全球智库信息集成服务系统

Most modern presidents have found that the transition from campaigning to governing presents a unique set of challenges, especially regarding their newfound national security responsibilities. Regardless of their party affiliation or preferred diplomatic priorities, presidents have invariably come to appreciate that they cannot afford to make foreign policy decisions in the same manner as they did when they were a candidate.

The requirements of managing an enormous and complex national security bureaucracy reward careful deliberation and strategic consistency, while sharply punishing the kind of policy shifts that are more common on the campaign trail. Statements by the president are taken far more seriously abroad than are promises by a candidate, by both allies and adversaries alike. And while policy mistakes made before entering office can damage a candidate’s personal political prospects, a serious misstep made once in office can put the country itself at risk.

These are the realities that President-elect Donald Trump will encounter when he enters the Oval Office. In building his national security team and designing his decision-making processes, he now has the opportunity to learn from those who have come before him, avoid their mistakes, and adopt their best practices.

Before President Trump takes office, Washington will be papered over with studies such as this one advising him on how to best organize his national security team.¹ Many of these reports’ recommendations will be somewhat redundant, which may cause some to wonder whether this field needs yet more tilling. We believe it does.

The similarities in recommendations across numerous expert studies are an indication that many practices in recent decades—across administrations of both political stripes—are generally considered by experienced policymakers to be inadequate to the demands of sound national security policymaking. Some suboptimal practices have been particular to specific administrations and presidential personalities, while others have been the result of a slow accretion of behavior patterns that transfer from one administration to another, even across party lines. As a direct result, America’s national security decisions and their execution have too often underperformed.

It is rare to find unanimity among leading national security policymakers across party lines and generations, but on the subject of national security policymaking, the authors of this joint report found general consensus on two central points:

• Process matters. Effective decision-making processes can go a long way toward facilitating successful national security policies, and dysfunctional processes can be the undoing of the best-intentioned plans and objectives. The importance of process is often underestimated by national security analysts who have not previously been policymakers themselves.
• It starts at the top. Unfortunately, most U.S. presidents in the modern era have been elected into office without deeply considering what processes might ideally integrate both their own personal preferences and the inherent needs of the large, complicated national security institutions that they will soon lead. Most incoming presidents have thus been forced to learn on the job and have adjusted their administrations’ processes accordingly over time.

It would be preferable, of course, for the next president to avoid this steep learning curve. And given the wide range of national security challenges that are likely to confront President-elect Trump in the first year of his administration, the cost to the nation of a too-steep learning curve may turn out to be unacceptably high. Our aim with this report is to help the incoming administration avoid this risk.

This report seeks to help the incoming president and his transition team identify choices that are often not deliberately made but are crucial to a well-functioning interagency process that provides information and tees up decisions in ways that support the president’s management style, rather than becoming an impediment to the president. The authors engaged with a wide range of high-level practitioners to determine best practices for making national security policy—including Stephen Hadley, Susan Rice, George Shultz, Madeleine Albright, and Henry Kissinger—and quote their comments at length. In our judgment, hearing the voices of the practitioners was the most valuable way to convey their lessons.

From these wide-ranging interviews emerged a general consensus on what steps President-elect Trump—or any future president in the decades to come—should undertake to demand processes that better inform policy choices and allow his or her decisions to be carried out more faithfully. While each modern administration has had its own sets of experiences and circumstances, its own moments of dysfunction, and its own comparative advantages in its national security decision-making processes, the degree of unanimity among the former leaders across these administrations is notable indeed. More voices help convey the urgency of improvement.

And while many recommendations are indeed similar, they are not all identical; where and why they differ is itself interesting and can inform an incoming administration looking for structural or procedural ways to improve government’s performance.

At the same time, it is all too easy to stipulate how an administration should be organized without factoring in the specific politics and personalities, the long-standing feuds and friendships, the excess or dearth of talent for positions, the competing priorities, the international crises, and the other urgent problems that will not wait on an administration to take shape. What strategist Carl von Clausewitz said of warfare is also true of national security policy: “Everything is simple, but the simple is exceptionally difficult.”² Running the national security policy process is a demanding job, more often done poorly than well due to the sheer degree of difficulty.

