The American Interest
Policy, Politics & Culture
object(WP_Session)#92 (5) { ["session_id:protected"]=> string(32) "b03955df1751efea0ffefe721781f7c5" ["expires:protected"]=> int(1414095877) ["exp_variant:protected"]=> int(1414095517) ["container:protected"]=> array(1) { ["ai_visit_counter"]=> int(0) } ["dirty:protected"]=> bool(true) }
Managing to Fail

The NSC/interagency system can’t cope with a multi-dimensional strategic environment.

Published on September 1, 2007

Most policy-relevant debate about U.S. national security focuses on strategy. Other important discussion centers on getting the right people in the right offices at the right times. The assumption behind all such debate, some of it very sophisticated, is that if we have the right concepts and experienced, competent people to implement them, the organizational structure of government becomes, if not entirely irrelevant, then certainly a less significant factor in foreign policy outcomes. The bias of experienced Washington insiders is clear: Good ideas and good people can overcome a creaky organizational structure, but even the best structure will fail if we lack good ideas and good people.

That may have been true in the past (or not). But it is true no longer. The higher organization of U.S. national security is broken. (By “higher organization” I mean the integration of the senior echelons of critical Executive Branch departments: State, Defense, the National Intelligence Directorate, Homeland Security, Treasury and Justice, to name the most obviously significant.) The National Security Council system and the interagency process we currently use do not function effectively—as has been manifest throughout the presidency of George W. Bush—not because (or not just because) of a deficiency of strategic acumen and experienced policymakers. The system itself is inadequate to the world as we find it. The old Washington insider bias needs to be turned on its head: The higher organization of national security is so dysfunctional that it almost doesn’t matter what strategies we select or how individually brilliant our policymakers are; they will fail.

The costs of structural dysfunction will get worse, too, because both problems and opportunities in the global environment are becoming increasingly diverse and multidimensional. Handling a nuclear Iran and fighting terrorists, rescuing Darfur and rebuilding failed states, managing the entry of new powers into the international order and protecting against the meltdown of the international financial system—all of these new challenges demand integrated approaches the current system cannot deliver, designed as it was for a Cold War world that no longer exists.

This argument is not entirely new. After the Cold War several observers reasoned that a new environment called for a new strategy. As National Security Advisor, Condoleezza Rice noted that, while the first national security decision memorandum of the George H.W. Bush Administration went out to 14 government departments and agencies, George W. Bush’s first went out to 41. The Hart-Rudman Commission (1999–2001) had precisely such a task as its mandate, and by most measures it did a good job. It called for: a homeland security department (though not the one we got) to rationalize domestic preparedness; major reform of the Executive Branch’s strategy and budget processes; and reform of the NSC and interagency processes.

Praised though the Hart-Rudman report was, few of its far-reaching proposals were taken seriously because, we are told, of “politics”—politics apparently being a catch-all term to describe everything from poor leadership and sheer dim-wittedness on the one side to cynicism and venality on the other. Alas, it took a national disaster for the commission’s homeland security proposals even to get a hearing. And “politics”, thus defined, is why we have seen no effort toward significant structural reform, despite the widespread conviction among former and current senior officials that the system doesn’t work anymore.

Out of style? A meeting of the National Security Council, October 2001.

It doesn’t have to be this way. No system is unchangeable, as our own history proves. Just imagine where the United States would be had Elihu Root, Theodore Roosevelt, Harry Truman and Dwight Eisenhower accepted the bureaucratic arrangements they inherited with the excuse that “politics” made change impossible. Their examples can inspire us, and at least some design principles for new organizational mechanisms may be gleaned from several large companies that have successfully faced the challenges of globalization in recent years.

A New Strategic Environment

Major changes in U.S. security organization have occurred in recent years, to be sure. Jointness in the armed forces has proceeded nicely. The Unified Command Plan has been redrawn as needed: A new Africa Command is being created. A host of agencies dealing with homeland security were organized into a new Cabinet department, and the intelligence community recently started its own version of a Goldwater-Nichols reform, with rotations around different intelligence agencies being a condition for promotion.

