Military commanders deciding how toreposition and withdraw U.S. forces in Afghanistan are now confronting decisionswhere mistakes could doom the war effort. NATO has achieved considerablesuccess after brutal combat in the southern Afghan provinces of Helmand andKandahar. Violent incident rates are running in excess of 20 percent lower than during a comparable time period in 2010. Local Afghan government, for all its weakness,is expanding into districts that were long controlled by the Taliban. Even recruitment of southern Pashtuns into local police and security forces is going up in some districts that last year saw the local population watching passively while Americans fought Taliban insurgents. While such recruitment for the army is still well below what is desired, there are scattered reports, both in print and from local sources, of larger numbers of Pashtuns joining local police forces.

However, violence in eastern Afghanistan is not dropping. The threat from the Haqqani  Network forces supported from bases in Pakistan has increased, as have Haqqani Network-originated spectacular attacks aimed at Kabul. The combination of apparent success in the south combined with the threat of eastern violence produce a strong argument for rapidly shifting forces east to mount a major campaign there before hard deadlines for U.S.troop withdrawals next year diminish the offensive power international forces can wield. Yet this military logic conceals critical political risks that deserve close attention.

The southern surge was intended to create conditions that would allow Afghan forces to take over population security and the expansion of governance and development. Claims by some serving and retiredAmerican generals that security gains in southern Afghanistan are irreversible seriously overstate the situation. None of my many Afghan contacts fully accept this view. Only half the mission is accomplished. Afghan Army performance has improved, and army units are supported by some police and a few small units of local village defense of varying political and military reliability. But nowhere have Afghan forces yet stood largely on their own.  Their ability to do so remains an unproven theory, not an established fact. 

The decisive battles for the south have also yet to begin. They will occur as U.S. forces thin out and insurgents try to regain control of the population. The Taliban's inability to confront Afghan security forces backed up by residual U.S. and NATO forces will not be the measure of success. Rather, success will only come when Afghan forces have the ability to maintain security for assistance workers, Afghan civil servants, and tribal leaders who have returned to their districts and cooperated with us and their government. All of them will be targeted by the Taliban, using threats and assassinations to intimidate others who might be on the fence. The struggle for control of the population will be the decisive battle.

The battlefield will also be psychological as much as physical. After 30 years of war, Afghans have a high pain threshold. If they believe they are on a winning side, they can and will put up with sacrifice, and replace assassinated officials. But if they become convinced that overall security is declining then we will again see local officials deserting their posts or living ineffectually in protected compounds.Tribal leaders will again flee to the cities. Confidence that has been slowly built in the south will be quickly destroyed. Worst of all, the word will spread rapidly that those who put their faith in improved Western- and Afghan-created security are taking suicidal risks, especially with the impending NATO force reductions. If this message goes forth it will undercut any possible military gains from repositioning forces.

These dangers do not mean that no forces can be withdrawn or repositioned. Indeed, transfer of control must occur if thewhole strategy of Afghan forces taking over responsibility for security by 2014 is to achieve credibility. Further, it is important to expand secure areas in the east and diminish the threats to Kabul. The need for troops in both the east and the south is real.  The time to make decisions has been reduced by President Obama's accelerated withdrawal schedule for 2012. Risk, as my military colleagues always remind me, must be taken somewhere. The point is that the risks must be considered in political and psychological terms far more than on a strictly military basis if we are not to waste the major gains of the last two years. Such considerations demand great prudence in two areas that must be worked out by our civilian and military leaders on the ground.

One is that turn over must be undertaken slowly enough that Afghan security forces can be tested when we still have the ability to correct after setbacks. War is a hard school taught by a capable and reactive enemy. There will be bad days. We must ensure that we retain the margin to work with our Afghan allies to rebound from problems, not let them be shattered by them.

Second, risk must be decided jointly with Afghan civilian and military authorities. They bear the ultimate cost of failure, and their confidence in the possibility of success is crucial to strategic credibility and their willingness to take the losses required to succeed. More progress has been made in the last two years than many Americans recognize. It must be solidified before it is excessively risked.

RonaldE. Neumann was U.S. Ambassador to Afghanistan from 2005 until 2007 and hasvisited regularly since. He is author ofThe Other War; Winning and Losing inAfghanistan.