The most important conclusion of this report came from Stephen J. Hadley, President George W. Bush’s second national security advisor, which we call Hadley’s Dictum: “Presidents get the national security process they deserve.” The president’s role is so determinative and personal influence is so all-encompassing that responsibility cannot rest anywhere else. His or her personality, personnel choices, the administrative routines they establish—whether by active management or simple default—overwhelm all other factors. The seminal advice from this report is that while there are better and worse ways to structure policymaking and execution, the optimal choices for any administration are those that work with the grain of the president’s management style. Textbook practices at stark odds with how the president is comfortable making decisions will result in circumvention of the formal decision-making process until practices are found that suit the president’s needs.

As Henry Kissinger has noted, it is theoretically better to use the Brent Scowcroft model of the National Security Council, or NSC, as honest broker, but most national security advisors and their staffs nonetheless become part of the debate. Many presidents have tended to have their immediate, in-house staff—the NSC staff—handle the most sensitive issues and negotiations because they are the most personally trusted, tend to have the greatest knowledge of and commitment to the president’s agenda, and are least likely to engage in leaks to the media and Congress hostile to the president’s leanings before a policy takes shape.

It takes an enormous amount of trust and unusually effective enforcers of the process for the interagency to work as the Dwight Eisenhower and George H.W. Bush administrations did. Nobody can do more to build such a culture of trust within an administration, or by action or inaction proliferate a culture of distrust, as can the president. Moreover, if a president cannot extricate himself from tactical decisions—either because it is the natural level of his engagement or because he prefers to reason from the tactical to the strategic—it unfortunately makes little sense to organize NSC meetings to decide strategic-level issues. The process must fit the president, not vice versa.

While researching this report, we were struck by how many crucial decisions about the process of policymaking on national security issues were the result of drift rather than deliberate decision. And as Condoleezza Rice has emphasized, opening the aperture for consideration of alternatives is a challenge once policy has been set. Presidents very often carried over practices and people from campaigns without conscious choices about whether those individuals would be suited to supporting the demands of the nation’s highest office. And yet all too often, they have found that the processes for organizing information and framing decisions that have served them so well in other circumstances, even in other executive leadership positions, are not always directly transferrable to the presidency.

It is our firm belief that this report can best contribute to the next administration by properly framing the questions that will most advantage the incoming president and his transition team as they are considering the staffing and workings of a national security team. With an in-depth understanding of the president and how he wants his administration to function, transition teams can make recommendations best suited to a well-functioning interagency that works with the president’s management style rather than struggling against it. This report concludes with a relatively long section on best practices that quotes many of the interviewees on subjects that they repeatedly raised.

We guided the process by posing questions but had no preconceived notions of what should be done differently: Our methodology has been to read widely of history and memoir and ask top policymakers to share their experiences and thinking about how to improve the processes to better serve the president, irrespective of who that is or their policies. Our analytic approach has been a devotion to economist Tyler Cowen’s first law, which stipulates that there is “something wrong with everything.”³ The authors sought to explain what is difficult or demanding in every approach recommended so that transition teams will have a catalogue of pros and cons associated with different approaches and can best optimize their practices.

In producing this report, we have been rigorously bipartisan, not only in the composition of our research team from both the Center for American Progress and the Hoover Institution but also in the people we interviewed. The research base consisted of former national security advisors, Cabinet members, deputies, and assistant secretaries and senior staffers regarded as especially attentive to structural and procedural issues. The authors have particularly sought the views of individuals who have routinely participated in Principals and Deputies Committee meetings across multiple administrations, both Republican and Democratic, at the White House and within the executive departments.

Lengthy interviews were conducted by the authors over the past year, typically in person in the interviewees’ offices. Those interviews were recorded and transcribed, and interviewees were offered opportunities to review those transcripts and clarify their intent so that we could ensure we were reflecting their views accurately. Unless otherwise noted, all the quotes cited in this report come from those interviews. A detailed list of those interviewed is included in Appendix A.

