Process Makes Perfect

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 resentments between the White House and multiple “operational” departments and agencies.¹²

Presidents must own the process

Richard Clarke, who served on the NSC staffs of Presidents George H.W. Bush, Bill Clinton, and George W. Bush, describes the Brent Scowcroft NSC as ruthless in policing the interagency. Deputy National Security Advisor Robert Gates would have NSC staff draft brief memoranda outlining options and identifying departmental positions on the issue for decision. In the Deputies Committee, Gates would ask participants to affirm that the strategic options outlined in the discussion papers accurately reflected their departments’ positions. If the papers did not, either the NSC staffers who drafted them or the department staffers who provided the input would be managed appropriately. Meetings were designed to force decisions at the lowest possible level rather than routinely push decisions to the highest level. Scowcroft himself described the process as “if they were at loggerheads, Gates simply had to say, ‘Well, let me talk to the president and see how things work out.’ You only had to do that once for people to understand how things ran.”¹³ Because Scowcroft was empowered by President Bush to run the process this way, no one doubted the NSC was the guardian of an efficient process that produced a maximum of agency participation in and clarity about decisions.

As Daalder emphasized to us, “One of the reasons Gates and Scowcroft are effective is because the whole system was effective.” Owning the process also means clearly defining for all senior staff at the beginning of the administration the acceptable and unacceptable behaviors of NSC principals and the NSC staff. Presidents typically use their first issued directive to outline the interagency process but do not enforce its practice. Former NSC Senior Director for Defense Frank Miller attributed much of the difficulty in the first George W. Bush administration to Cabinet secretaries not considering themselves subservient to the president.

The president owning the process is also important for managing the external relationships essential to national security policy. Hadley illustrates the difference by describing to the authors President Bush’s consultations with Congress about the 2003 invasion of Iraq and the 2006 decision to change strategy and surge forces in Iraq. In the first instance, President Bush left description of his strategy to national security advisor Condoleezza Rice, which Congress took as distancing himself from the approach. In the second instance, the national security advisor was present but not a participant in the consultation and the president himself explained his strategy and the force requirements, which Hadley felt was essential in garnering congressional support.

Because there are a range of successful models for organizing national security policymaking and execution, and the choices about organization and personnel appointments need to be specific to the president in order to work effectively, the authors pose below a series of questions for consideration. The process of deliberating on the questions will assist the new president and his transition team in making appointments and establishing processes that work effectively for him.

What are the circumstances that give the president a high level of comfort in evaluating options and making decisions?

Some presidents prefer the lonely weight of solitary decisions; others are almost prime ministerial in their desire for consensus. Some prefer to talk through evaluations, while others absorb information and argument better by reading. Some prefer formal and inclusive meetings that generate multiple ideas coupled with smaller, informal groups to weigh options and make decisions. Some prefer to deliberate in private and communicate decisions after the fact. Structuring a process that makes it harder for the president to think his way through to weighty decisions serves him poorly. Matching the process to the president’s actual proclivities may be the most important national security decision of the administration.

What role does the president want the Cabinet to have in strategic decisions: consultative, deliberative, or decision-making?

This is a decision only the president can make, but it is important for the transition team or the national security advisor to have a conversation that allows the president to honestly and confidentially assign roles to the Cabinet that will facilitate rather than aggravate his decision-making. President George W. Bush seems to have envisioned himself in the Reagan mode of encouraging active dissent among the Cabinet, yet he was most comfortable making decisions when the Cabinet had reached consensus ahead of time.

How does the president manage conflict among senior staff?

Some presidents prefer outright disagreement among their Cabinet. Even when that approach suits the president, it is often injurious to smooth functioning of national security policy because of resentment among the Cabinet. Elliott Abrams, who held foreign policy positions under President Reagan and President George W. Bush, told us of Reagan holding repeated Cabinet meetings on whether to intervene in Panama, not because he was debating the policy merits but because he was managing disagreement between Vice President George H.W. Bush and Secretary of State Shultz. Does the president see value in doing this work himself or prefer to have the chief of staff or national security advisor smooth feathers and ensure teamwork among the Cabinet? How the president manages disagreement that will inevitably occur among his Cabinet will be a main driver for sustaining trust in the interagency.

How much confidence does the president have in the judgment and managerial ability of the national security Cabinet members?

This will affect both the autonomy national security Cabinet members are given and the role they have in deciding national security issues. It will also affect the interagency process, since, as former Homeland Security Advisor Frances Townsend emphasized to us, preparation by principals for meetings is highly determinant of outcome productivity. For the Scowcroft model to work, it requires high-order performance not only by the NSC staff but also by interagency principals and deputies. A high degree of confidence in management by others will allow downsizing of the NSC staff and the holding of fewer and more strategic meetings among deputies and principals.

How will the president enforce decisions in the interagency?

President Eisenhower used statement of conclusion documents as enforcement: If any Cabinet secretary disagreed with the NSC statement, the Cabinet would have to reconvene. President George W. Bush seemed to tolerate high levels of noncompliance by the secretaries of both state and defense in his first term, as well as to utilize stoplight charts for grading policy execution in his second term.¹⁴ Most other presidents fall somewhere in between these two approaches. Ensuring that departments accept a common understanding of decisions and carry out their respective responsibilities in policy execution is essential; determining feedback mechanisms and empowering enforcers are important actions for a president to establish early.

How sensitive is the president to low-level bungling by departments?

If very, then either senior appointees will need to have very close knowledge of the president’s objectives or the NSC will need to have line authority to execute policies or wide supervision over departments’ implementation. If the president has a higher tolerance for mistakes that are attributable to departments’ work, then the NSC can better serve the president by ensuring interagency debate on strategic issues.

How will the White House protect the president’s need for confidentiality while deliberating?

Loyalty of NSC staff is essential to having the president’s trust; this argues for a senior staff known to the president. Kissinger said that sensitive tasks migrated to him because the president trusted his confidentiality. From his vantage point close to President Obama throughout the past eight years, Ben Rhodes described to us instances in which the NSC staff identified issues as presidential opportunities and then created processes around them. Work will migrate to competent people trusted by the president no matter their official jobs and responsibilities; the best practice would be to place those trusted, competent people in the relevant statutory positions to begin with.

In answering the preceding questions, transition teams can determine what types of organization and management of national security best suit the president’s own patterns of behavior, while still acknowledging the legitimate institutional needs of the bureaucracies that must execute the president’s decisions. The answers will help the White House establish the boundaries for the role of the National Security Council staff and its organization of interagency practices.

Over the course of our interviews, we perceived a wide consensus on best practices across presidential administrations for producing sound policies and effective implementation that are offered below for consideration. To emphasize that the views presented are not the authors’ alone, we have included longer quotes on each subject from many whom we interviewed.

Set the rules of the road

Within a month of the election, the president-elect should meet with the person selected to be the next national security advisor to examine alternative models for national security decision-making, review best practices from previous administrations, and establish guidelines for the incoming administration. As soon as practicable during the transition, the president-elect and the next national security advisor should meet with those expected to be nominated as members of the Principals Committee and Deputies Committee to discuss those guidelines and establish clear expectations on processes. Soon after the inauguration, the new president should issue written guidance on process to all relevant departments.

• “The president needs to take responsibility for the NSC system and see it as a team. Presidents get the national security process they deserve. Presidents can either pick the one they want, or if they ignore it, they are going to get a dysfunctional system, which they deserve because they have not taken ownership of the system to fix it. Part of the ownership involves presidents putting themselves in the centers of their systems.” – Stephen Hadley
• “What decisions does the president want to make? What decisions is he or she okay with his secretary of defense or secretary of state making? And where are their shared authorities? On what do both Defense and State need to agree? If you have some of those guidelines upfront, you’ll be able to answer the questions on maybe 80 percent of the issues. On 20 percent, people will still want to argue and contest, but by then you’ll have drastically reduced the level of confusion.” – Michèle Flournoy
• “Early on, the president and the national security advisor have to be clear on how the process is going to work. The national security advisor has to be empowered to call fouls and make sure that there are consequences for lack of coordination and lack of good faith. The national security advisor has to be invested in the whole process and has to make sure that the Cabinet members of the NSC are committed as well. That’s a ‘come to Jesus’ moment in the start of an administration. This is how we’re going to operate; these are the rules. We can always deviate, but we have to be clear when we’re deviating and why.” – Juan Zarate
• “From the very beginning, even before inauguration, the president has to have people looking at the way government is organized to execute policies, and not just at the policies themselves. … Unless the president is personally convinced that the organization of the national security policymaking system is a major factor affecting the odds of success and is personally interested, there is no other power able to substitute for that.” – Leon Fuerth
• “There have to be Cabinet loyalties at the highest level that then have to get transmitted down, to say we won’t be fighting each other. There needs to be an understanding that if you break the rules, you’re going to get fired. This establishes accountability. Nobody ever got fired from George W. Bush’s administration.” – Frank Miller
• “The president has to get principals together at the outset and say: ‘This is the way it’s going to be. You all have a right as Cabinet members to come directly to me, but if you do that all the time, the system’s going to break down. If you think your views are being suppressed and your department is being overrun, or you think that the NSC staff is doing your job, come tell me. I trust the national security advisor. And when he or she says that the president has said x or y, she’s speaking for me and not making it up. If you want to test that and call me, you can do that once.’” – Richard Clarke

Define the role of the national security advisor and the NSC staff

The next president and national security advisor should clearly define and communicate their vision for the preferred role of the advisor and his or her staff. It is important that Cabinet members, and by extension the thousands of people in their departments, have confidence that the national security advisor and staff will be honest brokers when differences emerge.

• “The key tension is that the national security advisor has to be on the one hand the honest broker and make sure that the options are clean, and on the other hand, the national security advisor has to be an advocate for policies he or she thinks are the right ones. The answer to the tension is to be transparent.” – Anthony Lake
• “When Colin Powell became the national security advisor, he came over to my office in the State Department. He said, ‘George, I just want to let you know that I’m a member of your staff. The National Security Council consists of the president, the vice president, the secretary of state and the secretary of defense. That council has a staff, and I’m the chief of it. Obviously, the president is my most important client, but all four of you are also my clients.’” – George Shultz
• “The national security advisor should be an honest broker and a systems administrator, someone who runs a process that ensures that the best and most insightful, helpful advice and options get to the president to enable the president to make better decisions. And that should importantly include the fair representation of dissent. In addition, obviously, the national security advisor will have his or her own views and will be able to offer those as personal staff to the president. But I think their primary role is to structure the process such that the president has the benefit of a full range of views before he or she is making a decision.” – Michèle Flournoy
• “The president needs to decide how policy is going to be made. There are two models. First is the traditional domestic policy model, whereby the president relies heavily on the White House staff to come up with a policy, then Cabinet secretaries are brought in to implement the policy. The alternative model is what I think is the right model, whereby the Cabinet secretaries are part of the policy development process. The president has to decide which model he or she is going to use.” – Stephen Hadley
• “The Cabinet has to believe that even if the national security advisor is an advocate for a particular position, that either they have access to the president or the president hears their views in an honest fashion. Even as an advocate, there has to be an aspect of that, that accords with the notion of being an honest broker.” – Rand Beers

Detail presidential access

Early on, the president should establish a regular battle rhythm of communication with the national security advisor. By modestly extending the time accorded to the in-person President’s Daily Brief on intelligence, for instance, the president will allow for a review of required national security decisions and upcoming Principals Committee and Deputies Committee meetings. Hadley described to us how he would come to see President Bush every day to let him know what was going on at the Principals Committee, for instance. If the next president decides, as some previous presidents have, to take the President’s Daily Brief in written form, then there should be a separate daily discussion scheduled with the national security advisor to meet this objective.

• “It would be very good if the president has a national security briefing every day. I think the intelligence briefing can be optional. We didn’t have a separate intelligence briefing for Reagan. We gave him the book every day, and he would read the articles that we highlighted. We did not have the staff prep for the book. President George H.W. Bush and George W. Bush obviously cared more about intelligence, the father having been the director of the CIA. Now, it’s become an expectation of the intelligence community to have a half hour with the president every day. But frankly, the intelligence briefing is optional.” – John Negroponte
• “One of the institutionalized meetings that can occur is the CIA briefer presenting the Presidential Daily Brief to the president, with the national security advisor in the room. … The preferred outcome is that the president meets with the national security advisor at least once a day, however you want to structure it. It actually makes sense to do it in conjunction with the PDB. This is the simplest way to do it.” – Rand Beers
• “In this administration, the President’s Daily Brief consists of a briefing from the intelligence community for 5 to 10 minutes and then a 20-minute discussion with the president and his top advisors. It’s a unique opportunity to flag critical issues for the president, to get his initial feedback and guidance and to prioritize.” – Tony Blinken
• “I saw it personally in the Bush administration, where the intelligence briefing turned into a policy discussion. The president would actually make decisions. Condi would say, ‘Mr. President, I think we really need to have a principals meeting on that,’ and Bush would say, ‘No, I just decided.’ That’s not the way to make decisions. You need the views of all the agencies. In the Obama administration, there is a daily meeting that has a short intelligence component but a large policy component, where a lot of discussion happens and a lot of the president’s views are formed. [Former national security advisor] Tom Donilon was very cautious on this, not to bring issues to the president until he was ready or he had all the inputs from all of the agencies. Tom was trying to protect the formal process from an invasion by the president too early.” – Michael Morell

Another question that should be defined upfront in the next administration is the degree of access that key Cabinet members will have to the president directly, outside of NSC meetings or staged public events. There are clear advantages to establishing routine patterns of meetings between the president and the secretaries of state and defense.