But these changes all focus on better integration within the big “chunks” of the national security establishment; they don’t affect the higher organization that links the chunks together. The larger coordinating superstructure hasn’t changed appreciably since 1947, the year it was designed for the Cold War. Consider for a moment the effective parameters in which it worked and evolved.

Starting in 1947, policy depended on a consistent enforcement of containment, nuclear deterrence and alliance leadership. The number and diversity of U.S. opponents was limited to Communist bloc countries. The variety of problems was also low: Strategic theory dealt with the relationship between conventional and nuclear forces, but did not need to handle the wider issues of terrorism, homeland security, the spread of missiles and counterproliferation. Nor did it have to manage as wide an array of non-military security threats ranging from failed states to climate change.

The best way to deal with a strategic environment characterized by a relatively low degree of diversity was to build strong departmental silos that specialized in the complexities of a relatively small number of major problems. In the Pentagon this meant separating nuclear and conventional operations and command and control in order to enforce deterrence. In the State Department it meant having a global diplomatic containment strategy. In the intelligence community it meant a huge, fixed-cost sensor system optimized for use against a slowly changing Soviet threat.

There are always variations and exceptions that challenge any coordinating mechanism, of course. Priorities must be chosen and turf tangles de-conflicted. However, broadly speaking, the required level of integration across the ranks of Executive departments and agencies was modest. Most problems that threatened to fall between the cracks of the bureaucracy could be solved in a single NSC meeting, or at deputy-level interagency meetings. Organizational problems had end-points.

Today’s environment is fundamentally different. The diversity of international actors is much greater. There are now nine nuclear weapon states and counting. How some of these states think about strategy and weapons of mass destruction is for all practical purposes terra incognita. The rise of China and India, the return of a more powerful Russia and a changing Japan also point to a new multipolar international order.

There are more and varied problems, too. From terrorism to climate change, solving these problems requires a greater range of skills in government decision-making. The individual silos that make up the establishment have been adapting, to be sure. The Pentagon takes terrorism as a serious threat, the State Department understands the political significance of Darfur, and so on. The problem is that most of these challenges require skills and energy from several departments working together to make any headway, yet the NSC/interagency system was never designed for that kind of integration.

The result is that the NSC system has invented workarounds using informal, ad hoc practices and bureaucratic entrepreneurs to get things done. This has worked, to a degree, but on really big integrative challenges—simultaneous warfighting and nation-building in Iraq, for example—workarounds just don’t work. The U.S. performance in Iraq demonstrates the glaring limits of such improvisation.

Why Organization Matters

It is amazing how many people in Washington believe that higher organization doesn’t matter much, that the right strategy and the right people will get the job done, regardless of what the organizational chart says. Clearly, such views are strongly held, especially among current and retired senior officials who draw on their personal experience in government to describe “how things really work.” These are people who have usually had their fill of futile government reorganizations—efforts that didn’t succeed, made no sense from the start, or whose purpose was merely to create the appearance of change after some disaster. So their sentiments are understandable, if misplaced.

Another reason bureaucratic veterans may discount the importance of organizational design is that bad design never threatened to put them out of business. Defense, State and Treasury are all organizational monopolies. They can fail at policy but they can’t be put out of business by competitors. Not so in the world of business.

Peter Drucker argued that one of the main ways large corporations prove to be more productive than small shops and entrepreneurs is by out-organizing them. Companies such as DuPont, Sears, Toyota, IBM and Siemens harnessed their workers’ skills in corporate structures to achieve their strategic objectives.11.
Drucker, Managing for the Future (Dutton, 1992), chapter 13.
Without an effective high-level structure, the modern corporation could not exist. Thus Toyota fought its way into the American market in the 1970s with reliable, low-priced fuel-efficient cars using a structure called the Toyota Production System, which coordinated its corporate structure into its manufacturing. Other automakers knew what the market wanted, but only Toyota had the organizational system to actually build cars to match the market assessments.

Structure is such a key competitive weapon in business that it is hard to imagine its being unimportant in national security. The same goes for military history: Note the genius of the German General Staff in fighting well while outnumbered; Halsey’s structuring of the Pacific Fleet; Eisenhower’s organization of the war in Europe. Indeed, there is a fascinating interplay between organizational innovation in big business and in big militaries, and it provides us a useful source of insights about how to structure strategy in “big governments”, too.