The first scheduled interview for this project was with Sandy Berger, who served as national security advisor in the second term of the Clinton administration and as the deputy national security advisor in the first. Sadly, Berger died a few days before we were to have sat down with him. He had strongly encouraged that this report be written, stressing how important good process was to good policy. He particularly liked our approach of letting people who have held these demanding jobs explain what worked and what did not in the time of their responsibility. Berger actively guided our preparations and the types of questions we planned to ask interviewees. We sincerely hope he would approve of the result, which is dedicated to him.

Presidents often underestimate how important it is to get the process of policymaking right. Other than Dwight Eisenhower, no modern president has ever entered office having run an organization as large and complex as the federal government. Indeed, many have never run anything larger than a congressional staff office. Those assuming the presidency have likely never reported to a board of directors as divisive as the U.S. Congress nor had customers as diverse, demanding, and opinionated as the American public. They have experienced the media scrutiny that comes with running for president, yet they are often surprised that it is possible for the intensity to increase even further once in office. Most presidents have never had to make life and death decisions before, nor were they responsible for precedents that would reshape international practice or materially affect the world’s largest economy. There may be no genuine preparation adequate for the job.

Still, it is a fact that most presidents have dramatically overestimated their ability to put their personal stamp on the working of the executive branch. This is nowhere more true than in national security policy. The development of many of the essential U.S. capabilities, such as major weapons systems, intelligence access, the caliber and training of career personnel, the depth of trusted foreign relationships—all have lead times far in excess of a single presidential term. Many key appointments—such as Federal Reserve chairmen, FBI directors, and senior military appointments—are staggered to provide depoliticized continuity across administrations. Congress determines spending levels and allocations and can frustrate the president with policy riders and stalled confirmations even when the president’s political party is the majority. Moreover, the military; the foreign service; and the intelligence, economic, development, and law enforcement communities each have deeply entrenched and differing cultures and incentive structures that are difficult to affect from the White House.

Yet most presidents think that because they are commander in chief, the interagency will be immediately responsive to their commands. This is only true if a president structures the process of policy formation to encourage institutional buy-in and then carefully supervises its subsequent execution—something few do. By far the more common practice is for a president to come into office without having much considered the optimal structure that balances the way he likes to receive information and make decisions with the legitimate institutional needs of the organizations he now leads. Presidents therefore fail to define the process clearly or demand that senior appointees abide by that prescribed process. As a result, they most often end up reflecting in their memoirs that they felt trapped by an ineffectual national security policy process that failed to produce a wider range of choices, was unresponsive to direction from above, and boxed them in by denying the time and range of choice they desired.

Solving this problem begins with developing a solid understanding of the functional needs of the key national security agencies, especially the U.S. Department of State and the U.S. Department of Defense, or DOD. Some of these needs are common to all such institutions. For instance, large organizations such as State and DOD work best when they receive clear strategic direction from the White House that is itself the result of an inclusive and deliberative process. At the same time, each department and agency has its own cultural idiosyncrasies. To make a musical analogy, the military is like a symphony orchestra, and at its best, the foreign service can be like a jazz quartet. They both can sound beautiful. But while one is large, highly specialized, led formally, and follows a plan, the other is smaller, more flexible in its roles, led informally if at all, and rewards improvisation. An effective policymaking process recognizes these differences to meet the institutional needs of both communities simultaneously.

But if there was only one correct answer to this conundrum, then this report would read as a checklist rather than a series of questions and findings. As former Deputy Secretary of State James Steinberg emphasized to us, “Each administration requires a process which is suited to the style of decision-making of the president.” And as Ambassador Nicholas Burns points out, “You’d have to say Nixon was successful with a system so different than Bush; it’s not just the system, it’s the fit for the president.”⁴

Trust is the fundamental metric for determining how to structure the process

Outside analysts—and those in the various departments—often critique White House micromanagement of the Cabinet. And these analysts are almost uniformly critical of instances when the national security advisor and the NSC staff have gone operational by undertaking execution of policy instead of restricting themselves to staff functions. Indeed, operationalizing policy execution in the White House has famously resulted in outcomes deeply damaging to presidents, whether by making tactical decisions on bombing targets in the Johnson administration or by running the Iran-Contra operations in the Reagan administration. As several interviewees stressed to us, the conventional wisdom that the NSC staff should not be operational really emerged from the Iran-Contra scandal experience.