• “The NSC advisor sees the president every day, as part of the briefing. So he has much greater proximity. And there’s a temptation to overuse that privilege. So President Reagan suggested, ‘Why don’t you and I have two private meetings a week?’ That helped, but after a while, I told the president we should include the NSC advisor. He could just be an observer, otherwise I had to go down after the meeting and brief him. It was an insight into this very factor. The two of us would sit there and I would present my agenda of strategic issues, problems, and opportunities. These meetings also helped because then when I spoke, everybody knew I was on the same terms as the president. I was speaking for the president.” – George Shultz
• “I think that relationship and proximity of the national security advisor to the president can begin to cause certain animosity between the national security advisor and secretary of state. … What needs to happen more is to have the national security advisor and the secretary trust each other enough that if the secretary wants to be in the room alone with the president, that should be allowed. … There needs to be a process where you have the privilege of writing something to the president that does not go through the bureaucratic system. There has to be some way that a Cabinet secretary has access to the president.” – Madeleine Albright
• “The president’s regular meetings with Secretary Kerry and Secretary Carter are important because this is their chance for them to come back to an issue. It’s important process-wise to have regular engagements scheduled where non-White House principals can go directly to the president. They might have a different point of view. Usually Susan Rice and the deputy national security advisor are there during these meetings, but I’m not in these meetings. The utility of them being there during these meetings is that the president is usually not going to follow up in a way that the national security advisor is. My own view is that this should be the norm, but there should be an opportunity for them to have a one-on-one as well.” – Ben Rhodes
• “Presidents need to deal with their Cabinet secretaries directly. It was fine with me when the president wanted to talk to his Cabinet secretaries one-on-one. As national security advisor, I wanted to encourage good relationships between the president and his Cabinet secretaries.” – Stephen Hadley
• “A successful ‘system as a whole’ requires attention to president-Cabinet secretary relations. Otherwise, departments just become less useful appendages. Or even sources of trouble. Effective involvement of Cabinet departments is also key to good implementation.” – Robert Zoellick
• “I believe that they should have a regular meeting, but the national security advisor should be present. I think that if you don’t have enough confidence and trust within people at that level, that the secretary of state feels that on a regular basis she needs to talk to the president without the national security advisor there, then you have a problem. Then that’s not fixable. There may be times, as a matter of prerogative, where there’s an issue that they feel they just need to talk to the president alone about, and the national security advisor doesn’t need to be there. But don’t make it a pattern that Cabinet members meet with the president without the national security advisor present, or the vice president if the president wants them there.” – James Steinberg

Consider politics and communications

Another one of the early rules of the road should define the mechanisms by which any domestic political concerns should or should not be incorporated in the national security policymaking process. There is clear consensus that political advice should come to the president outside of the formal national security policymaking process.

• “It’s important that the NSC process helps the president to make informed strategic choices and that his political advisors provide the domestic context. That should never be part of the analysis done by the NSC process. This has been violated from time to time, with bad results. The national security advisor should help the president in as nonpolitical of a way as possible to sell the president’s policies on the Hill. Of course, this has a political dimension to it, but if the national security advisor is perceived to be highly political then, ironically, the national security advisor is less useful to the president to make a political point.” –Anthony Lake
• “I would be careful with too many political and communications folks participating in the front end. They may need to be in to understand what is happening, but I worry about the formulation of policy driven by political instincts or factors on the front end versus allowing the president to weigh such factors once the national security community has laid out its best set of options and strategies. … If you know and think that the people at the table are the political guardians and that they are there to watch the parameters of the debates, then it begins to shape the way options are thought about and talked about.” – Juan Zarate
• “My standing instructions from Vice President Gore from the very beginning were ‘tell me what you think is best for the country and leave politics to me.’ That needs to be a president’s instruction to his national security people. … Whenever a discussion at the Principals or Deputies began to move in the direction of what appeared to be political concern, either Tony [Lake] or Sandy [Berger] would stop it. They would intervene: ‘We are not going to go there. This is not what we are here to discuss.’ The integration of political concerns and reasoning about the best national policy in a given situation—those things were essentially left to the president. The purpose of the system was to make sure that the president received what was needed to make that integration successfully.” – Leon Fuerth
• “There needs to be an understanding between the chief of staff and the national security advisor about what the swimming lanes are. That is difficult and different in every administration. I think the best administrations are those where the national security advisor is the national security guy and goes directly to the president. The chief of staff can sit in on those meetings, but it’s not the chief of staff’s job to be the second national security advisor.” – Richard Clarke

At the same time, many whom we interviewed stressed that the communications strategies that drive global perceptions should be incorporated into national security policymaking. The distinction between incorporating a communications strategy and incorporating domestic political concerns is a narrow one, however. A clearly communicated policy from the outset of an administration on this issue will help avoid the risk of misperceptions down the road within the agencies about perceived politicization of national security policymaking.

• “The important point is to incorporate messaging into the policymaking process. The traditional practice is that you work up policy and after that’s done, you then figure out a messaging plan for that policy. But what has worked well for us is when messaging considerations are incorporated into the development of policy so that we are anticipating important questions. How is this step going to be received, amplified, or misrepresented by different audiences? Does it contradict other things that we are doing? How does it fit into our broader foreign policy? The more you think through how this action is going to be perceived, how we are connecting it to a narrative, the more coherent that policy is going to appear to the wider world. … Sometimes, people view the focus of messaging as a domestic-political interest, but I actually see it more as your ability to provide a framework for the entire world, not just here at home.” – Ben Rhodes
• “It is essential that the chief of staff be a part of the DCs [Deputies Committees] and PCs [Principals Committees]. You are not there to develop an unviable policy recommendation that will not be supported by anybody. The chief of staff represents the president and brings in all those different perspectives. So I have no problem and I am affirmatively in favor of that.” – James Steinberg

Effectively manage the interagency process

There is consensus among those interviewed that the next president should maintain the fundamental decision-making structure that has served the office well since the end of the Reagan administration: NSC meetings, Principals Committee meetings, and Deputies Committee meetings, all informed by thorough analysis and structuring by subject-specific interagency working groups, known by different names and initials in each administration but most recently as interagency policy committees, or IPCs.

• “The process that works the best is a process that really step-by-step tees things up for the president. It wasn’t used as effectively as it should have been in the Obama administration. When we, the deputies, got something that had been through an IPC, we had a more effective deputies meeting. And when we just had a deputies meeting that had not been teed up by an IPC, we didn’t have as effective a deputies’ meeting. It was very important to have a meeting at the senior director level and assistant secretary level. These are the people that have a better substantive grasp on the issues than anybody else.” – Michael Morell
• “You can’t forget the value of the NSC system, just because in the moment it does not work at all. That coordinating role, synthesizing role, the role of making sure that the president’s decisions are read and implemented.” – Madeleine Albright
• “The structure of a policy coordination committee, Deputies Committee, and Principals Committee, and of study and decision memos, has largely been in place since 1989. It was a great compliment when the Clinton team kept it on in 1993 and a great compliment when Bush 43 and Obama kept it on. The system as it is now structured has proven the test of time of very different presidents who have had very different approaches to national security policy formulation and execution. Be careful of tinkering with the system that has served a bipartisan set of national security decision-makers and presidents very effectively.” – Robert Kimmitt
• “The nature of the foreign policy and national security issues that we face today is that they are far more interagency and multidisciplinary. … There are far more agencies at the table. … We have had a robust NSC role, but it has been one out of a necessity of coordination because there have been more agencies operating in the national security space given the nature of 21st century threats. … The NSC is the only coordinating body that can do that.” – Ben Rhodes

Not all interagency meetings are for the same purpose. It is important from the outset to distinguish between the three types of interagency meetings at every level: policymaking; implementation oversight; and crisis management.

Policymaking

There is widespread agreement among our interviewees that formal meetings to establish national security policy at the principals’ or deputies’ level should be well-organized in advance, limited in instance and duration, and decisional in outcome. They should choose between well-defined strategic options and not be used for general information exchange or unstructured discussion. Formal meetings should be held to decide on the preferred national strategy for the issue at hand, which necessarily includes interrelated but distinct decisions about ends, ways, and means, while providing guidance on acceptable risk thresholds and defining appropriate interim goals and objectives.

• “For something to come to principals, in almost every instance, it needs to be for the purpose of decision. We can’t be staring at problems and chewing on problems. Once in a rare while there may be a specific reason to do that—but there needs to be a very clear cut reason. … A well-prepared principals meeting is usually the product of work that’s been done well at those levels.” – Susan Rice
• “I think the deputies are there to make sure that good staff work has been done, to provide real guidance to the work, and ensure that the table has been set properly for the principals and the president. They are kind of the gatekeepers to help tee up thoughtful decision-making. First to make sure that options haven’t been unduly constrained or that there isn’t a dimension that has been left out. To make sure that things have been well-structured and the right questions have been asked. It is surprising, but sometimes what seemed to be important and obvious questions on the deputies’ level never even got asked at the working level.” – James Steinberg
• “You want options worked well through the IPCs before it gets up to the deputies so that the deputies are not restarting. Too many DCs started a discussion of an issue from scratch, with some deputies inserting new ideas for the first time, rather than building on work that had already been done. Deputies Committee meetings became graduate seminar discussions rather than focused consideration of options that had previously been analyzed at lower levels. You have to solve that problem. There have to be better papers with clear issues teed up for decision, and more preliminary work.” – John Bellinger
• “If there is a lesson learned, it is to start from an understanding of what our objectives are so that we’re making decisions about what we want to accomplish as opposed to simply making resource decisions. In retrospect, the Afghanistan decision process in the fall of the first year was not ideal. Because the resource question surfaced before anything else. The resource request drove the policy process more than the policy objective. It was means driving ends.” – Ben Rhodes
• “Back when I started in government, there used to be a requirement for multiple policy options. That’s pretty much gone away. The conventional wisdom is when there are options, people take positions, and that leads to controversy and that leads to public discussion about fighting within an administration or an opportunity for the press to be more involved. The other thing is, people don’t like to be on the losing side of an option discussion. So there is a struggle to find a consensus option, even recognizing that it may be lowest common denominator. I do not think it is an improvement. I think that government is better served by a discussion of options at each level depending upon an issue.” – Rand Beers

One of the common themes that emerged from our interviews was that to best ensure that options are fully explored and well-defined, deputies should insist on far more interagency work being done at the assistant secretary level before discussions take place at higher levels.

• “Everything needs to be pushed down. … Delegating does not mean fire and forget. Delegating means asking people to take the lead, go and convene the appropriate meeting, get a consensus or a split decision, and then come back to me before action is taken. Then I can make a decision and take this up to the president. You get confidence in people that they understand the president’s objectives so they are not going to come to you with weak options. They need to run a fair, open, and inclusive process so that I’m not going to start getting phone calls complaining about how NSC people are running the meetings.” – Stephen Hadley
• “We had a hybrid process on ISIL that was an empowered IPC but still on a lower level than the deputies’ level. It came about from a combination of experimentation and relationships with individuals. The day-to-day churning was done in the super-empowered IPC, which was very helpful to the whole process.” – Tony Blinken
• “Be sure you have a good threshold for what issues should reach the level of the Deputies Committee. When I was the deputy secretary of state, I frequently ran into the situation when they were complaining at the White House that I wasn’t coming to the Deputies Committee meetings. I couldn’t come to the deputies meetings when there were three or four meetings a day. There has been a debasement of the currency and nomenclature. The subject matter of the Deputies Committee meetings must be sufficiently worthy of the attention of those respective deputy Cabinet secretaries. This is a truly serious problem.” –John Negroponte
• “The third layer below the principals and the deputies isn’t really empowered now to structure decisions. It’s a huge waste of people’s time, and it’s enormously frustrating. That I think is the worst of the problems. … They need to be empowered and know what their deputy is going to say. You just have to enforce that. The natural tendency, of course, is that discipline will break down.” – Richard Clarke

One of the ways that such discipline breaks down is if a perception builds throughout an administration that the policymaking process at lower levels is irrelevant. This can happen if there are major disconnects between the discussions at different levels or if there is a widespread view that NSC staff members are not acting as honest brokers between agencies. In such circumstances, departments often “lean back” and refrain from active participation, much to the dismay of the White House.