One of these insights concerns the mutual effects of strategy and policy execution. It is self-evident to most people that it is strategy that drives structure and hence execution: One first decides the goal and the basic organizational methods to achieve it, and only then does one put the strategy into motion. This thinking may be self-evident, but it is also wrong. Experience in both business and war shows instead a subtle interplay between strategy and structure, which, if overlooked, leads to a grand strategy so abstract as to be useless. As IBM CEO Lou Gerstner put it when taking over IBM in 1993, “the last thing this company needs is a vision.” Gerstner understood that one could change structure in such a way as to clarify and expand strategic options; in other words, that innovating one’s means could reflect back on one’s ends.

In the 1980s IBM gave away services as a means to sell big computers. Services were considered “sales support” for the hardware business. Gerstner thought IBM could make a lot more money selling services than by giving them away, so he greatly expanded a separate services division that IBM had set up earlier. By making it easier to buy services without buying associated hardware, service sales began to take off. With a separate services department, IBM could consider strategies that were hardly imaginable before and led to a rethinking of the entire corporate strategy.22.
Gerstner, Who Says Elephants Can’t Dance? Inside IBM’s Historic Turnaround (HarperBusiness, 2002).
That rethinking, in turn, put IBM in a position to see the potential of the Internet. IBM Global Services saw that wiring Starbucks stores for Internet access opened up a whole new region of strategy to which the old IBM would have been structurally blind.33.
See John R. Patrick, Net Attitude (Perseus, 2001).

In theory, IBM could have done all of this by ignoring structure and approaching its challenge purely in terms of strategy. Processes could have been set up within the hardware organization to cross-sell services. But only in theory: Power dynamics in real organizations work against functional innovation within the same structures. Without changes in higher organization, the connection between means and ends often isn’t clear. Only when IBM changed its structure, busting out of the hardware silo, did it open up a rich new menu of strategic possibilities.

The Lesson of Iraq

The war in Iraq offers the starkest example of the dysfunction now built into current U.S. national security organization. Nearly everyone agrees that far more than military power was needed for the United States to have achieved its stated goals in Iraq. Even without an insurgency, job creation, policing, electricity and health services were going to be as critical as security to overall success in standing up a viable democratic Iraqi polity. The insurgency just made an already daunting agenda far more urgent. Virtually every study of modern insurgency, too, concludes that force alone cannot defeat an insurgency once it has set social roots.

The U.S. government has spent about $60 billion so far on reconstruction in Iraq and gotten little for the investment. With the same set of institutions—DoD, USAID and their contractors, all coordinated by the NSC system—we could have spent four times as much money and done no better. An insurgency cannot be defeated by a rigid, linear plan: first, security; second, jobs, policing and health systems; third, democratic politics. It can only be defeated if these three features are achieved to some degree in parallel. But it is hard to see how that can happen when the U.S. Army and Marine Corps are doing the “economic stuff” they were never trained for because no one else is available to do it.

The problem is much deeper than a failure to “plan.” It is a design failure extended over many years, not to have thought about which U.S. capacities would be needed over a range of likely contingencies. Contrary to common belief, there was much good planning in the months before the war to address the post-conventional phase of the conflict, but those plans made little practical difference because no one was there to implement them. These plans bore some resemblance to the good ideas in the 1970s and 1980s auto industry about fuel efficiency and low prices. Many companies had these ideas, but only Toyota structured itself to be able to implement them.

The deficiencies of USAID are a particularly sad case in point. In the 1960s and 1970s, USAID was on the cutting edge of a multi-track approach—integrating political, security and economic resources—to economic development under demanding conditions of insurgency. Some readers will remember John Paul Vann, a retired Army officer who, as a USAID official in Vietnam, saw the need for a multi-track approach to counterinsurgency.44.
See Neil Sheehan, A Bright Shining Lie: John Paul Vann and America in Vietnam (Random House, 1988).
After the Cold War, however, Congress cut USAID to a shell of its former self. It has largely become an outsourcing mechanism, and most of its experienced staff is gone. USAID officials now know a great deal about government contract regulations, but the organization itself has forgotten what it learned over more than three decades, often the hard way, in Vietnam and Latin America. USAID’s decline stretched out across several administrations of both parties. It went largely unnoticed, and the implications were never apparent until USAID was called on to deal with Iraqi reconstruction—and simply couldn’t.