And yet most modern presidents have also—sometimes, quite successfully—decided in certain instances to entrust the development and execution of their most critical and riskiest policies to their national security advisors and their immediate staff. Crises tend to be managed from the White House and have included secret overtures as well. The surge in Iraq, the Israel-Palestinian peace process, negotiations to end the Vietnam War, and the openings to China and Cuba were endeavors led by national security advisors—or by presidents directly. The extent to which a president relies on the NSC staff exclusive of departments for the development or execution of policies can be understood as the result of the president’s confidence, or lack thereof, in the workings of the rest of his administration. As former U.S. Ambassador to NATO Ivo Daalder noted to us, reflecting on the Obama administration’s reputation for centralizing national security decision-making: “The Scowcroft model requires high levels of trust, and President Obama didn’t have it.”

Personnel decisions are policy decisions

There is simply no substitute for good people. “National security is a team sport,” Stephen Hadley stressed to us. He went on to note, “When making national security appointments, the president needs to think about the process as putting together a team. It needs to be the president’s team. It ought to be people the president knows and has confidence in personally.”

That it is diabolically difficult to tell who will prove successful in high-level interagency coordination jobs additionally clouds the process of making good choices. But there are characteristics common to people who have successful track records. They are committed to strong policy processes, instead of only to particular policy outcomes. They understand what level of authority is appropriate for different decisions to be made. They listen for departments’ concerns and craft integrated approaches as policy alternatives. They frame strategic decisions for Cabinet consideration and insist that the full range of strategic options be reviewed, not just the ones perceived to be preferred by the president. They use their authority to ensure implementation of decisions once made, review effectiveness, and propose adjustments as required. If part of the president’s campaign team, they have the rare discipline to be able to shift from campaigning to governing.

There are understandable reasons why any president might select Cabinet members that he did not already have reason to trust deeply. For example, they may bring specialized expertise that is valuable for the position. They may represent politically important domestic constituencies, critical for a successful presidency. They may provide geographic, racial, age, or gender balance among advisors. They may bring a prior adversary inside the proverbial tent. But rather than being a model to emulate, Lincoln’s famous “Team of Rivals” is the exception that proves the rule: Appointing untrusted persons to the Cabinet almost always increases the risk of interagency dysfunction. At the very least, such appointments will likely result in additional supervisory responsibilities from the NSC staff. More typical is that a department led by someone untrusted by the president will find itself marginalized in the interagency progress. And as Secretary Madeleine Albright noted, “These are huge agencies, and if they are not included, they either sit on their hands or they do something that you do not want them to do.”

If a president does select Cabinet members who have difficulty working effectively together or subordinating their activity to the president’s agenda, the role of national security advisor becomes especially important. As President Reagan’s last national security advisor, Colin Powell defused friction in the wake of Iran-Contra by emphasizing that he worked for the statutory Cabinet members of the NSC, not just the president. He thus held a brief in-person meeting in his office each morning with the secretaries of state and defense to ensure common focus.

Bureaucratic rivalries are endemic to the interagency, inherent to departments with overlapping roles and responsibilities, and pre-exist any political personnel a president might appoint. This is why it is critically important that presidents select national security personnel who actively seek to minimize these rivalries, rather than those who exacerbate them. There is no substitute for candidates with deep experience in the executive branch, preferably across multiple agencies. Ideally, personnel decisions would be designed to create effective teams, both among the Cabinet and again within key sub-Cabinet subject area groupings. Very little will be accomplished if the relevant NSC senior director, assistant secretary of state, and assistant secretary of defense responsible for a specific issue area distrust each other personally and are unable to work together effectively.