• “If there is a sense that the NSC staff will make the decision, then you wait to be told what the decision is. It has to be a more collaborative spirit of government.” – Frank Wisner
• “The State Department needs to lean in and put ideas forward. When you lean back, no one in the White House will wait around. And that can turn into a cycle: If information is not coming from the agencies to meet the demand signal, then the NSC staff will have to fill the vacuum. If the NSC staff is providing the information, then the agencies will lean back.” –Tony Blinken
• “I can recall many occasions when Secretary Baker’s activism had to prod the NSC. … Baker’s confidants often worked offline with NSC staff partners to shape policy. This was certainly true for most activities at the end of the Cold War.” – Robert Zoellick

Implementation oversight, not execution

Our interviewees had a great deal to say about the important distinction between the execution of national security policies—something that the NSC staff should largely refrain from, except in very narrow circumstances—and the structured oversight of departmental implementation—a legitimate role for the NSC staff to lead. But even that oversight function can be controversial if it extends beyond the operational level and into the tactical level, whether that be clearing diplomatic cables or approving bombing targets.

These lines are fine ones, however, and different individuals—and institutions—might prefer to draw them at different places and perhaps also for different subjects. This can easily lead to accusations of White House micromanagement, a criticism that has been lobbed at every modern NSC staff but has become especially common more recently. The solution is for the president to be clear at the outset of the administration on how he sees the proper role for the NSC staff in the oversight of the execution of presidential policies.

• “Every president says that the NSC should be strategic, not tactical and operational, and I hope every national security advisor says the same thing. It shouldn’t happen. Micromanaging leads to loss of loyalty and secrecy.” – Anthony Lake
• “The NSC staff has limited functions. Support the president in their unique role in national security and foreign policy. Help write the presidential speeches, plan trips, give talking points before meetings, or handling phone calls. Champion presidential initiatives and drive those initiatives through the interagency process. … Coordinate tasks that require multiple agencies. Oversee and not perform the execution. … The president needs to insist on the limited staff role for the NSC. If the Cabinet secretary isn’t performing, the solution is not to take the tasks away him or her and give them to the NSC—the solution is to have the courage to replace the Cabinet secretary and not substitute staff for line.” – Stephen Hadley
• “The NSC performs an indispensable coordinating function. The most important practice for the NSC is to avoid the temptation to micromanage the implementation of decisions. This is where I feel the NSC can stumble and frankly continues to stumble.” – John Negroponte
• “The policymaking process is not just making policy, it’s overseeing it. So the deputies and principals meet on the implementation of policy, and I think that’s a critical role. I don’t think you need a big staff to do that. Where I think the NSC staff is less equipped is the actual implementation and the execution itself of policy.” – Frances Townsend
• “The NSC staff should not be in the business of implementing and executing. They do have an important role in oversight. So whether it’s every six months or whatever the tempo is, checking in, seeing how the implementation is going is appropriate. … Given the complexity of the government and the tempo of events, if you try to bring everything to the center for decision you will be overwhelmed, inefficient, paralyzed, and totally reactive. … You have to figure out what you can power down. What decisions are you comfortable with being made at a lower level once there is a policy framework in place? It doesn’t mean that you don’t empower people without a check, that the ambassador has to agree or the regional assistant secretary has to agree.” – Michèle Flournoy
• “The other problem is lack of implementation oversight. When you make decisions, and then they move on expecting the government to function, which it doesn’t, unless you make it. … It’s the NSC’s job to make sure the president’s policies are being implemented. That doesn’t make the NSC operational. … The NSC’s job is also to make sure that the president’s decisions must be reflected in the budgets. When the budget requests hit OMB, if they don’t have that reflected, then OMB and NSC look at that together and say, ‘nope, that doesn’t reflect the president’s policy decision over here, we’re sending it back, saying fix it and get it back to this number.’ What almost every NSC has gotten wrong—with the exception of Sandy Berger’s and Tony Lake’s—was that they had no budget process and policy for those resources. It was ridiculous.” – Richard Clarke

Specific examples offered of NSC staff operating outside their lane include conducting routine diplomacy with foreign interlocutors, tactically directing covert or military operations, and personally communicating with operators in the field.

• “We went through a difficult period during the Iran-Contra issue where the NSC staff was actually running operations, let alone taking foreign travel and speaking to foreign dignitaries. That has, by and large, been something that people have said that shouldn’t happen. The line between running a process in Washington and becoming a diplomat for one’s country is one that has changed over the last number of years. There may be situations in which it’s very important to have the person who meets a foreign government be the person who is the direct advisor to the president. But that should be the exception, not the rule. Or you just undermine the ability of departments and agencies to do the job.” – Rand Beers
• “The NSC staff is not a place where we have a stable of negotiators with foreign governments. There is no reason why the State Department couldn’t have negotiated the normalization with Cuba. I do not agree with the idea that it can only be secret if it is done out of the NSC. I have seen that before with Kissinger. The role of the NSC is to advise the president and coordinate the bureaucracy. Taking on these operational roles causes the organization to grow. Why hire talented people across the river if you’re going to do it yourself?” – John Negroponte
• “I don’t think the NSC should do a lot of travel or meet foreigners. Sometimes, sure, you might want a junior person to travel with the secretary of state and secretary of defense. This is fine. Having a director-level person tag along on some of those trips is okay, for information sharing, information flow. But it’s not the job of a special assistant to the president to go off and negotiate with foreign leaders. There are Senate-confirmed assistant secretaries of state and Senate-confirmed ambassadors. That is their job. Don’t like what they’re doing? Fire them. But don’t do it yourself.” – Richard Clarke
• “That does not mean that on diplomatic missions that are headed by different agencies there shouldn’t be someone from the NSC staff that goes along and listens and is part of the team because decisions have to be made on the ground. But I do not think that the operational lead should be in the NSC staff, I just do not think that that is what the job is. If you have somebody in the NSC staff that you think needs to be doing that, then send them out to the agencies, as opposed to having them do it from the NSC staff. There is plenty to do in the NSC without taking on that, and there is no comparative advantage for the NSC to be doing that.” – James Steinberg
• “A trap that the NSC staff can fall into is going over the head of the departments. Whenever I caught somebody in the administration talking to a special operator in Iraq, I would kindly remind them that they shouldn’t do this. They would swear to do it never again, and a few months later they would be doing the same thing. People at the NSC need to understand the ground rules. If they are not getting information that they need, then they must tell the people they deal with every day to get them that information.” – Sandy Winnefeld
• “I think it’s appropriate to have an NSC process that establishes the policy parameters and objectives, the right and left limits, the guidance for the use of force. But for execution, we have a military chain of command that has a legal mandate, that goes from the president through the secretary of defense to the COCOMs [combatant commands], and that’s how you should be running things.” – Michèle Flournoy

For the vast majority of cross-cutting issues, the primary responsibility to lead the implementation of policy decisions across the interagency should be given to a lead agency and a specifically designated official, most often the State Department in the person of the relevant regional or functional assistant secretary. For country-specific implementation, best practice is to delegate the management of interagency execution to the chief of mission. Because other agencies have at times been reluctant to take direction from the State Department, even on matters of strict implementation rather than fundamental policymaking, a key role for the relevant NSC senior director is therefore to consistently reinforce the authority of the designated lead across the interagency.

• “The new president will have to decide if it is going to be a centralized model or a distributed model where people are going to be authorized and empowered to implement policies. Of course, they get generalized guidance from the center, but then they are delegated with the implementation and execution of that guidance. The guidance has to be simple, not this micromanaging.” – Stephen Hadley
• “I am a fan of having a lead federal agency that is equipped to choreograph the elements of response. They are being enabled by the White House with meetings when necessary to make hard decisions or to move the chess pieces on the board more effectively. Ultimately, it is that model that begins to move to allow us to deal with the multiple crises and multiple things that are across agencies but where you don’t want the White House managing day to day. … Jointly hosting meetings at the NSC, so that State might be quarterbacking along with the White House. The White House brings authority, and the agency brings the expertise and bulk of equities and resources. In that process, you empower the person who is the lead on the issue.” – Juan Zarate
• “What we have done is that we conflated the idea that there needs to be an interagency process within the idea that it has to be run by the NSC staff, and I think that is correct in policy formulation, but I do not think it is correct in policy implementation.” – James Steinberg
• “When it works best, the policymaking process comes out of the NSC, and the implementation migrates out to an agency. It is a problem when the NSC gets too entangled with implementation because it’s too time consuming and in some ways disincentives agencies from taking ownership of an issue. They think that the NSC will be looking into that.” – Ben Rhodes
• “Looking at [Richard] Holbrooke’s experience in Bosnia, everyone responded to what he said because despite his title [assistant secretary of state], people knew he had the backing of the White House. When it came to Afghanistan, people didn’t have that, so it was a very different experience.” – Frances Townsend
• “Decisions that are consistent with presidential policy and intent should be pushed down to implementers. … You can appoint a lead agency, even if it requires multiagency support. The department will then appoint a program manager. That person should have a lot of authority in making decisions, as long as they are consistent in the policy. Their decisions have to be transparent and available.” – Richard Clarke

Crisis management

It is critically important that not every national security issue be managed as if it were a time-sensitive crisis. As Rand Beers aptly put it to us, “If you want to operate at a crisis mode in the moment of crisis, everything can’t be in crisis mode.” Or as current National Security Advisor Susan Rice explained: “We have to distinguish between ‘hot crises’ and the normal policy process. When you have ‘hot crises’ you don’t necessarily have the luxury of working it through in a systematic way. Of course, you have to manage issues such that those become the exception and not the rule.”

One useful mechanism to get ahead of foreseeable crises is by asking the principals or deputies to take part in a tabletop exercise, or TTX. According to Sandy Winnefeld: “One of the most effective things we ever did was a TTX. The first TTX we ever did with the NSC staff at the deputies’ level was on what happens if Israel strikes Iran’s nuclear program. … An exercise like this makes it very clear early on to all the players what was going on. … You can’t get a deputies meeting at 1:00 a.m. and say, ‘My God, what are we going to say about this, and how are we going to move forward?’ After that, we did a number of TTXs. They were hugely valuable and necessary.”

Working groups chaired by the relevant NSC senior director routinely engage in crisis management on issues that do not have the potential to affect vital U.S. national interests. But for those with the potential to have such effects, best practice is to have the Deputies Committee play this role.

The first purpose of these meetings is to share information. The Deputies Committee should therefore err on the side of inclusivity so that every potential agency, subordinate organization, or field office can provide what it knows. The second purpose of these meetings is to make quick decisions as required, even if information is incomplete, and give clear guidance to the departments and the working-level IPC on the range of options subject to further quick, formal review. The third purpose of these meetings is to manage the public narrative about the crisis and the United States’ role in addressing it.

• “It’s inevitable that if you have a big crisis, you are going to need a coalescence of leadership to understand what’s happening and to then ensure that what needs to be executed is actually being executed. … If deputies are getting together to simply understand what is happening and to ensure that the departments and agencies are doing what everyone says that they are doing, then that’s ok.” – Juan Zarate
• “Crisis management needs to be done by the White House. Because when it is a crisis, it is the president’s reputation on the line. There is not a lot of time for messing around. Crisis defined by people getting killed or us using military force, a finite period of time involved. I think the White House has to have a trained crisis manager who is your go-to guy for crisis management. Crisis management is an expertise. To ask all of your general players to have that expertise too is not realistic.” – Richard Clarke
• “The most important thing in effectively managing a crisis is clarity about who is in charge of running this process. Because it’s very easy for Cabinet secretaries to get ahead of the president. And so it’s very important that the president and the White House make very clear the limits of the authority and how far the Cabinet member can get in front. And there is tension because it’s all unfolding so quickly that even well-intentioned people risk getting in front of the president, and you’ll have a very angry White House that’s then trying to dial them back.” – Frances Townsend
• “I think the rule should be if you can create the time and space during a crisis, you should start it at the lowest level. Because otherwise you can have an issue go straight to the principals, and now they become action officers. Now, there are certain actions that occur and you have no time, so you have to make judgment calls. For example, the president has to react in the next four hours and what is he or she going to say? Then you call the seniors.” – Michèle Flournoy 

Given the requirements for the principals and deputies to conduct formal meetings to make policy, oversee its implementation, and respond to crises, there can be understandable pressure to schedule a large number of such meetings. Nevertheless, one of the most strongly expressed opinions by virtually all our interviewees was that the number of these meetings has increased significantly during the past two administrations, which has not served either president well.