This isn’t the fault of USAID, but of the macrostructure that allowed its decline to happen. In government and business, it’s a fact of life that some departments will fall off in performance. Higher management’s job, in large part, is to determine and monitor critical abilities that must not be allowed to deteriorate. The U.S. government failed to do this. The NSC system did not break USAID’s fall and, worse, didn’t see capabilities preservation itself as one of its responsibilities. The NSC sees its job as providing smooth administration for the White House. Its focus is on the next good move rather than on the relationship of moves to each other over time. A business analogy would be if Toyota’s Production System fell apart, with the CEO’s staff claiming it had more “urgent” matters to handle.

It isn’t even quite fair to blame “the CEO”—that is, the President of the United States. Presidents inherit a macrostructure, and it is unrealistic to expect a typical president to make sense of developments that have accreted over many years, long before he entered office. The problem is that higher organization is no one’s responsibility. The current default coordinating mechanism handles problems of unorganized complexity by dropping them on the president’s desk. The next president will face the same challenge, but without a structure to point out the consequences of a decline in the performance of important capabilities, practical mechanisms for horizontal integration, and some way to pull all the pieces into a comprehensive picture, he or she will not succeed either.

The Environment Demands It

So far, the U.S. government’s failure rate due to faulty higher organization has not been lethal. Few other large governments have done any better than the United States, certainly not the late and not very great Soviet Union. Besides, the challenges that arose between the cracks of our silos were not existential in nature, either singly or in combination. But the failure rate is rising because a wide range of security problems requires a sharp increase in cross-department integration, and the current structure just cannot do this.

New power centers are emerging in India, China and arguably Russia. Dealing with them is probably the single most important issue facing the United States. To manage this new international system we need strategies, of course. But we also need structures that reveal hidden issues. Without such structures, strategy is liable to meander, being dependent on political personalities or the outcomes of bureaucratic struggles.

Take the question of whether the U.S. government should sell the F-22 stealth fighter to Japan. This is an important policy decision, not least because it affects China. Beijing could see such a sale as a U.S. effort to rearm Japan, in turn provoking an arms build-up in East Asia. But the sale would have far wider effects, impinging on how China, India, Russia and Japan see the global balance of power, not to mention the effects of technology proliferation on the global market.

For our purposes, the point is not to debate the wisdom of such a sale. Our concern is that the 1947 system cannot reasonably be expected to assess that wisdom. State, Defense and CIA are organized essentially by geography. Yes, State does have a Bureau of Political-Military Affairs and Policy Planning, and Defense has an Office of Net Assessment, but these non-geographical bureaus do not have the clout or the day-to-day operational momentum of their regional counterparts. For that reason it is hard even to uncover the broader impact of a military sale, or of any advanced technology, in this structure. Such geographic organization once made sense, but it can no longer handle the integration of policy for China, India, Japan and Russia as a group.

The NSC system is supposed to be able to rise above geography because it is small enough to cross-talk on a regular basis. But the NSC is itself organized largely geographically, too, having been modeled after the geographic bureaus of its departmental clients. There is no “Bureau of Multipolar Order” in any of them, or in the NSC either. Multipolar strategies are simply not likely to be well conceptualized in the current set-up, just as IBM could not see its own opportunities when services were straightjacketed in a hardware division.

Global businesses have faced a similar challenge. When U.S. business first went overseas, the idea was to hire country managers, people who spoke the language and understood the local context. Country managers were the experts and were left pretty much alone. A corporation’s “global strategy” consisted, in effect, of individual country strategies stapled together.

No longer. Successful companies have developed global business strategies that allow them to master the challenge of international supply chains, which requires effective cross-border integration. The emergence of Brazil, Russia, India and China—the so-called BRIC markets—also called for a more integrated approach. Since similar developments are taking place in each of these countries—demand for infrastructure, consumer products, credit cards and mobile phones—there were economies of scale and knowledge to be gained from looking at the BRICs as an integrated whole. (This, actually, is the origin of the concept of the BRICs.)