Presidents can further mitigate risk of ineffectual appointments by training individual appointees and teams of interagency counterparts at the beginning of an administration—that this is almost never done is one of the most glaring oversights and easiest fixes for improving the process of national security policymaking and execution. Given the central importance of trust in interagency policymaking, lessons might be drawn from business management research that suggests that the concept of “psychological safety” is key to successful group dynamics.⁵

Most presidents underestimate the challenge of effective communication within their administration

Politicians consider themselves effective communicators, for how can one be elected to national office without communicating effectively? Yet many, even most, presidents struggle to have their priorities understood within the federal bureaucracy and get the work of departments aligned to those priorities. And most do not consider how effective a specific candidate might be at internal communications when making Cabinet selections.

In day-to-day management of national security, good communication allows policymakers to come into meetings having the same correct assumptions and to leave meetings with a common understanding of the next steps. Internal communication is driven by the model of interagency decision-making. The two foundational models of national security process are the Scowcroft and Kissinger models. Scowcroft is the exemplar of a neutral and transparent process of policy evaluation; Kissinger exemplifies an opaque and highly centralized policymaking process with the national security advisor as the president’s confidant and policy surrogate. Both models rely on principals making strategic-level decisions and departments empowered to execute policy. In practice, the Kissinger model often leeches so much information and responsibility away from departments that confused policy execution can easily be the result. It can also lead to a more activist Congress, urged on by departments and individuals uncertain about or opposed to policies they did not have a hand in crafting. The same logic applies to re-litigating policy decisions in the media: Incentives increase for working outside the system when the system is not perceived to incorporate departments’ concerns.

Historian Philip Zelikow, who participated in the 9/11 Commission, has previously emphasized that, “All successful presidents have a shakedown period in which they figure out how to compose and lead a team that is good at governance.”⁶

Secretary George Shultz pointed out to us that lengthy confirmation processes also push decisions to the White House early in the course of an administration, which can become a set pattern if not consciously pushed back on as departments are staffed. In recent years, it has become more common for tactical-level decisions to be directed by the president’s staff in the White House. This may partly be driven by the revolution of instantaneous communications that make confidentiality more difficult and immediate reaction seem urgent, as many recent senior policymakers explained to us.

It is also partly driven by the burgeoning size of the NSC staff, particularly after incorporating the enlarged Homeland Security Council. At its high point, the total number of staff members was reported to be more than 400 people, vastly larger than it was a generation ago.⁷ The number of people currently reporting to the homeland security and counterterrorism advisor alone is roughly the same as the size of entire NSC staff that worked for Zbigniew Brzezinski or Brent Scowcroft only one generation ago.⁸ Talented people armed with the president’s mandate are seldom passive in the conduct of their work, so they find problems to solve—even if it is not their proper role to do so. There is widespread consensus that this number got too high, and most recently, under National Security Advisor Susan Rice, the staff has undergone a deliberative process of reductions, coming down 13 percent to date.

Former Under Secretary of Defense for Policy Michèle Flournoy stresses that the White House needs to clearly communicate to the departments what decisions the president wants to make personally; which types of interagency decisions should be delegated to principals, deputies or below; and which types of decisions should be left to the individual agencies. The most recent two administrations have tended to raise a large proportion of policy decisions to higher levels than was done previously, with deputies very often a consultative rather than a decision-making committee. President George W. Bush allowed interagency processes in his first term that precluded most decisions below the principals.⁹ President Obama is reported to hold Cabinet meetings to solicit advice and then make many key national security decisions alongside a small group of longtime trusted aides.¹⁰

The higher the level of decision-making, the more rigorous a White House must be to ensure effective communication within the government. Otherwise, departments will continue to contest policies that have been decided and hesitate to carry them out speedily in case the information they are operating on proves inaccurate. While it ought to be the case that decisions at high levels allow for harmonization of individual policies across broad spans of decision-makers’ authority, in practice, it is more often the cause of uncoordinated execution as information is passed through different interagency funnels to the person responsible for carrying out specific mandates. In the first term of the George W. Bush administration, ineffective communication resulted in the president’s decisions not being carried out because Cabinet members did not mobilize their departments to execute policies.¹¹ In the Obama administration, ineffective communication has led to r