The most valuable resource for the executive branch is the finite time of the senior-most members of the administration: the president; the principals; and the deputies. Moreover, the unrelenting demands on the time of these senior officials also come from their need to personally execute and oversee key policy implementation while more generally managing their own departments and agencies, some of which exceed the size and complexity of Fortune 100 companies.

• “Deputies would be in there for meetings for hours and hours on end. It comes at a cost. If you look at departments like State, Defense, and Homeland Security, at these really huge agencies, those deputies also have an important operating function. In terms of the private sector, you can think of them as the COO. They have to make sure good order and discipline, things are moving along, and budgets are getting done. I think we have to come to grips with the burden on deputies.” – Frances Townsend
• “The White House fails to understand that people are running agencies. The White House doesn’t understand that the deputy director of the CIA has to come prepared to all these deputies meetings and also has an agency to run of thousands of people.” – Michael Morell
• “I don’t think there should be deputies meetings every day. They need to run their departments. You can’t have the disruption of going back and forth to the White House constantly.” – Juan Zarate
• “The time demands for NSC meetings were a perennial problem in the Bush administration, and I assume it may have gotten worse. The deputies and principals had departments to run, and they all have heavy travel schedules. We often had four DCs a week, and one or two PCs. The pace of four DCs a week was a killer for the deputies and resulted in participation by lower-level officials because deputies were often traveling or had other commitments. Deputies who had farther to travel to the White House had to give up huge parts of their day. They often want to participate by video, but that is not as effective.” –John Bellinger
• “I think the times when we were meeting every other day on something, those were often fairly unproductive because the deputies got into the management of the execution details or displacing their staffs in trying to develop and refine options. Usually when that happens, the deputies have migrated to doing work that is typically the responsibility of a line leader.” – Michèle Flournoy

Most of those interviewed for this report shared the view that a goal should be for formal Deputies Committee meetings to occur no more than three times per week and for formal Principals Committee meetings to take place no more than once per week. Formal National Security Council meetings should take place as needed but typically once every two or three weeks under the normal press of business. Many interviewees noted that the past two administrations both routinely exceeded this rate.

• “On any given policy formulation process, there ought not to be a lot of deputies meetings. There needs to be a deputies meeting in the beginning to provide the broader framework and a deputies meeting to review the work product that has been developed as a result of that and perhaps another one if the work is not ready to go to the principals, and that is it. On the implementation side is where most of the interminable deputies meetings happen. The reason why we ended up with these incredibly large numbers of deputies meetings that go on forever is because there has been an unwillingness to delegate lead responsibility for policy implementation. … I am a firm believer that wherever possible, you should try to get the day-to-day management out from the deputies.” – James Steinberg
• “During my 14-month tenure as deputy national security advisor, I chaired 98 deputies meetings, an average of less than two a week. Of course, there were even fewer principals meetings. Too many of these meetings drain time and concentration away from the other pressing responsibilities of Cabinet and sub-Cabinet officials and supporting staffs.” – John Negroponte
• “In a good system, 50 percent of the interagency decisions should be made at the assistant secretary level and below, 25 percent of the decisions should be made at the deputies’ level, 15 percent at the principals’ level, and only about 10 percent should go to the president for discussion and decision in front of the full NSC. … Too much centralization at the higher level causes a lot of things to fall through the cracks. This also causes people at higher levels to become tactical rather than strategic in their decision-making, and they lose the forest for the trees.” – Robert Kimmitt
• “Two deputies meetings a week and one principals meeting a week. Like court dockets … what cases and what files are going to come up. Bob Gates did a pretty good job of this when he was the deputy.” – Richard Clarke
• “The more frequent the meetings are, the less time people have to prepare for those meetings, and the participation level will drop. … I think that there is nothing wrong with scheduling a Deputies Committee meeting timeslot every day. But just one meeting. For principals, I would say, once or at most twice a week, but only if you need them.” – Rand Beers
• “It’s ridiculous to have two or three deputies meetings per day. Three deputies meetings per week, one or two principals meetings per week, and one NSC meeting every two to three weeks—that’s all you need. You should push the decision down as far as you can. There should be more IPC meetings than deputies meetings, more deputies meetings than principals meetings, and more principals meetings than NSC meetings. You may need more meetings if there is a crisis or you are working through an especially difficult policy issue, like the Iran nuclear negotiations.” – Michael Morell
• “It would be great to have no more than three DCs a week and two PCs a week.” ­– Sandy Winnefeld
• “I’ve tried to conduct Principals Committee meetings with predictable regularity. We have three slots set aside each week. Sometimes they need to change because something else has bumped it, like a meeting with the president, or because something is so time sensitive that we may to meet more than the norm. But having those predictable slots and planning around them has been a useful tool.” – Susan Rice

Best practice is for the formal schedule of anticipated Principals and Deputies Committee meetings to be drawn up and continuously updated, distributed to the departments and agencies and to extend out at least one month, preferably longer.

• “What you want to do is have a baseline calendar of strategic meetings that are focused on advancing the president’s proactive priorities or dealing with big strategic items that can be anticipated and are not going away. We didn’t have meetings about China unless we had a summit or a major bilateral meeting coming up. You want to have a regular order of business based on the national security priorities of the president, and then the add-ons are what you need to do to manage crises, and that will ebb and flow with the nature of events.” – Michèle Flournoy
• “I convene my deputies every week, and we look out at the following three to four weeks. And we look at what the Deputies Committee is working on and what we expect will need to come to principals, what issues we think we’ll need to have an NSC meeting on, and we plan that out over the course of the next several weeks.” – Susan Rice

When Principals or Deputies Committee meetings are scheduled significantly more often than best practices would allow, the predictable result is that the meetings are often not properly structured; the officials at the table are not properly prepared; and the discussion is not well-focused or in the end decisional. Moreover, contributing directly to all of the above, the rank of the participants at these meetings often degrades to those unable to speak authoritatively for their departments.

• “You don’t want to hold a deputies meeting to show that you care about something. The test of something being important should not be whether there is a deputies meeting. Sometimes people can fall into that trap.” – Ben Rhodes
• “I think there are some complaints I have heard about this administration and about the last one—though less so about earlier ones—and that is that there are too many meetings without outcomes. And too much time spent by principals and their deputies absorbed in meetings. … You would have these Deputies Committee meetings that had no point, people pontificating like C-SPAN was covering it. We don’t need to hear each other talk. These positions should be well-established before you go into the meeting.” – Richard Clarke
• “The effect of having too many Deputies Committee meetings is that the secretaries delegate downwards. If you push the level of representation down too low, is that person really empowered to speak for the whole organization? The truth of the matter is, he or she is probably not.” – John Negroponte
• “Having more meetings is a characteristic of the White House trying to micromanage foreign policy. Micromanaging to me means getting into the tactical side. A disciplined White House identifies the strategic issues that the president wants an NSC lead on, while having confidence in the personnel, processes, and practices set up elsewhere. This will probably result in fewer DCs over time. I would like to have those under secretaries and deputies back in their departments doing their jobs, working with counterparts, and have meetings at the White House be both exceptional enough and regular enough that they know they are attending a purpose-driven meeting and not just a continuation of a discussion.” – Robert Kimmitt
• “Too many DCs, despite lip service about crisp decision-making, would turn into group gropes. … Smart people would come in and appear to be focusing on the issue for the first time.” – John Bellinger
• “Too often, deputies were asked to basically do the work that I think would be more appropriately done at the interagency working group level. … You have to empower the working groups. If the deputies are pulling it up to their level, then you get what you ask for.” – Michèle Flournoy
• “The IPC needs to put together all-of-government options. There were way too many times when the policy discussion was State, Defense, and CIA all saying: ‘Here is what we can do about the problem.’ And the deputies were piecing all of this together. That’s not the way policy should be made. The options need to be strategic. Each strategic option needs to be defined in a way that will contain components from each agency, if the IPC does its job right.” – Michael Morell

The new media environment can contribute to the pressure to hold meetings more often. As Ben Rhodes described to us, “A reality that any administration is going to have to wrestle with is that the media environment is so much more aggressive. … There is no discipline that is going to change that.” This new media environment brings information to the awareness of top policymakers much faster than before, provokes domestic and international audiences in ways that can be destabilizing, and makes it much more difficult to conduct U.S. diplomacy and military operations without leaks.

• “In the last 25 years, the single biggest change is the onslaught of information, and the challenge has become how to deal with all this incoming information. The immediacy of receiving information and pressure to do something about it. You now have to know things right away and do something right away. This has had an impact on the system, and it undermines the long-term policymaking process.” – Tony Blinken
• “The media creates political imperatives that are ignored at peril to the political standing of an administration, and unfortunately, the beast requires feeding, and when you’re interested in taking a long-term view, you get blamed for having nothing. So you have to carve out time for the longer-term effort. And you have to be disciplined about that in terms of the allotment of time, even though you are probably spending more time chasing the daily news clipping.” – Rand Beers

Informal processes

Ad hoc informal communication mechanisms allow NSC staff to better manage the formal interagency meetings by clarifying department positions in advance and to ensure that they reflect the views of the principals.

• “One of the most underappreciated jobs is the chiefs of staff positions in these agencies. You just need someone who knows where their boss is but also knows how to call the right people in the building. It’s harder for us to do it from the White House without it being kind of an awkward White House versus agency dynamic. But if you can call someone and say ‘I’m beginning to get the sense that your boss might be in a different place… Can we talk about how to work through that?’ Often this conversation has to take place with someone outside the chain of the deputy assistant secretary or the assistant secretary. … Incoming people should consider how to create this. The chief of staff’s office is flexible. If the chief of staff has relationships at the White House, it’s a lot easier. It’s ideal if they have relationships with their secretary and White House and they are a trusted interlocutor in both directions.” – Ben Rhodes
• “The other piece to enable you to get a clear decision on-time and have a real debate was a function of talking to everybody in advance to the meeting who was going to sit at the table. So when they walked in, I knew pretty clearly on any given issue what the views at the table were, what I had to be sure got discussed at that meeting, and where the differences would be. I would have probably had to tease out or figure out in my own mind where I thought consensus could be drawn and where the sharp divisions were going to be that where we weren’t going to bridge. And even if you did not have the unanimous view, the purpose of these meetings is not a unanimous view, but it is really clarity. Talking to the principals and having a sense of where the principals were at before they walked into the room was enormously time intensive on the part of the convener.” – Frances Townsend

Beyond noting the utility of ad hoc communications, our interviewees also stressed the importance of developing a structure of informal meetings that operates alongside the formal national security policymaking system. Administrations have repeatedly discovered the utility of having a parallel series of informal meetings and discussions at each level of interagency decision-making, designed to identify issues, set agendas, exchange information, coordinate plans, and resolve implementation issues.

• “Developing personal relationships is very important, and it is important that informal meetings take place. Like the ABC [Albright, Berger, Cohen] lunch or during the first term, when I was up in New York as U.N. ambassador, Tony Lake had these meetings that included the chairman of the joint chiefs of staff, the CIA, State, Defense, and me. They would be up in the national security advisor’s office without benefit of the staff. … The problem then becomes how you make sure that whatever was said in the informal is somehow transmitted in some way to staff and then followed up to see whether anything happens.” – Madeleine Albright
• “When Colin Powell was made national security advisor, one of the first things he did was tell the secretary of defense and the secretary of state to meet him in his office every day at 7:00 a.m. to go over the order of the day. … These types of meetings prevented any serious problems from accumulating between the departments. It was a very clever device on Colin’s part, and it worked very well.” – John Negroponte
• “I used to have a Tuesday afternoon session where I would serve tortilla chips, cheese dip, and soft drinks and we would deal with the hard operational issues. It was just the NSC principals and my deputy—no other plus ones. These were the best conversations we had because they were very candid and civil and no one was worried about leaks.” – Stephen Hadley
• “In addition of the formal system of interagency working groups, it is significant to have around-the-margins, informal weekly meetings. Our informal weekly meetings were helpful in better informing the formal process in order to anticipate the direction of the discussion. On occasion, you can resolve issues without having to go through the deputies process.” – Tony Blinken
• “In the Clinton administration, deputies met for lunches once a week, and principals met for breakfast once a week. I met in both of those venues. They were very good for identifying over-the-horizon issues, providing you did not use them to short circuit the external interagency system. If you let that happen, you are building problems for yourself.” – Leon Fuerth
• “In my tenure, we had a weekly deputies lunch, and we’d rotate who would host it. It was the vice chairman, Policy, State, and NSC. And I called it the brush-clearing meeting. … If we can work it out, then we don’t have to take the rest of the deputies’ time on it. And then there was agenda setting. What are the deputies not talking about that we need to talk about? How do we lift our gaze beyond the day-to-day issues, and what are the longer-term issues we should be working on? It was also an opportunity for us to say, ‘Look, this is something my boss is really worried about and they want a serious discussion to take place, so let’s put it on the agenda.” – Michèle Flournoy
• “We as deputies would do a deputies breakfast once a month. People would take turns hosting them. There was no agenda. We would just talk, and they were very effective in identifying issues that needed to be addressed, and what issues weren’t on the formal agenda. Formal deputies meetings would flow out of these breakfast discussions. … Saturday morning deputies meetings are a best practice, either doing a general discussion or a step back. These are incredibly valuable. No agenda; we are just going to talk about where are we on certain issues. You can talk two to three hours, not time constrained, and you’re not constrained by an agenda. New ideas flow out of these meetings.” – Michael Morell
• “We had breakfast. Sometimes it was once a month, sometimes it was every other month. It was a nonagenda breakfast, but everybody would bring things in that they wanted to talk about, and it was really collegial and bonding. It started when Denis McDonough was the deputy national security advisor and Mike Morell and I asked him to start meeting informally. … These meetings were really effective and also gave us some time to socialize a bit.” – Sandy Winnefeld

One of the advantages of these informal settings is that senior officials can express views that might differ from the institutional positions of their departments.

