Can Anti-Karzai Rebels Maintain Unity?

One analytic issue that I think hasn’t gotten enough attention is whether there’s any realistic chance that the anti-Karzai forces in Afghanistan could maintain unity if they keep making advances. Traditionally what’s happened in Afghanistan is that factions with their backs up against the wall find ways to work together, find strength in unity, start making gains, and then as they do better their alliances start falling apart. Read this 1989 New Republic article from Robert Wright about Afghanistan:
One sign of how little serious thought the Reagan administration gave to an outright rebel victory is how it spent its money. A disproportionate amount of the arms we’ve distributed among the seven rebel factions has gone to the very last guy we should want to lead a post-victory Afghan government: a murderous anti- American fundamentalist named Gulbuddin Hekmatyar, who heads the Hezb-i-Islami (Party of Islam). Hekmatyar most recently made news in July, when his fighters slaughtered 36 members of a rival mujaheddin faction and tortured a few of them by pouring gunpowder in their eyes and lighting it. The alibi for our strong support of Hekmatyar is that it was a concession to Pakistan’s Interservices Intelligence agency (Pakistan’s CIA), which serves as the conduit for rebel aid and has its own reasons for liking him. True enough. Still, had the intentions of our policy-makrrs extended much beyond inflicting prolonged misery on the Russians, we probably could have divvied up the money with more foresight.

The recent tendency of freedom fighters to murder and torture one another helps explain why the collapse of the Najibullah regime–once assumed to follow automatically from a Russian pullout–hasn’t happened and doesn’t appear imminent. What little unity the rebels ever had was a product of three things: (a) a hated foreign intruder (the Red Army); (b) a hated puppet dictator (most recently Najibullah); and (c) the fact that Western aid, funneled through a shell rebel coalition based in Pakistan, was made contingent on a show of unity. Now (a) the Russians are gone and (b) Najibullah has launched a kinder-gentler offensive that includes being nice to his urbanite subjects and offering peace and large chunks of turf to rebel commanders who will Ieave him alone. That leaves (c), which doesn’t seem to be doing the trick.
Today, Karzai is in office and backed by American forces and money. And Gulbuddin Hekmatyar’s forces are cooperating with the Taliban against Karzai. And Hekmatyar still has close ties to the ISI. If our troops weren’t in Afghanistan, would Hekmatyar keep working with the Taliban against Karzai in hopes of becoming second-banana in a Taliban-dominated Afghanistan? Or would Karzai find himself facing new incentives to strike a deal with Hekmatyar, and would Hekmatyar repeat his life-long pattern of stabbing his once-allies in the back in hopes of seizing more for himself?
To put the point in a more general way, we often seem to be assuming that the equilibrium state in Afghanistan is a stable, centralized government. Thus either Karzai and the United States will win by creating a stable, centralized, Karzai-run state or else the Taliban will win by creating a stable, centralized, Taliban-dominated state. The evidence, however, suggests that the equilibrium state is for Afghanistan to be run by rival gangs of warlords configured in a complicated panorama of shifting alliances. Any “winning” coalition is likely to collapse under the weight of victory.