Splitting the difference is no way to win a war.
If you clicked on a link to read this column, you were likely influenced by a combination of factors: the title (often misleading or sensationalist), an accompanying picture (I bet someone in a uniform), where it appeared on the FP homepage (top is better than bottom), or how it was described in a tweet (short and pithy helps). Advertisers, psychologists, and behavioral economists have long been aware that the social construction of options influences both the choices of decision-makers and their outcomes. Because it is impossible to make decisions in a vacuum, free from external influences, we rely on shortcuts like recommendations from friends, historical precedents, name recognition, or simply cool pictures.
How choices are framed for decision-makers can have negative or positive repercussions — assuming that the person has some ranking of values or preferences. In their landmark 1981 paper, "The Framing of Decisions and the Psychology of Choice," psychologists Amos Tversky and Daniel Kahneman presented social science experiments that found people reversed their preferences based on manipulative "variations in the framing of acts, contingencies, and outcomes." In the book Nudge: Improving Decisions about Health, Wealth, and Happiness, Richard Thaler and Cass Sunstein endorsed the role of "choice architects," engaged in libertarian paternalism that preference certain decisions by, for example, arranging healthy foods prominently at eye level in a school cafeteria to make it more likely children choose to eat them.
And in one memorable scene from The Simpsons movie, Environmental Protection Agency chief Russ Cargill lays out five folders (labeled one to five) on the desk of "President Schwarzenegger" in the Oval Office. He tells the commander-in-chief: "Well, I’ve narrowed your choices down to five unthinkable options."
Schwarzenegger: Ok, I pick 3!
Cargill: Try again.
Cargill: Go higher.
Cargill: Too high.
Cargill: You already said 3.
Cargill: There is no 6.
Cargill: Double it.
Cargill: As you wish, sir.
Although national security discussions are (hopefully) taken more seriously than the Simpsons would convey, such cartoon satire does have a kernel of truth in how decisions are framed to the president and/or secretary of defense — collectively, the national command authority permitted by U.S. law to authorize the use of military force. Similar to everyday choices like which brand of cereal to buy, how senior military officials present and characterize military options strongly influences policymakers’ decisions.
Recent history is rife with examples of this phenomenon. According to David Halberstam, in the George H.W. Bush and early Clinton administrations, chairman of the Joint Chiefs of Staff General Colin Powell "would show his lack of enthusiasm by giving [senior civilians] a high estimate, and they would quickly back off. The figure never went under two-hundred thousand troops." Largely as a result of how the choices were framed, the proposed massive ground presence was never seriously considered by civilian officials, who summarily dismissed it as "the usual two-division, $2 billion option."
In August 1992, the Senate Armed Services Committee requested testimony from a senior official from the Joint Chiefs of Staff at a hearing on Bosnia. There was zero appetite among the Joint Chiefs to get involved militarily in Bosnia, and Powell directed his principal military assistant, Lieutenant General Barry McCaffrey, to convey this to the committee. To prepare, McCaffrey drafted his presentation and slides sans approval from the Office of the Secretary of Defense or the White House. When committee senators asked what force level would be required to end the violence in Bosnia, McCaffrey replied, "It would be around 400,000 troops." According to McCaffrey, Powell later asked, "Where the f#&k did you get those numbers?" He replied, "I made them up, based on my understanding of the parties and the situation." Nevertheless, the anchoring effect of the highball estimate ultimately succeeded in shelving the ground presence option.
In fall 1998, the White House tasked the Pentagon with developing options for halting the Serbian ethnic cleansing in Kosovo. In return, the Pentagon estimated a requisite deployment of 175,000 to 200,000 NATO troops. As a senior administration official stated: "The numbers came in high. No one said yes, no one said no; it was taken off the table…It was a complete eye-roller." When the White House reconsidered ground options in Serbia after three months of air war, the initial proposal effectively prevented any serious debate, since "there was only one option by then that the Joint Chiefs would support."
At the same time, during Clinton administration debates over whether to capture or kill Osama bin Laden, senior civilian officials believed that military leaders were intentionally oversizing the options on the table. National Security Adviser Sandy Berger stated flatly: "[The military] didn’t want to do it [attack al Qaeda]…There was just no enthusiasm and creativity." His deputy, James Steinberg, recalled that the civilian advisors "were not at all happy with the military’s options for going after Bin Laden." Meanwhile, Vice Admiral Scott Fry, director of operations for the Joint Chiefs of Staff from 1998 to 2000 and the man who often presented such options, argued, "We could never impress upon the civilians in the National Security Council how far Afghanistan was from a staging base or carrier group."