• “There’s a certain tension in the processes as you move up the chain between the role of the people at the table as representing the institutional perspective of their agencies and their role as advisors to the president. … The ability to be flexible and accommodate the perspectives of others is enhanced when you don’t have to do that in front of the people that you’re nominally representing. Informal discussions also allow you to have and discard ideas, to have a kind of no-fault discussion where you can have the flexibility to try out new things without owning them. And obviously, the more you’re in a small group with a degree of trust among the individuals, the more free-thinking and the more no-fault the conversation can be. Finally, and regrettably, is the problem of public leaks. Which is just a fact of life that there’s an exponential relationship between the number of people in the room and the likelihood of the material getting out of the room. Informal small group discussions help deal with this.” – James Steinberg
• “The formal meetings, in my experience across various administrations, are less practical in flexibly addressing the toughest issues than informal meetings. The principals are there sitting in front of their folks and in front of all the other principals. How do you go back if you’re the secretary of state and say ‘I sold out the State option?’ Which is why on Bosnia, I started having informal lunches, which were not policymaking lunches but were explorative. Near the end of the process, Clinton came in and joined a lunch for a while and told us what he wanted. If he had done this at a principals meeting, he would have humiliated them in front of the others.” – Anthony Lake

While policy issues are working their way through the formal and informal processes, the national security advisor’s responsibility is to keep the president informed. In soliciting the president’s views informally, a delicate balance should be struck between ensuring that the process fully explores all relevant options and that it does not waste time producing recommendations that will be contrary to the president’s strongly held preferences.

• “If there is a significant priority, we will occasionally give him an indication where the debate is taking place in deputies and get an initial steer from him. There is some utility in that so you’re not sending people essentially down a cul-de-sac. You want to have some sense of the president’s instincts.” – Ben Rhodes
• “The national security advisor should tell the president what he or she is working on and the issues the principals are pushing, to get his initial steer. Otherwise, the bureaucracy could produce a position that the president will simply reject. So the national security advisor has to be able to say early in the process: ‘This is simply a nonstarter.’ But this should never become a way of killing new ideas or enforcing orthodoxy.” – Anthony Lake
• “I think for the Deputies Committee to work, you need some degree of clarity coming from the president—what’s the policy direction and what are the policy guardrails. This should be informing the entire process before anything gets to the deputies.” – Juan Zarate
• “A president who has a particular view on a policy issue, his staff is going to know whether they ask him or not. … If you go into the meeting and you say, ‘The president’s view is X,’ then that conversation is going to be a different conversation. You’re likely to affect the tone, tenor, and subject matter of the conversation. It was important for me to know because as the conductor of an orchestra, I was going to orchestrate a conversation. I could do that better if I knew the president’s view. It didn’t mean I was pushing people in that direction, but it allowed me to listen in a different way. So after a meeting I would talk to the president and say, ‘You know, you’re inclined in this direction, but this deputy made this point. Had you thought of that?’ It could change his view. It has a lot to do with your view of your role and your personal relationship with the president of the United States.” – Frances Townsend
• “As long as that doesn’t completely cut off debate about other issues. … It’s one thing [to know] if the president leaned one way or had a gut reaction. That’s an interesting data point, but that doesn’t mean that his initial reaction should invalidate or stop more fulsome consideration of the issues. Then you’re really biasing the process in a way that is not helpful.” – Michèle Flournoy
• “Another thing that was frustrating for the deputies was, it was absolutely clear that the guy who got all the electoral votes already knew what he wanted to do. … If the president has a strong view on something, then it’s better to start the meeting that way. You are not playing any games, and nobody is going to be shy about pushing back. … Get the boss’ view on the table and work with it. Then you are not wasting people’s time. … Just be honest with people, it’s absolutely crazy not to be.” – Michael Morell
• “It is incumbent upon the national security advisor to make sure that the president knows the particular issues and to see if there is any initial guidance. … Then, I think as that process continues, you have to keep the president informed because his or her views may be changing.” – Rand Beers

Background papers

Those interviewed stressed the importance of the quality and timeliness of the background papers that the NSC staff distributed in advance of Deputies and Principals Committee meetings. Unfortunately, many complained that these materials regularly arrived too late.

• “The most important part of the deputies meeting is the pre-reads. This gives people the chance to prepare. It’s really important for people to have a piece of paper that provides the background to what it is the NSC staff wants to accomplish in the meeting and the meeting agenda. The most frustrating thing to a deputy and an agency was they either did not get the pre-reads or it didn’t come until a couple of hours before the meeting. So it’s absolutely useless.” – Michael Morell
• “There were always complaints about the papers not being circulated in advance. The NSC tried to require circulation 72 hours in advance, which was important. For things to really work well, you need to give people advance notice and papers.” – John Bellinger
• “The number of times where people at the deputies’ level or the principals’ level were walking into the meeting reading the memo in the car, maybe, or the famous table drop where now you are supposed to respond to something you haven’t seen before the meeting … it’s just not a good way to run a process.” – Michèle Flournoy
• “To have an informed discussion, you must have the materials in time enough to be informed and thoughtful, and if the papers are out only half an hour before the meeting, then you will have people just reacting off the top of their head.” – James Steinberg
• “The less crisis involved, the more time in advance you should have the read-ahead. Your organizations should also be working together at the action-officer level so the staff already knows what’s coming. The staff should have already drafted what they are going to say to you. I want to be able to take something home before, recognizing that there are some instances where that doesn’t work. This should be the exception and not the rule.” – Sandy Winnefeld

A best practice is for background papers to be distributed to participants one week before a scheduled meeting. This allows each department the time required to circulate the papers to all relevant headquarters and field offices, to assess whether its preferences have been appropriately incorporated into the papers, to resolve any remaining differences of opinion within the department, and to draft and clear department-specific preparatory materials for the principal or deputy.

• “The existing rules of when the papers need to be sent out are violated on a daily basis, and that’s partly because meetings are event driven and it’s hard to stick to timelines. In policy-driven meetings, ideally you want to build enough time so each building can have their own internal deliberative process before the meeting. It works best when you are able to bring in all the stakeholders in the building together and discuss the questions that we need to answer and have enough time internally to try to deliver a consensus or provide options for the secretary. This process requires a week to 10 days to get done and should be convened by the deputy and under secretary.” ­– Tony Blinken
• “You need to send it out with enough time to let people read and digest it. … Where time allows, you’d like the materials a week ahead. … For example, when a given paper would come to DOD, we had a number of stakeholders that needed to take a look at it. The chairman needed time to meet with his staff. The secretary needed time to get the advice of his staff. There were policy dimensions both regional and functional. There were legal questions. There was a range of stakeholders who needed time to process the materials, provide their advice back to the principals, and for the principals to pull it together and decide on their own positions. It takes time. It doesn’t happen in a 10-minute car ride.” – Michèle Flournoy
• “The paper should have an opportunity to be vetted internally by the departments. That’s more like three or four days. … The process within the department simply takes more time because they are not as flat as the White House. And to put down ideas, to have them reviewed by all of the policy people who ought to be involved but also by the legal staff to make sure that the process meets the requirements of the law and the Constitution. And then depending upon the issue, some resource people either on the personal side or on the financial side. This means that you’re not going to get a good product if you force the preparation of the materials. Because the White House is so flat, and because paper can move quickly in the White House, there is an expectation that departments and agencies ought to be able to move as quickly as the White House. It is just not true.” – Rand Beers
• “If you have too many meetings, it’s not going to be easy to staff every one of them properly. You need a good background paper. It’s good to have a paper and an intelligence briefing to set the informational framework. … You can’t just call meetings like that, unless it’s a crisis. Give people several days, ideally a week to prepare for a meeting. If you want it bad, you’ll get it bad. To tell somebody we are going to have a meeting in two or three days puts too much pressure on the system to produce product quickly. People need three to four days’ notice.” – John Negroponte
• “Ideally, you’d want to get the papers out a week before, that’s nearly impossible sometimes. At a bare minimum, 48 to 72 hours. That’s important because you want the bureaucracies themselves to be able to ingest and respond to a paper. Forty-eight to 72 hours is a crunch, but it gives the bureaucracy and the experts within a department or agency a chance to look at the paper to opine. It’s also a chance for voices within a department that for some reason haven’t been given visibility to the issue … to weigh in. The principals are given a little bit more time to talk to their people.” – Juan Zarate
• “The preparation of incredibly busy Cabinet officials is the single greatest determinant of outcome and productivity. You have to get the papers out, and you have to get them cleared and in the hands of the principals with enough time for them to actually absorb it and talk to the staff about it. I think you want it 48 hours or more in advance because you want to give them a chance to not just read it but to talk to people on their staff.” – Frances Townsend

This background paper should be drafted by the lead senior director on the NSC staff, with appropriate guidance from deputies or principals. It should clearly define the full range of potential mutually exclusive strategic options and identify department preferences for each option.

• “I didn’t pay enough attention probably to the papers [at the beginning of the administration]. In this period, we were moving awful fast and we didn’t have the formalized studies they had in the Kissinger period. For every Principals Committee meeting, Henry had a standard format for the memo. I think I didn’t, and in retrospect, it’s one of the things the president could do to encourage more strategic thinking. For papers, my recommendation would be to identify the following: What’s the issue? What’s the context? What’s the objective or existing policy? What are my options to pursue the policy? What are the assumptions? It would be interesting for the president to describe the format of the type of paper he or she wanted.” – Stephen Hadley
• “All of these things need to start with a discussion by deputies or principals of the broad strategic framework—What are our objectives? What are our constraints? Where are the opportunities? Where are the risks?—but without a detailed, formal discussion of options. Some initial thinking that outlines the big trade-offs. What are we trying to achieve here? What are the constraints? What are the things that could interfere with it? By answering these questions, deputies and principals can provide overall guidance to the process. That then makes the work of developing formal options and evaluations much more useful because it is not blue sky, it is not a fill-in-every-box kind of exercise. At that middle level, formal papers that really look through and deal in some depth with the choices, the opportunities, the likelihood of success, the risks, and the wild cards are really quite important to create a very comprehensive view within that framework of what the choices are and what the benefits are.” – James Steinberg
• “I think the policy discussion has to be more systematic and more structured. It is key to identify the interests of the United States regarding a particular issue. Sometimes, we were spending a lot of time on issues that we don’t have a big interest in and not nearly enough time on issues we do have significant interest in. These issues should be right up front in the policy discussion. The other structured part in the policy discussion includes identifying options and the risks of the options. We are really good at saying here is why this policy option makes sense, but we fail to identify the risks. … Have we thought through the potential unintended consequences? Another important part of the structured policy discussion is to ask: ‘What if we are successful, then what?’” – Michael Morell

Participant preparation

Department representatives at formal interagency meetings must be in a position to speak authoritatively for their agency. Principals should provide guidance to their deputies before every Deputies Committee meeting. Deputies should provide departmental guidance to their assistant secretaries while the interagency working group processes are underway.

• “When you had a situation where the [CIA] director’s view was different than my own, I would talk to him about it before the deputies meeting. The last thing I wanted was to take a position and have the director go to the principals meeting and say that he has a different view than his deputy. That does not look good, and it is inefficient. When I took a position at a deputies meeting, I knew or I was confident I knew that the director was on board. This is a best practice. It is very important for deputies and heads of agencies to have good communication.” ­– Michael Morell
• “You want representatives from each department, agency, and office to be empowered. When we sat down at the Deputies Committee as a small group, [Robert] Gates knew that there was no gap between Secretary Baker and me, and I spoke for Baker.” – Robert Kimmitt
• “Some deputies interpret their jobs differently, but I always felt most empowered at the table if I could definitively say, ‘I know that this is where Secretary Gates is on this issue.’ Other deputies had their own voice and didn’t think their job was necessarily to represent their department or even their secretary.” – Michèle Flournoy

Indeed, many whom we interviewed expressed similar complaints regarding the perceived lack of internal coordination within national security departments and agencies.