Business has thus responded to globalization with a new emphasis on global integration. For example, GE cut its 11 strategic business units and moved them into six larger divisions: GE Infrastructure, GE Commercial Finance, GE Industrial, and so forth. GE Infrastructure sells aircraft engines and power plants to all of the BRICs, and likewise for other divisions. GE wanted to bust out of the old country and product silos to build a mechanism that brought to bear all corporate resources for the key big markets. Again, a change in structure allowed and propelled a change in strategy.

Global integration isn’t the only challenge presented by the security environment. In 2007, the Bush Administration sent two aircraft carriers and a large naval force to the Persian Gulf as tension rose over Iran’s nuclear program. In addition, the Office of Terrorism and Financial Intelligence in the Treasury Department placed financial sanctions on Iranian companies involved with their atomic program. It pressured Western banks not to deal with Iran, and further efforts are now underway to “unplug” Iran’s banks from the international bank payments systems, a kind of excommunication from one of the world’s most critical networks. While all this was going on, the National Security Agency may have been, shall we say, “testing” Iran’s vulnerability to cybernetic disruptions.

Again, our purpose here isn’t to determine whether these are good or bad policies. The point is that a policy of escalating pressure is being applied against Iran, involving at least two and probably more than two different pressure points. Yet it is by no means clear that we have a mechanism adequate to coordinate a strategy involving the Defense Department (civilians and at least one uniformed service), the Treasury Department, and at least one arm of a functionally differentiated intelligence community.

Escalation, whether against Iran or any other state, involves a complicated set of interacting considerations. These range from psychological operations to the construction of “off ramps” should the adversaries not want to go all the way to a major confrontation. To divide this case into separate strategies of military escalation, on the one hand, and financial warfare, on the other, overlooks the points at which both strategies can complement one another. A failure to integrate the different moving parts of the policy risks sending the wrong signals to Iran. It could even inadvertently produce an Iranian “explosion” against U.S. forces.

In the Cold War there was pretty much only one kind of escalation: military escalation—in the forms of alerts, deployments and the like. Financial warfare against Moscow was impossible because the Soviet Union was “off the grid” of the world economy. As the scale of the world financial system has grown, however, the opportunity for financial warfare escalation has as well.55.
See my “Financial Warfare”, Orbis (Fall 2007).
U.S. financial sanctions have put the screws to Pyongyang, among others. Other financial measures have been taken against Syria. Financial warfare, in short, is a growth business.

There are still other avenues to ratchet up the pressure—cyber warfare, for example. Computer network attacks against Estonia earlier this year were timed suspiciously soon after the removal of a Soviet-era military monument from the square in Tallinn. Were the cyber-attacks an officially sanctioned manifestation of Russia’s increasingly heated rhetoric? Who knows for sure? Again, the point is that even a weakened state like Russia can now mount multidimensional escalation strategies.

This state of affairs raises two questions: How, given the current high-level organization of U.S. security, would the United States respond to a multidimensional attack against it, and how would the United States mount its own multidimensional attacks? Who would handle multidimensional escalation in the U.S. government? Can the current system even foster coherent thought about such a strategy? As things stand now, the only way the NSC system can manage incoming or outgoing multidimensional attacks is to disaggregate the elements and assign them to existing silos. The White House/NSC system would give the Pentagon responsibility for the military part, Treasury responsibility for the financial part, NSA (or some other agency) responsibility for the cyber warfare part and so on. But this won’t work. Anyone with experience in large organizations knows that unity of effort and unity of command are two very different things. Escalation shouldn’t be decentralized. Leaving escalation to individual agencies, or only loosely coordinating them with tacit understandings and informal review, is intrinsically dangerous. Escalation requires taking calculated risks, yes, but the calculations have to be done.

With the growth in global finance and the pervasive use of cyberspace, the very character of strategic escalation is changing, requiring new kinds of horizontal coordination. “Winging it” isn’t likely to work any better than improvising economic reconstruction has worked in Iraq. The same applies to de-escalation, a topic that has received even less attention than escalation. Just as the decline of USAID was “missed” by the NSC system, the “escalation and de-escalation business” could easily be overlooked or dumped on an unprepared president’s desk with a plea that he (or she) “figure it out.”

