. Currently, Afghan units are stationed where the Americans are, in hundreds of small bases, mostly in populated areas. Some American officers say that the Afghans will find it difficult to disperse themselves as fully, because of problems with supplies and communications. Once the coalition forces leave, those officers say, the Afghans are likely to consolidate their units on bigger and fewer bases. If that happens, the Afghans could end up ceding large tracts of territory to the Taliban—much as the Afghan Army did after 1989.

Filkins interviews several Afghans in and outside of the Afghan government who all candidly admit that each of the rival factions that fought each other prior to the 2001 American invasion are actively preparing for the resumption of civil war when U.S. forces leave.

Afghan and American officials believe that some precipitating event could prompt the country’s ethnic minorities to fall back into their enclaves in northern Afghanistan, taking large chunks of the Army and police forces with them. Another concern is that Jamiat officers within the Afghan Army could indeed try to mount a coup against Karzai or a successor. The most likely trigger for a coup, these officials say, would be a peace deal with the Taliban that would bring them into the government or even into the Army itself. Tajiks and other ethnic minorities would find this intolerable. Another scenario would most likely unfold after 2014: a series of dramatic military advances by the Taliban after the American pullout.

“A coup is one of the big possibilities—a coup or civil war,” a former American official who was based in Kabul and has since left the country told me. “It’s clear that the main factions assume that civil war is a possibility and they are hedging their bets. And, of course, once people assume that civil war is going to happen then that can sometimes be a self-fulfilling prophecy.”

One Afghan, Abdul Nasir, made this point quite clear:

These days, Nasir said, the nineties are very much on his mind. The announced departure of American and NATO combat troops has convinced him and his friends that the civil war, suspended but never settled, is on the verge of resuming. “Everyone is preparing,” he said. “It will be bloodier and longer than before, street to street. This time, everyone has more guns, more to lose. It will be the same groups, the same commanders.” Hezb-e-Wahdat and Jamiat-e-Islami and Hezb-e-Islami and Junbish—all now political parties—are rearming. The Afghan Army is unlikely to be able to restore order as it did in the time of Najibullah. “It’s a joke,” Nasir said. “I’ve worked with the Afghan Army. They get tired making TV commercials!”

A few weeks ago, Nasir returned to Deh Afghanan. The Taliban were back, practically ignored by U.S. forces in the area. “The Americans have a big base there, and they never go out,” he said. “And, only four kilometres from the front gate, the Taliban control everything. You can see them carrying their weapons.” On a drive to Jalrez, a town a little farther west, Nasir was stopped at ten Taliban checkpoints. “How can you expect me to be optimistic?” he said. “Everyone is getting ready for 2014.”

In the process of his interviews, however, Filkins did discover one, unsettling truth:  the most effective force against the Taliban so far have been local militias.

The most effective weapon against the Taliban were people like Mohammad Omar, the commander of a local militia. In late 2008, Omar was asked by agents with the National Directorate of Security (N.D.S.)—the Afghan intelligence agency––if he could raise a militia. It wasn’t hard to do. Omar’s brother Habibullah had been a lieutenant for Abdul Rasul Sayyaf, one of the leading commanders in the war against the Soviets, and a warlord who helped destroy Kabul during the civil war. The Taliban had killed Habibullah in 1999, and Omar jumped at the opportunity to take revenge. Using his brother’s old contacts, he raised an army of volunteers from around Khanabad and began attacking the Taliban. He set up forces in a string of villages on the southern bank of the Khanabad River. “We pushed all the Taliban out,” he told me.

The Taliban are gone from Khanabad now, but Omar and his fighters are not. Indeed, Omar’s militia appears to be the only effective government on the south side of the Khanabad River. “Without Omar, we could never defeat the Taliban,” a local police chief, Mohammad Sharif, said. “I’ve got two hundred men. Omar has four thousand.”

The N.D.S. and American Special Forces have set up armed neighborhood groups like Omar’s across Afghanistan. Some groups, like the Afghanistan Local Police, have official supervision, but others, like Omar’s, are on their own. Omar insists that he and his men are not being paid by either the Americans or the Afghan government, but he appears to enjoy the support of both. His stack of business cards includes that of Brigadier General Edward Reeder, an American in charge of Special Forces in Afghanistan in 2009, when the Americans began counterattacking in Kunduz.