• “There were times when the Principals Committee meeting was a completely different experience from the Deputies Committee meetings. … It’s a tough problem when you have a centrally controlled hierarchical process of meetings where the senior leaders who are very busy are not involved at every step of the way.” – Sandy Winnefeld
• “There has to be communication within departments as to what they are trying to achieve in these meetings. I was in meetings at times when deputies and even principals clearly had not talked internally. If the principals or deputies are not communicating internally, they don’t have the benefit of the best thinking in their departments, people who are presumably more focused on these issues than they are. … I think that becomes a major deficit and becomes a clog to the system because it then gives the entire process a feel of artificiality. … A problematic outcome is that you get a principal weighing on in the 11th hour with some contrarian point of view to that of his or her department to that point. This view could have clearly been injected early on, and it would have helped the process. But now it’s getting thrown into the 11th hour, perhaps in front of the president. There was no sense of how the policy process was working, and there was no confidence in the people that the secretary was sending to the meetings.” – Juan Zarate
• “In an ideal world, the IPC and DC would be fully empowered and speak with authority from their buildings. But the reality is that most principals have no idea what is happening in the DCs, and by definition, the issues get pushed up. … You need to encourage people coming to the IPCs and DCs to get guidance from their principals and get an understanding of what their principals can and cannot commit to. … In advance of each meeting, get the stakeholders within your department together and have a consensus-level meeting, especially at the deputy level.” – Tony Blinken
• “There were too many DC meetings where a deputy would come in and say we have not been able to talk to the secretary about this specific thing, or the secretary would come in and say this is the first that I have heard of this. … Inside every agency, it is really a fundamental thing for the under secretary or deputy secretary to clear with their principal what they are going to say at deputies. … If there is not the lash-up at every level, and particularly between deputy, under secretary, and the principal, then DCs become useless because the deputy or under secretary is not speaking for the department. A lot of stuff does not have to go to a Principals Committee meeting if you can actually get it agreed at the deputies’ level.” – John Bellinger
• “There have been times in this [Obama] administration where there has been a perception that the president was resistant to the view of the Defense Department or State Department, when in fact, the deputies had a very strong view that may have been the building view, but the principals were more skeptical and at a different place. It’s important that the principals have a good sense of where their own buildings are. They need to have staff who are keeping them connected down into their building. Because they are often out traveling and doing other things. It’s very important for them to know where the discussions have been so that they are not coming into a principals meeting hearing for the first time what their deputy’s view was.” – Ben Rhodes

Determining who attends

For the vast majority of meetings, only principals should sit at the table in Principals Committee meetings and only deputies should represent their departments in Deputies Committee meetings. Only in rare exceptions should subordinates be allowed to participate in lieu of their superior. It is best practice for the principal or deputy to be allowed a single “plus one” to sit in the back row—unless the subject at hand is particularly sensitive. These backbenchers typically led the policymaking process for their departments during the interagency policy committee process, so they are in a position to provide expertise when asked, though their primary role at the Deputies Committee or Principals Committee meetings is that of a note taker and internal back-briefer.

• “I think a best practice of the Obama administration was to invite deputies to principals meetings, and the president invited deputies to NSC meetings. Deputies became the backbenchers. In the deputies meetings, we had talked about these issues in depth. Deputies almost by definition have a much deeper understanding of an issue than the principals do. It was valuable to have deputies in the room for the principals meeting. They only talked when they were asked.” – Michael Morell
• “Plus ones can be very important, particularly if you’re dealing with a complicated issue, a really technical issue, or a dilemma where the conversation is truly going to benefit from an expert perspective. So I have no objections to plus ones.” – Michèle Flournoy

However, concerns about the perceived proliferation of backbenchers, often people unknown to those at the table, were also expressed by many of those interviewed.

• “In general, there are too many people in the room. Participants in the meeting are a lot less forthcoming when there are more people in the room, particularly on sensitive topics. There are a lot of times when there were things I did not say because there were too many people in the room. I would talk to John [Brennan] or Denis [McDonough] later, but a lot of people did not get that opportunity. I found meetings without plus ones a lot more effective than meetings with plus ones. I saw the value of having a plus one, including my own. … But I found we had a much richer discussion when there were no plus ones because of the dynamics. … People were more willing to question other institutions.” – Michael Morell
• “If you are in the Situation Room and the staffer that wrote your brief is sitting behind you, and you are arguing with people, and you think, ‘They wrote this but I disagree with it, but am I not doing my job if I do not defend it.’ So I think one cannot forget the human dynamics of a lot of this.” – Madeleine Albright

Similarly, the national security advisor and deputy national security advisor should allow themselves only one plus one rather than the especially large number of NSC and White House staffers who would otherwise seek to sit in on the meeting.

• “You don’t want to go into a meeting where you are told it is only a principal plus one and then there are 15 White House people sitting there. You feel like you’ve been ambushed.” – Frank Miller
• “What I used to worry about was all the 30-somethings in the room. You don’t know who these people are. Why are they here? And what’s their purpose? As opposed to a combatant commander or an ambassador being there?” – Sandy Winnefeld
• “I think it raised some resentment when some people were told you can’t bring a plus one because it’s supposed to be a small, closed meeting and you walk in and there are 15 people from different parts of the NSC staff.” – Michèle Flournoy

Our interviewees felt that in-person meetings are always preferable to participation by secure video conference, or VTC, and having fewer meetings overall would encourage officials to participate in person whenever possible. Those who chair interagency meetings should take care when inviting field officials to participate in Principals or Deputies Committee meetings via video conference that the multitude of additional voices does not result in a less efficient conversation.

• “In-person meetings are much better. If I were the deputy national security advisor, I would say everyone has to be here, unless you are overseas. It is very difficult to have a rich discussion over a VTC. People who are not there in person tend to be less prepared. They are totally winging it.” – Michael Morell
• “I think that having anybody on a screen stilts the conversation because nobody is quite sure who else is in the room. And in fairness it’s very hard for the person on the screen to ‘read the room.’ And so other than when it’s really required, I’m not a big fan. You might need the ambassador who’s halfway around the world for a time, but I think that there’s a point in the conversation where you no longer need those folks up on the screen. Now, when I’m sitting on public board, we go into executive sessions. We have management in the room for the substance of the conversation, and then in order to be able to have a candid conversation among the board members, we skinny down to an executive session. So now, in retrospect, I wonder if there’s not a way to have some key people up on screen to set the scene for the conversation, and then going to a smaller conversation among principals.” – Frances Townsend
• “I think there are certain times when you want a field perspective. When you’re about to take on a high-risk military operation, you actually want to get some perspective from the field. When you’re about to evacuate an embassy, you have to hear from the ambassador. … But too often it becomes a habit. … At the deputies’ level, there has to be one person who is the deputy for a department and whose job is to understand the range of views within the department and bring those views—and especially the secretary’s view—to the table.” – Michèle Flournoy

Managing meetings

The chair should ensure that meetings held at these senior levels start and end on time. While perhaps an obvious point, many interviewees described it as a consistent source of frustration in each of the past two administrations.

• “Meetings have got to run on time. If you develop a reputation for running late, the people you want—the deputies, the principals—they stop coming. Instead, they start sending people because they can’t afford to lose time.” – Frances Townsend
• “No meeting is good if it lasts more than an hour. The meeting has to be well-organized enough so you can carry it out in a one-hour period of time so you can have outcomes.” – John Negroponte
• “Meetings must be able to start on time and finish on time. These are incredible busy people. The only excuse for a deputy national security advisor or for a national security advisor not being there on time is because the president wanted to see them. This is the only possible excuse. All of these people have things to do. It’s the responsibility for the meeting chair to manage the meeting in a way so that it ends on time.” – Michael Morell
• “One of my biggest pet peeves was to not have on-time meetings. Very often, the people who end up running the NSC process are people who have never led a major organization. Therefore, they don’t all understand how to run a meeting or how to run decision processes, and they don’t have a respect for time. What they don’t realize is that the people sitting around that table have run things and they are leaders. Their time is precious. When you’re sitting there at the table and the meeting hasn’t started and it’s 15 minutes late, this isn’t just a professor being 15 minutes late. This is the clock running on senior executives who have better things to do. So the NSC needs to respect their time. There will be unavoidable instances—for example, when the president calls someone to the Oval Office and he or she can’t get out of it. But if you’re finishing your email in your office and are late to the meeting, that is completely unacceptable.” – Sandy Winnefeld

To improve time management, the portion of the Deputies or Principals Committee meeting accorded for intelligence briefings should be limited. Instead, participants should have been briefed on this material by their own agencies prior to the meeting.

• “The intelligence portion of the deputies meeting went on forever. The policy part is much harder. The intelligence portion needs to be constrained in time. This is the job of the national security advisor or the deputy national security advisor. They need to cut off the intelligence portion and get to the policy portion. In this regard, it would be helpful if the read-ahead included not only papers from State or DOD on the policy questions but also included key intelligence pieces that are related to the issues.” – Michael Morell
• “The classic rule is that the intelligence community is to provide an intelligence estimate, not a policy recommendation. That is both a very sound principle and nonsense. There is no such thing as a purely factual report. You have to be sure that the assumptions that the intelligence community is using are the same ones that you are using when making this decision.” – Frank Wisner

Consensus is a desirable outcome but not required. Too much of an emphasis on consensus risks leading to language that papers over disagreements instead of highlighting them for higher-level consideration and adjudication.

• “I think what everyone would like to have is consensus, but consensus has come out as a result of having heard everybody’s views. It shouldn’t be the case where the national security advisor knows what he/she wants and gets others to agree with his or her positions. This is not consensus; consensus comes out from having discussion. Often you can do that; if you cannot, then you do need to take it to the president, and how the president deals with this depends on the president. For example, President Clinton liked to have us argue in front of him, and he was comfortable with that.” – Madeleine Albright
• “When an issue comes up for decision, the national security advisor may have a tendency to go and try to forge a consensus among the principals, and then present the consensus to the president. The alternative to that is for the national security advisor to emphasize the preparation of clear options and choices for the president. The president needs to make clear which model he or she wants to follow. … The national security advisor’s instinct to ‘please the teacher’ is trying to forge a consensus, when consensus sometimes is not the right outcome. Presidents should err on the side of making decisions rather than pushing for consensus.” – Stephen Hadley
• “Right now, we have a system that tends toward the tyranny of consensus, which means that the objective becomes finding what everybody can agree on as opposed to presenting the president with real choices and options. … Even if you don’t agree with the dissent, the hearing of it and a fair assessment of it may actually improve the quality of your decision-making.” – Michèle Flournoy
• “Debate is necessary and healthy, but the entire process can’t be held hostage to consensus. You have to have the courage when the table doesn’t agree after a couple of meetings to go to the president and say, ‘We haven’t come to a consensus and here are all the factors. Mr. President, we need some guidance,’ rather than waiting for everyone to finally agree. The danger for waiting until everyone agrees is then you have some watered-down solution that doesn’t work. This administration has sometimes had a tendency to do that.” – Sandy Winnefeld

Summaries of conclusion

Everyone interviewed for this report stressed the importance of the official summaries of conclusion, or SOCs. As Frank Wisner told us, “You need to make sure that summaries of conclusions are done quickly and accurately. It has to be absolutely clear because with national security matters, people’s lives are at risk.” A best practice, therefore, would be that in addition to the chair reviewing the conclusions verbally at the end of the meeting, a written summary of conclusion should be distributed within 24 hours. To expedite this process, it is recommended that the relevant NSC senior director produce a draft of the summary of conclusion in advance.