Prior to the U.S. invasion of Iraq in March 2003, debates over force levels spilled over into the public sphere. After Army Chief of Staff General Eric Shinseki testified that a post-war occupying force would require "something on the order of several hundred thousand soldiers," Deputy Secretary of Defense Paul Wolfowitz immediately dismissed the estimate as "wildly off the mark." Campaign planners found subsequent proposed options silently rejected by Secretary of Defense Donald Rumsfeld with no explanation or "force caps" (a ceiling for permitted ground troops) beyond "rework the plan." As a result, the size of the U.S. ground invasion of Iraq shrank substantially and — combined with inadequate post-conflict stabilization planning — ultimately played a role in the emergence of the insurgency.
In 2005, CIA and special operations teams in northern Pakistan developed intelligence that provided "80 percent confidence" about the future location of senior members of al Qaeda, including Ayman al-Zawahiri. Over time, the plan to kill or capture these militants grew to include somewhere between 150 to several hundred special operators and CIA operatives. Secretary of Defense Donald Rumsfeld called off the raid at the last moment because of its unwieldy size and less than 100 percent guaranteed actionable intelligence.
(Six years later, President Obama would authorize the raid that killed Osama bin Laden with fewer U.S. special operators (79 and a dog), less certainty that the al Qaeda leader would be in Abbottabad ("50-50"), and significantly farther into Pakistani territory (100 miles). However, although Obama is lauded for authorizing the operation, it was probably less politically risky than Rumsfeld’s decision, largely because Special Operations Command (SOCOM) had dramatically improved and routinized such raids. As Admiral William McRaven described the killing of bin Laden: "We did 11 other raids much like that in Afghanistan that night. From a military standpoint, this was a standard raid and really not very sexy.")
In fall 2009, over the course of three months, the Obama administration debated how many additional U.S. troops to deploy as part of an Afghanistan "surge." The military reportedly developed options between 40,000 and 80,000 troops, as the Joint Chiefs of Staff, regional combatant commander General David Petraeus, and commander of U.S. forces in Afghanistan General Stanley McChrystal coalesced around the lower end of the spectrum. Vice President Joseph Biden and Chief of Staff Rahm Emmanuel sought a smaller and cheaper option of no more than 20,000 troops focused on conducting counterterrorism missions. According to journalist Steve Luxenburg: "In the end, Obama essentially designed his own strategy for the 30,000 troops, which some aides considered a compromise." When military officials repeatedly attempted to add forces, much to the president’s frustration ("Why do we keep having these meetings?" he asked an aide), he personally wrote a six-page "terms sheet" that limited the size and scope of U.S. forces and operations in Afghanistan.
As many of these examples suggest, civilian officials are often deeply ignorant of the factors that limit military effectiveness like geography, logistical demands, actionable intelligence, or adversarial strategy. Meanwhile, the perspective of many senior uniformed officials was best described by Richard Betts thirty-five years ago: "The military’s natural professional impulse is toward worst-case contingency planning for any conceivable disaster." One way to limit worst-case contingencies is to "plus-up" the force package. "War is not a game to be won 6 to 5. You want to win 21 to 0," as another director of operations on the Joint Staff told me in an interview.
In his excellent book, Armed Servants: Agency, Oversight, and Civil-Military Relations, Peter Feaver analyzed the principal-agent problem, which argues that civilian principals have the authority and military agents have the expertise. The belief that military planners provide inflated estimates of the forces required to achieve an objective is an example of what Feaver terms "shirking" — something that undermines the ability of civilians to make future decisions. This includes generals dissuading the active consideration of certain options or intimidating mid-level civilians under the guise of merely offering their "best military advice."
To bridge the gap between relatively uninformed civilian principals and expert military agents, trust must be built through the back-and-forth bargaining process that develops and hones military options. When senior military officials perceive that their civilian bosses are unfocused, unserious, or unaware of the costs associated with using force, they are less likely to develop creative (and risky) options, resulting in "oversized options" that render any use of force politically impossible. Similarly, when civilian officials sense that the military is reticent or inflexible, or refusing to accept that the White House wants to "do something" — assuming reasonable risks and costs — military planning suffers. While there is no "correct" number for the forces required for a campaign or mission, it is only through trust and dialogue that strategy — the combination of military and non-military means to achieve an outcome — are generally successful.
The framing of military options remains particularly relevant today as the United States considers what force levels to keep in Afghanistan after 2014 — assuming permission from Kabul. As the Wall Street Journal reported this week, U.S. officials noted that "preliminary military recommendations" include maintaining "6,000 to 15,000 troops for training and counterterrorism missions" in Afghanistan, and that "senior U.S. officials said they expected the president to authorize up to 10,000" — roughly splitting the difference, just as he did three years ago.