This is a strange twist in U.S. strategy.  While the State Department, the White House and U.S. commanders in Afghanistan all blather about the progress of the ANA and the expected success of the transition to Afghan security forces, the American Special Forces seem to be busy setting up militias all over Afghanistan, perhaps in the grim realization that these militias are the only native force capable of actually eradicating the Taliban.

This would be welcome news indeed, if it is true, because it finally faces the truth that Afghanistan today simply cannot function as a modern, centrally governed nation state.   This is not to say that it never has in the past (clearly it has) nor that it will never function as one in the future, but only that the combination of religious fanaticism, the interference of Pakistan (and Iran to some extent), the ethnic divisions and the drug trade militate in favor of local control.    And this seems to be the main trade-off where militias effectively keep out the Taliban but bring their own set of problems:

Kunduz Province is divided into fiefdoms, each controlled by one of the new militias. In Khanabad district alone, I counted nine armed groups. Omar’s is among the biggest; another is led by a rival, on the northern bank of the Khanabad River, named Mir Alam. Like Omar, Alam was a commander during the civil war. He was a member of Jamiat-e-Islami. Alam and his men, who declined to speak to me, are said to be paid by the Afghan government.

As in the nineties, the militias around Kunduz have begun fighting each other for territory. They also steal, tax, and rape. “I have to give ten per cent of my crops to Mir Alam’s men,” a villager named Mohammad Omar said. (He is unrelated to the militia commander.) “That is the only tax I pay. The government is not strong enough to collect taxes.” When I accompanied the warlord Omar to Jannat Bagh, one of the villages under his control, his fighters told me that Mir Alam’s men were just a few hundred yards away. “We fight them whenever they try to move into our village,” one of Omar’s men said.

U.S. policy in Afghanistan, then, must take a hard look  at our national interests.   The primary, national interest for the United States in Afghanistan is to ensure that international terrorists do not find safe havens from which to plot and launch attacks against U.S. interests.   If militias will ensure that their territory will not be used for Islamist terror activities, that is enough.   We may be able to exercise some leverage over warlords and militias by doling out more arms and money to those who refrain from humanitarian abuses, but it is not in our national interest to force-feed an entire nation on Western morality and values as we have done for the last 11 years.

In fact, some Afghans appear to be leaning in the direction of a decentralized approach:

One political change that might prevent civil war, some opposition leaders say, would be the imposition of a federal system in which power would devolve to the provinces. Such a move could essentially cede dominion to the Taliban in the south and the east but protect the rest of the country. In 2004, when the new Afghan constitution was ratified, under American supervision, the central government, in Kabul, was given extraordinary powers, including the right to appoint local officials. The hope then was that a strong central government would unite the country.

If a federal system were to be adopted, some Afghan leaders say, it might matter less to the Tajiks and other minorities if the Taliban were allowed to govern Pashtun provinces in the south and the east. (How it would matter to the Pashtuns, and particularly to Pashtun women, isn’t much discussed.) As it is, many of the most prominent leaders of Afghanistan’s minority groups appear to be preparing for civil war.

While I disagree that the U.S. needs to “cede dominion to the Taliban in the south and the east” (there is no reason to think that Pashtun militias could not be set up in these regions as in other areas), the fact remains that Afghanistan is headed for division one way or another.  Current U.S. policy is wishful thinking and a criminal waste of American lives and treasure.

As a parting thought, it is even possible that by empowering local militias and tribes in this fashion, the U.S. may be able to deal a severe if not fatal blow to the Islamists across the border in Pakistan.  It is axiomatic that insurgencies work in both directions.   If Pakistan can support, for example, the Haqqanis in infiltrating into Afghanistan, so, too, can the U.S. support Pashtun militias on the Afghan side of the border to infiltrate and take away territory from Islamists in the FATA.   This provides U.S. policymakers with a unique lever in the tense relations with Pakistan.   This approach also allows a dramatically smaller footprint for U.S. forces in Afghanistan, enough one would expect, to deprive Pakistan of its logistical choke-hold.

If there is a new Administration in 2013, a new approach in Afghanistan is at least possible.