• “Summaries of conclusion are essential because the only way to drive the process, especially if you want to delegate, is to have a clear record of what the understanding is and what the expectations are. So the more you want to drive the process out of perpetual meetings and really engage it, the stronger your summaries of conclusions are. They should be done right after the meeting because that is when it is fresh.” – James Steinberg
• “Summing up at the end of the meeting of what was decided is really important so that there is no miscommunication. … This is important for many reasons. Having it on paper for historical reasons is really important. The public, via the ultimate writing of history, has the right to understand how decisions are made. It is also important from an institutional perspective. I would come back from a deputies meeting and debrief my guys, but 2 out of 10 times I couldn’t because of my schedule. It was important to have a piece of paper that laid out what was decided at the meeting.” – Michael Morell
• “Summaries of conclusions are usually a work list: who got assigned what; what is the exact assignment; and when is it due. … The summaries of conclusions should be written going in. And sit down immediately after the meetings and think about how we ought to change this draft. By the time people get back to their offices they should already have the summary. … The summary of conclusions should be out in an hour. If you write it in advance, it will be. … I would love to write the summary of conclusions during the meeting. Throw it up on the board, have somebody sitting there writing it as the meeting goes on.” – Richard Clarke
• “Really, in this environment where we have meetings so frequently on the same subject, the summary of conclusions of the meeting on Monday can be OBE by the meeting on Friday. Because you had a meeting on Tuesday and Wednesday. That doesn’t make any sense.” – Rand Beers
• “The summaries of conclusions need to come overnight and no later than noon the next day. The summaries of conclusions are important because they make sure that everyone has heard the same thing from the meeting. Our action officers would comb through them to make sure that there wasn’t any misunderstanding and to also look for deadlines. It wasn’t always clear at the end of the meeting who was responsible for what and who had the lead for something.” – Sandy Winnefeld
• “Within 24 hours, the summary of conclusions should be out. Especially if you’ve got a sequence of meetings building onto each other. … Summaries of conclusion can be very important as drivers of actions if they are seen as legitimate and reflective of the reality of the meeting. Where the NSC process has hurt itself is when they become seen as interpretive documents or more expansive than what was actually reflected in the meeting. … The larger the agency, the more the summary of conclusions matters to the agency.” – Juan Zarate

Indeed, several interviewees commented on the need to ensure that these summaries sometimes needed to be corrected.

• “It’s desirable that the summary of conclusions resembles the conversation that took place at the meeting, which isn’t always the case. They must be a thorough and faithful rendition of the meeting.” – John Negroponte
• “Sometimes there was very little congruence between the paper and what actually was decided at the meeting. The summaries of conclusions need to be much more accurate and reflect what exactly was decided at the meeting. They should say, ‘Here are the policy conclusions, and here are the do-outs.’” – Michael Morell
• “It’s very important to document the outcomes of the meetings because different people hear different things. It’s human nature. What I objected to is when the NSC staff uses the SOC to push a position in a way that doesn’t reflect what happened at the meeting. Rather than play the honest broker role and say ‘We didn’t reach agreement’ or ‘Everybody agreed on this, but Defense dissented, and here is their dissent,’ it is written it as ‘everybody agreed’ or ‘some people dissented but we all agreed.’ If you’re going to give the SOC real standing and quality, you have to be the honest broker. The amount of time my staff had to spend trying to correct SOCs to reflect the actual meeting was enormous.” – Michèle Flournoy
• Getting it out fast is key, otherwise people waste time going off in the wrong direction, based on erroneous information. It also allows State and Defense to say, ‘No, that’s not the summary of the meeting, you left out key information.’ There ought to be a process whereby if the summary of conclusions is wrong, that you get out a new one relatively quickly.” – Richard Clarke
• “The summary of conclusions gets disseminated to the principals, who have an opportunity to edit. But like most consensus documents, it is entirely watered down by the time it has gone through the clearance process. Because the summary of conclusions is so mushy, there is no process learning. … All of those people who participated in the process from the beginning of the policy discussion don’t really have the benefit of learning what happened at the different levels and how that decision was made so that they might incorporate it to the next level.” – Frances Townsend

There is debate, however, on how much descriptive information should be included in the formal summaries of conclusion.

• “The more difficult question is beyond the conclusions—how much of the rationale should be included? This is more controversial. From a historian’s point of view, this is incredibly valuable to not only understand what was decided but also understand why it was decided and what was the underlying reasoning and logic. For the departments, it is also easier to make thoughtful decisions of implementation if you understand the debate and rationale. On the other hand, that is often unrealistic. There is resistance to doing that because confidentiality is extremely difficult to maintain, and in some cases, there is agreement on the course of action but not the reason for it.” – James Steinberg
• “One thing worth mentioning is the contentious, litigious environment in a post-Clinton administration environment, everyone was very conscious of what went on a piece of paper to be minimalist about it. I don’t necessarily think that was good. Everybody worried about it. You were very conscious about the Presidential Records Act because historians are going to pick this up. … I do worry that all the paper, minutes, memos, all of it, was sort of edited and sanitized and written in a way that does not lend itself to what you would think of as a best practice. Views of agencies on particular issues are not clearly articulated in detail. … But most of the time, the people attending the meeting understand what is not written on the paper, who has which view. I don’t know how you overcome that or how you convince people to go back to a more written historical record given the environment.” – Frances Townsend

Presidential decisions

When issues require an in-person NSC meeting, many who we interviewed argued that the president should consider holding back his final decision during the meeting in order to continue his own personal deliberations. Doing so also provides an opportunity to manage the narrative as the new policy is rolled out.

• “In terms of the decision-making, President Obama will listen, hear everyone out at the NSC, and then make his own decision, but not in the meeting itself. If President Obama announces a decision at the NSC, everyone goes back to their agencies and tells everyone in their agencies, and you lose control of how the decision is announced and communicated to other agencies. This is an information management issue.” – Ben Rhodes
• “Almost always, the principals get a feel for the president’s view, but he rarely will actually make the decision while sitting at the table. And that’s because when he walks out of the room, he’s then talking to his staff. He’ll ask his staff their views, and then ultimately he’ll decide. And I think it’s important to leave the president the freedom to do that, to change his or her mind.” – Frances Townsend
• “President Clinton never made a decision in front of us, and that is an important part. … I think that it is a positive, it allows the president to have heard all of this and then go make up his or her mind. Maybe talking to the national security advisor again or the vice president.” – Madeleine Albright
• “At some point, the president has to go off and decide. It is not necessarily going to be at the table in the presence of the participants, but the principals need to feel legitimately that major decisions where they had concerns, and in which they will have a role in executing, are ones in which they have had input and that their input has been seriously considered. It may be that things don’t go as they might like, but the minimum requirement is that they will have been heard. … They have to feel that the discussion in which they participated as principals and deputies is truly in play when the president makes a decision. The president can draw others into it if desired—but the principals must have the feeling that their views are in the president’s mind when decisions are taken.” – Leon Fuerth

In order for the NSC staff to play the optimal role described above, the national security advisor should structure time for internal discussions with the NSC staff to encourage information sharing across directorates and drive strategic planning.

• “I would have Saturday meetings with our strategic planning cell, and we would talk about broader strategy issues and have more ‘down the road’ looking meetings.” – Stephen Hadley
• “You have to schedule and carve out dedicated times for internal meetings. We would have Saturday sessions. What Steve Hadley would do would be to call a deputy or a senior director, and he would have an issue he wanted to talk about broadly. It would be much more informal in his office. There was more time simply to just talk. He would sometimes set the agenda, and sometimes we would. The desired outcome would at times be to decide a policy process or to explore a new strategy or approach.” – Juan Zarate
• “The important thing for the NSC staff as a whole, on the senior level, is to meet together a lot. The information sharing is extremely important. Issues cut across directorates. My view is to have a meeting at least every other day with the full senior NSC staff to see where we are, what is on the plate, what we are working on, what has been going on in the meetings and what the president is focused on. The second thing is your door has to be open to the senior staff, and you just have to see them whenever they need to be seen. Especially the deputy has to be there constantly and be seen as pushing the day to day and making the team function.” – James Steinberg
• “The national security advisor has to manage and lead the staff. I don’t think most national security advisors think about that at all, unless they have run large organizations, like Scowcroft and Gates. People who have run large organizations think about that, people who haven’t don’t think about that. Sometimes, you would have guys at the head of the table who haven’t run anything. Maybe staff director of a congressional committee. … I don’t see a lot of management of the NSC staff. I see senior directors and directors who don’t know what’s going on and who can’t get their view up the chain of command. … There is a daily meeting chaired by the number three, not the NSA or DNSA. It makes absolutely no sense.” – Michael Morell

NSC staff size

Those interviewed for this report were nearly unanimous in their view that the size of the NSC staff should be reduced. Many stressed, however, that this does not mean that the NSC staff should be expected to shrink back to the modest number of people who worked directly for Henry Kissinger. As Tony Lake told us, “As the world has become extraordinarily more complex, the demands on the NSC system have grown more. Every national security advisor has complained that the next national security advisor has increased the NSC staff. But has it grown because it needed to grow?” But at the same time, there is general concern about the potential downsides of the growth seen over the past two administrations and recognition that the more recent staff reductions taken late in the Obama administration should continue under the next president. Given the strong views on this subject, we have decided to quote from our interviews more extensively.¹⁵

• “The NSC staff had grown more than I thought was appropriate, and that’s why I conducted a substantial review process. As a result, we’ve already downsized over the past year about 10 percent of our personnel. We are continuing to do that through managed attrition and looking very carefully at where we might need more people and where we can do with fewer and reallocating and downsizing in the process.” – Susan Rice
• “What’s happened now is totally out of bounds. The current NSC is too big and will lead to conflicts like it did during the Iran-Contra affair.” – George Shultz
• “I think that when you grow the NSC to a size that is historically out of the norm, you’re going to increase the temptation for a larger NSC staff to move into the business of execution or micromanagement. [Former Deputy Defense Secretary] John Hamre used to say, ‘If you want to make a staff more strategic, cut it in half.’ So I think a smaller NSC staff is one that’s necessarily more focused at the strategic level and on oversight rather than on execution.” – Michèle Flournoy
• “When I came on board at the very beginning [of the George W. Bush administration], we thought the NSC staff had gotten too big, so every directorate was asked to cut staff. … Directorates were in the process of downsizing when 9/11 happened, and many directorates were allowed to keep staff to handle the post-9/11 period.” – John Bellinger
• “The White House in the 21st century feels like it has to centralize more and more of not just the decision-making but the policy execution and choreography. Who’s in charge of cyber? Who’s in charge of CVE [countering violent extremism]? It’s a little bit of everybody. So that means it’s actually the White House. There is more political exposure, more media, and more sense of crisis management. These trends, in addition to creeping bureaucracy, led to an NSC staff that is bigger, more junior, and with more elevated titles. The large numbers and title inflation aren’t helpful because it damages the gravitas you need for the president’s advisors.” – Juan Zarate
• “The NSC staff’s job is to make sure that ultimately the president gets all the options, all the information, and all sides of the issue. That’s the important job that I think only the NSC can do. … This kind of NSC is a small NSC. You either have people who have interagency experience—I am shocked that there are people at the NSC staff who don’t have interagency experience—or you train them.” – Richard Clarke
• “When you have so many people on the National Security Council staff, you are then engaged, in my judgment, in a very dangerous situation. Dangerous, because the NSC staff defines the assumptions, draws the conclusions, presents them for decision, and then executes them—the executive bodies of government, with their inherited expertise, are largely marginalized. And if you have the same person defining the policy, making the recommendations, and executing it, then the policy is going to be defective. … I believe a process that doesn’t permit the career services to have a full crack at the problem before it moves on and the necessary political feed takes place is bad process.” – Frank Wisner
• “The larger the size of the NSC apparatus, the more likely it is that you are working for a president who has a view of the primacy of his staff and presidential power. On the other hand, the smaller the staff, the more likely you are working for a president whose view is—the authority and responsibility lies in the Cabinet agencies that report to him.” – Frances Townsend
• “I think the enlargement of the NSC is driven by the information age we are in today and the increase in the number of issues that have to be dealt with so much more quickly. Often, it is difficult to get the necessary information from agencies quickly. Again, the demand signal from the president. Ideally it should be supplied with information from the agencies, but the pressure is put on the NSC staff to respond. Twenty years ago, you had distance from issues and the time to deal with them that you just don’t have today. The NSC must make sure that the personal equities of all the agencies are represented, and this information has to be delivered to the president.” – Tony Blinken
• “The transition allows there to be the time to figure out if the current structure is the one you want. I think the next time is going to be unbelievably hard with 400 people there and trying to figure out who should stay and who should go. Raise this now because this is going to be the issue. The current staff will say that the world is much more complicated and you need all these people.” – Madeleine Albright
• “The increase in size carries the risk that the NSC is becoming tactical. You mitigate this risk by leadership, by the national security advisor defining to the staff its role and what’s not its role, and then enforcing that as you go through your day-to-day activities.” – Michael Morell
• “The larger the NSC staff is, the more the NSC staffers have to fill their time. The more they have to fill their time, the more they intrude into the actual work of departments and agencies. I do think that NSC offices ought to be relatively small, staffed by people with sufficient status and managerial skills to run processes and present options. … Most departments and agencies and people at the White House think that the NSC staff is too large right now. … When I first started working at the NSC, the Near East office, which was from Marrakech to Bangladesh, had a senior director and three directors. Now there are at least three senior directors, each with numerous directors reporting to them, and one additional special assistant to the president. And the staff are holding constant meetings. When I started, we had interagency working groups and they were chaired by senior directors. … Now this whole notion of sub-IPCs is just all over the place.” – Rand Beers
• “My own personal view is that the NSC staff is much too large. I am told by people overseas that they really don’t pay too much attention to delegations unless they are headed by people from the NSC staff. That just would not have passed muster in previous administrations. … The staff needs to be large enough to serve its coordinating and advisory functions but not so large that it has time on its hands to become operational or tactical. Striking that balance has to be a reflection of the system and process that the new president wants to set up. I would err on the side of smaller rather than larger.” – Robert Kimmitt
• “The larger the staff, the more the temptation to become operational. You’re supposed to make policy and oversee implementation. You’re not supposed to implement it yourself. I think that tends to get lost the larger we build staff. The larger the staff, the larger the layering.” – Frank Miller

NSC staff structure

As alluded to by Frank Miller above, our interviewees expressed widespread concerns about the relatively recent, excessive layering of the NSC staff. The NSC staff has always had two tremendous comparative advantages when compared with the national security departments and agencies: its proximity to the president and its organizational flatness. The proliferation of deputy national security advisors and the layering of senior coordinators ranking above some senior directors have all contributed to the creeping bureaucratization of the NSC staff, diminishing the flat organizational structure that served previous presidents well.