Another example from the corporate world may be instructive here. There have been truly massive changes in the way companies manage risk in a more complex financial and global business environment. Doing business in the BRICs is vastly more challenging than doing business in traditional markets. The laws, intellectual property protections, liability and risks to corporate reputation are all very different. Business learned that it was dangerous to do risk management in these markets as it was done in traditional ones.

Businesses learned, too, that the very act of drawing a risk map of a new venture or strategy is a useful exercise in itself. Following the criminal indictment of Arthur Andersen in 2001, the accounting firm Ernst & Young established a new global unit to manage risks and ensure common standards of quality in all its divisions. The Ernst & Young risk group ensures that processes for reporting, government relations, and protection of reputation are used around the world. It meets regularly with country managers and with the top global leadership.

The Ernst & Young group is a separate team overlaying the divisions; it is not made up of senior representatives from the major countries reporting to the CEO through his staff. That insures an unobstructed view of the global practice, mirroring the insights of Edward Deming, the quality control guru who decades ago advocated the need for quality to be a separate reporting function. More and more companies are using this model in risk management, with some even designating a chief risk officer, or CRO, who reports outside existing structures to headquarters.

While risk management offers new thinking and techniques that are increasingly central to business, this revolution has barely touched scholars or practitioners of international security affairs. The closest we have ever come to this is “B-team” intelligence assessments at the CIA. Government still has a lot to learn from successful business practices, properly adapted to its own particular purposes.

Principles for Change

Simply rearranging the actors in the current system cannot supply the greater degree of integration needed in today’s environment. That would be analogous to paving over old country cow paths as a way to modernize a transportation system. Reshuffling NSC committees to include counterterrorism, counter-proliferation, alliance management and the like won’t get at the structural dysfunction of the current system: authority and information still flow from the NSC to the silos through representatives who work in them.

What’s needed is a new superhighway. Devising a new organizational design depends on the strategy of a new administration. It also depends on Congress, which has ultimate authority on major changes in government. There are four basic design principles that should clearly apply in any such effort.

Strategy without structure is a delusion. Any proposed grand strategy must have a carefully articulated set of institutional arrangements to execute it. The vast majority of what is called “national security strategy” is actually strategy formulation, while only a tiny fraction of such thinking deals with execution. Visions and grand strategy are fine, but without organizations that can implement them they are more pipe dream than policy.

As interdependencies increase, mechanisms for horizontal coordination must also increase. This is the core challenge: how to get better horizontal integration among departments and from national capabilities. Businesses have done this by breaking out of their old silos of line and staff. Customary reorganization proposals call for strengthening the NSC/interagency process. We need the very reverse. We need to replace major parts and functions of the NSC system with specialized teams overlaid on top of the silos, assigning them specific tasks and operational responsibilities. To illustrate, a small team formed to manage some task, such as multidimensional escalation or risk management, would not be made up either of emissaries from the silos or of the NSC. It would be independent and, of course, would report to the president or an assistant. The team would not be captive of the silos, nor would it be watered down by the process demands of the interagency system.

This is the trend in the corporate world. Companies are turning to fluid sets of teams that can break out of the old silos. The team-overlay approach provides a needed decentralization and focus. Each team, in theory, is a miniature of the whole security bureaucracy, through its appointment of individuals with problem-domain and silo knowledge—military, financial, cyber warfare, for example. But the teams are stripped down to just those capacities to focus directly on a particular problem. That way they stay small and effective. By decentralizing authority to overlay teams, decisions can be made more quickly and with better local knowledge. This approach flattens the organization to manage important processes and cuts down on issues getting diverted or badly delayed in NSC/interagency space.

The U.S. government has done this before, on a limited basis. The genesis of Plan Columbia is a good example. The White House will need to stand up many more such teams, particularly for problems that are either very important, time critical, or those with a potential for explosive eruptions in violence. Other problems can be left for a stripped down, better balanced NSC/interagency system.