• “I have a certain degree of agnosticism about the number of NSC staff, but I am not agnostic about the layering. I think it is terrible. There is no reason for layering. As a practical manner, this should be a staff of a small number of experts in each area who are peers as opposed to empires. I think there are too many deputies, too many layers, and I think that defeats the purpose of what the NSC is about.” – James Steinberg
• “There is no good reason to have multiple layers in the White House staff. It ought to be responsibly flat. I think smaller and more senior is better. Numbers are not going to win this game necessarily.” – Frances Townsend
• “We have constructed a much larger and more layered national security structure since 9/11. But it is important to remind ourselves that the challenges we are facing are certainly not tougher than containing Stalin and Mao at the start of the Cold War or surviving the Cuban Missile Crisis in 1962. Ultimately, the NSC system is only as effective as the president that leads it.” – Nicholas Burns
• “The other part that is odd at the moment is the number of deputy national security advisors there is.” – Madeleine Albright
• “The impact of the proliferation of deputy national security advisors is that you get a proliferation of meetings and demands on the Deputies Committee members, most of whom are also charged with significant management responsibilities in their home agencies. I understand that you may want more than one deputy NSA from a span-of-control perspective, especially if a certain area like counterterrorism is intense and has its own high op tempo. I understand that argument, but I think the proliferation of deputies should be the exception, not the rule. We should aim for one and accept two if you have to, but only for a compelling reason.” – Michèle Flournoy
• “Susan Rice has some strong views on the fact that today we have too many layers in the NSC and that senior director positions must be filled by senior people or people of stature. The counterargument, however, on the coordinator layers is that arguably there are so many different equities in these spaces such as Middle East and nonproliferation, it’s helpful to have people within the NSC to bring the different strands together.” – Tony Blinken
• “The thing that was interesting to me was I could never get an organization chart for the NSC staff. Part of it was that they were so sensitive of the criticism they were getting for being too big that they didn’t want to show an organization chart. I just want transparency from the NSC staff.” – Sandy Winnefeld

NSC staff personnel

The consensus best practice, expressed often in our interviews, is for NSC senior directors to possess a personal stature that matches their positional stature. That is to say, if they are to lead work at the assistant secretary level, they would benefit from already having been an assistant secretary, ambassador, or the equivalent—or being widely perceived as being qualified to hold such a position. It is advantageous for senior directors to have previously worked in more than one department or agency. It is rare for a senior director to be successful in their role without having worked in government at all prior to joining the NSC staff.

• “The other thing I found when I came that was concerning to me was that the seniority of the senior directors had diminished rather significantly. … Senior directors are the counterparts of assistant secretaries. And directors are the counterparts of deputy assistant secretaries. If you have junior foreign service officers in those roles, they don’t have the experience, the breadth, the juice to engage effectively with their assistant-secretary-level counterparts in the interagency, and that puts the NSC at a point of weakness. … I have really tried to ensure that the people to whom we are giving these very important roles are people who have the background to carry it off. They have experience not only on the substance of the portfolio but also experience in managing the interagency and working in the bureaucracy. I would hesitate to put somebody in as a senior director who had not served previously in government. I can’t say that I would never do it, but it’s a pretty high bar.” – Susan Rice
• “We had very high-powered senior directors when I was on the NSC staff. Senior directors were ambassadors. You have more experienced people with broader and longer perspectives who are going to be able to prepare these issues. Having senior people will also help prevent issues from going up the chain. Now, you’re sending issues up what previously would have been an issue for a senior director.” – John Negroponte
• “You want a senior director and a director to be extraordinarily substantive. They need to know the issues in depth and should have vast experience working on these issues. It’s the way it used to be in the government. For example, assistant secretaries were steeped in their region. We have lost this over time. I don’t know why, but we have. Senior directors have to have stature. Deputy national security advisors have to have stature. You get that through experience and the jobs you have had. The senior director position is not a stepping stone for senior jobs, it is a follow-up from one of those jobs. When I am looking at State Department people on the NSC staff sitting there in meetings, I expect to look at ambassadors.” – Michael Morell
• “It’s hard to have an effective NSC staff without a national security advisor and deputy who have served on the NSC staff or at least participated regularly in NSC meetings on behalf of a department. You will have a better senior director if they have been a director before so they are not just reading the think tank papers but have seen for themselves what works and what doesn’t. … If you are serving on the NSC staff for the first time and you don’t know how to make the system work, then all you can do is sit there and bang on your high chair and say this is what the president wants.” – John Bellinger
• “Be careful about hiring people on the NSC staff who are in their first position in government because they may know more about human rights or nonproliferation than anyone else, but they have to have that interagency experience.” – Robert Kimmitt
• “You have to have senior people who are subject matter experts with enough depth, expertise, and networks to actually be in a position to be a source for ideas and to be able to check and run the bureaucracy. They must also be able to provide candid advice to the president and the national security advisor. So the deputy national security advisors and the senior directors ideally are seasoned professionals in their space. … This person must have extensive experience in actually running some element of a government agency to understand how policy works.” – Juan Zarate
• “I think it’s very important for someone to have prior agency experience or prior government experience. You need an appropriate level of experience and seniority. If you’re running an interagency working group, you’re calling together Senate-confirmed assistant-secretary-level personnel, you need to have the stature and the heft to do that. It’s not so much a question of age as it is about asking, ‘Does this person have the experience and command respect among their interagency peers to lead the interagency process in a given area?’ And that is important to make the level below the Deputies Committee work.” – Michèle Flournoy
• “I don’t think there is a place for really junior people on the White House staff. I have nothing against young, junior people, but I just don’t think that is the place to have them other than logistics, support, and administration. From my perspective, those I have benefited the most from having worked with on the White House staff were people who had been out and done things in government, think tanks, who had been out in the field, implemented policy or programs, who had had the opportunity to think long and hard and write and do field work.” – Frances Townsend
• “It is important for people who have stature and are respected to be in these jobs. I don’t agree that junior jobs are for smart but not knowledgeable and inexperienced people.” – James Steinberg
• “You can have very smart people with very smart ideas, but it is much harder to find very smart people with very smart ideas that can go out and get things done. When you can marry the two, you have the ideal person. People who understand the art of government, who can bring people together, know how to package the decisions, know who to trust and lean to. That precious combination of intelligence and practicality, those are the right kinds of people for the NSC staff.” – Frank Wisner
• “Each administration is a mix of fresh blood, often new to high policy. Their energy is important. But you need to have seasoned people who understand intuitively how the system operates. You need to have the capacity for creative re-examination of the facts, an ongoing capacity to have a second look at what the hell is happening.” – Leon Fuerth
• “The senior director has to be someone of sufficient and independent stature that an assistant secretary would come to a meeting called by the senior director. … Should that person have been an assistant secretary before, or should that person have been a senior military officer or an ambassador in order to run those meetings? All of those are valuable stature issues. They have to be meetings that are worthy of people of that level.” – Rand Beers
• “The rule of thumb I thought had pertained was that the senior director and special assistant to the president has been an assistant secretary. … A lot of the implementation problem and a lot of the confusion and waste of time is because the NSC people don’t have experience in the departments. They don’t know what to ask and who to ask, making them learn on the job at the expense of everyone involved in the process. … Condi [Rice] came in and said the NSC staff is not the place where you learn on the job. I don’t understand the rationale where you make somebody the special assistant to the president who has never worked in any department. It ought to be a given that the national security advisor and the one deputy national security advisor—I don’t think we ought to have more than one—and all the senior directors should have in the past been part of the NSC system. That doesn’t mean they should have worked in the White House, they have had to either be NSC staff or represented a department in the NSC system.” – Richard Clarke

Federal departments should be encouraged to make their very best personnel available to serve on the NSC staff and then make superior positions available for them upon their return. Unfortunately, that often is not the case.

• “At the State Department, I was deeply involved in the personnel process for a variety of reasons. I can’t tell you how many meetings I went into where we were selecting career foreign service officers to be nominees for ambassadorial positions, consuls general, or deputy chiefs of mission, and I would hear someone say either from the regional or personnel bureaus that a particular candidate’s career was going great, but then he or she spent three years ‘off track’ at the NSC or at the Defense Department.” – Robert Kimmitt

Formal training on best practices in management, policymaking processes, and presidential support should be provided to incoming NSC senior directors and directors. The current Obama administration initiative to provide such formal training should be continued in the next administration.

• “Our focus on ways to reform and restructure the NSC resulted among other things in new initiatives which provided additional and more structured orientation for staff, a revamped system for training, an updated handbook on NSC processes and resources, and an increased emphasis on professional development. We took all of these steps in an effort to ensure we are fulfilling the NSC mission as effectively as possible.” – Suzy George
• “There has to be initial training. When I was appointed assistant secretary of state, I had to go to a class. It was Jim Baker’s idea. I thought I knew how this all worked—but no, I didn’t. There was stuff I learned in that class. … I didn’t get any training when I came on the NSC staff.” – Richard Clarke
• “What I hope to do is to be able to hand off to my successor an NSC staff and process that is well-oiled and functioning and that represents my best recommendation on how to do things. I want to leave a turnkey operation that has predictability, that has paper flow, that has a process that’s working. I want to give my successor the benefit of not having to worry about that on day one. They should inherit a well-functioning apparatus that can sustain him or her until they have the bandwidth to decide to relook it.” – Susan Rice

Our intention in undertaking this work has been to assist a new administration in making deliberate choices about how to structure the process of supporting the president on national security issues. We engaged with high-level practitioners of both parties and let their voices carry the recommendations of best practices for structuring and running the policy process. Our strong belief is that transition teams best serve the president by matching processes to the president’s own management style and advocating for senior appointments who will consider national security policymaking a team sport. As they design their internal processes, they should apply best practices drawn from the experiences of their predecessors. We hope this report will assist in the difficult and important work of transition.

Kori Schake is a research fellow at the Hoover Institution. She is the editor, with Jim Mattis, of the book Warriors and Citizens: American Views of Our Military. She teaches “Thinking About War” at Stanford, is a columnist for Foreign Policy magazine, and is a contributor to “War on the Rocks.” Her history of the Anglo-American hegemonic transition is forthcoming from Harvard University Press in 2017. She has served in various policy roles, including on the National Security Council at the White House; in the Office of the Secretary and Joint Chiefs of Staff at the Department of Defense; and on the Policy Planning Staff at the State Department. During the 2008 presidential election, she was senior policy advisor on the McCain-Palin campaign. She has been profiled in publications ranging from national news to popular culture including the Los Angeles Times, Politico, and Vogue magazine. Her recent publications include: “Republican Foreign Policy After Trump” (Survival, Fall 2016) and “National Security Challenges for the Next President” (Orbis, Winter 2017).

William F. Wechsler is a Senior Fellow at the Center for American Progress, where his work focuses on counterterrorism and U.S. national security policy in the Middle East, South Asia, and Africa. His most recent position in government was deputy assistant secretary of defense for special operations and combating terrorism. Prior to that, he served as deputy assistant secretary of defense for counternarcotics and global threats. During the Clinton administration, he served as special advisor to the secretary of the Treasury, director for transnational threats on the staff of the National Security Council, and special assistant to the chairman of the Joint Chiefs of Staff. He has contributed chapters to two edited volumes and has been published in The New York Times, Los Angeles Times, Foreign Affairs, and the National Interest. He is a graduate of Cornell University and received a master’s degree from Columbia University’s School of International and Public Affairs.

The authors could not have written this report without the help of several research assistants from the Center for American Progress and the Hoover Institution, including Ayla Bitmouni; Christina Huchro; Shaun Kleber; Andrea Sarahi Zaldumbide; and especially, Sangita Kanumalla.