These changes would require significant reorganization of the current structure. Congress would have to play a critical role, of course. It is important to get the current debate about organizing U.S. security away from the customary calls for greater cooperation in a system that has reached the limits of its integration abilities.

Smaller, more focused groups outperform larger, diffuse ones. This maxim takes the above principle and applies it to all levels of the national security hierarchy. Small groups are not only important at the pinnacles of national security, but down the line at the working level, as well. An additional failing of the current system is its inability to enforce cooperation at lower levels in the hierarchy.

One of the major reasons for Israel’s recent success against suicide bombers is from redesigned teams at the front line of defense. These teams have a boundary-spanning role. It is not uncommon, for example, for a Mossad officer to be in charge of a small unit military action involving simultaneous air and ground strikes directed at terrorists. The idea of an intelligence officer commanding military assets owned by other services adopts the same logic used in the above principle, but applies it at the point of contact of actual operations.

Another lesson to be drawn from this and the preceding principle is that any reorganization that tries to use the same structure but adds broad, diffuse goals to it is very unlikely to be effective. Some have advocated the creation of a counterproliferation committee in the NSC, for example. But as important as counterproliferation is, it tends to be diffuse. To some people it means cooperation with international organizations like the UN and the IAEA. To others it means military strikes. Overlay teams need to be small and highly focused if they are to be effective. An NSC counterproliferation committee would make things worse.

Horizontal skills and behavior need to be driven down into the silos. Companies have found that horizontal integration also requires changes in behavior. It takes people willing to look at problems in a certain way; entirely new career paths will be needed to make this work. Currently, it isn’t clear that there is any framework for pulling together, say, the way financial sanctions work in tandem with military moves, or the way others are likely to respond to multidimensional escalation.

A troubling development in this regard is the recent tendency of major government departments to create their own specialized universities. In addition to the Defense Acquisition University, a National Intelligence University and even a Homeland Security University have been proposed. Such narrow in-house education may only reinforce the culture of the silo, or worse, institutionalize rule compliance rather than problem solving. Most leading schools of management and public policy in the United States are moving in the opposite direction. Business schools, for example, are trying to break out of their traditional silos of accounting, finance and marketing to teach students better integrated approaches to problem solving.

Plenty needs to change if the U.S. government is to have any hope of keeping pace with a changing world. That said, hierarchy and channels have their place, too. Many problems are best dealt with by using the existing system, and certainly we should not make the mistake of equating better horizontal integration with a network organizational structure. In the media, and especially inside the Capitol Beltway, “networks” are now offered as a universal solution to nearly all problems of bureaucracy. Some even posit a cosmic struggle between our bureaucracies and “their” networks, with the implication that the U.S. security establishment needs to become networked like al-Qaeda. That’s just bunk. The unbridled substitution of networks for hierarchies will create organizational chaos, not efficiency.

Again we can take a lesson from business best practice. In business it is hard to discern what a network structure really means in practice. But we know what it doesn’t mean: It is not about connecting everyone to everyone else, and it is not about moving all power to the “edge of the organization”, to cite the latest “in” jargon, to those in direct contact with the problem. Both notions would have brought down IBM, and would most likely lead to unwanted war if escalation decisions were delegated to field commanders.

Higher-level organization involves a mix of strategy, people and structure. No one best answer works in all contexts, and government obviously differs from business. But government does need an answer that will work for its context, because the present structure cannot cope. The problem is not “too hard to think about”, and “politics” does not prevent change, unless we let it. Then “politics” becomes just another word for lazy.

Drucker, Managing for the Future (Dutton, 1992), chapter 13.
Gerstner, Who Says Elephants Can’t Dance? Inside IBM’s Historic Turnaround (HarperBusiness, 2002).
See John R. Patrick, Net Attitude (Perseus, 2001).

See Neil Sheehan, A Bright Shining Lie: John Paul Vann and America in Vietnam (Random House, 1988).

See my “Financial Warfare”, Orbis (Fall 2007).

Paul Bracken is professor of management and political science at Yale University and co-editor, with Ian Bremmer and David Gordon, of the forthcoming book Managing Strategic Surprise: Lessons from Risk Management & Risk Assessment (Cambridge University Press).