In recent years we have been witness to a portentous competition between two determined but dissimilar rivals on the international scene. In one corner we have al-Qaeda, founded in the early 1990s, the transnational Islamic terrorist organization led by Osama bin Laden. In the other corner, we have the government of the United States of America, established in 1787, at present the most powerful state on the planet. The key question defining this competition is this: Who has the more agile organization? Al-Qaeda, in planning and executing a terrorist attack, or the United States, in planning, developing and executing the measures to stop one?
Let’s look at the record. Sometime during the spring of 1999, Khalid Sheikh Mohamed visited bin Laden in Afghanistan and asked if al-Qaeda would fund what came to be called the “planes operation”—the plan for suicide attacks using commercial airliners. (Mohamed had been mulling the plot since at least 1993, when he discussed it with his nephew, Ramzi Yousef, one of the terrorists behind the first World Trade Center bombing and the attempted Philippine-based effort to bring down a dozen U.S. airliners over the Pacific in 1995.) Bin Laden agreed, and by the summer of 1999 he had selected as team leaders four al-Qaeda members—Khalid al-Mihdhar, Nawaf al-Hazmi, Tawfiq bin Attash (also known as “Khallad”) and Abu Bara al-Yemen.
These four team leaders entered the United States in early 2000 and started taking flying lessons that summer. The so-called “muscle” hijackers, the 15 terrorists tasked with overpowering the crews on the targeted flights, began arriving in April 2001 and spent the summer preparing for the September 11 attack. So from the point in time that a government contracting official would call “authority to proceed” to completion, the operation took approximately 27 months.
Now let’s track the U.S. response. U.S. officials began debating options for preventing future terrorist attacks immediately following the September 11 strike. Congress took a year to debate the statute establishing the Department of Homeland Security. George W. Bush, who originally opposed creating a new department, changed his mind and signed the bill into law on November 25, 2002. A joint House-Senate committee finished the first investigation of intelligence leading up to the attack in December 2002. The 9/11 Commission issued its report on July 22, 2004, recommending among other things the establishment of a Director of National Intelligence and a new National Counterterrorism Center. President Bush established the NCTC by Executive Order on August 27, 2004.
Adoption of the Intelligence Reform and Terrorism Prevention Act, which embodied most of the Commission’s other proposals, took another three months. The measures it authorized—including the creation of a Director of National Intelligence—lay fallow until a second commission, investigating intelligence prior to the war in Iraq, issued its own report four months later. The new Director was sworn in on April 21, 2005. Total response time, charitably defined: about 44 months, and implementation continues today.
Obviously, planning an attack and adjusting defenses to prevent a subsequent attack are not comparable tasks. Still, it is hard to avoid concluding that organizations like al-Qaeda are inherently nimbler than governments, especially large and highly bureaucratized governments like ours. As things stand now, terrorists can size up a situation, make decisions and act faster than we can. In military terms, they are “inside our decision cycle.”
Recall July 7, 2005, for example, when terrorists bombed three London Underground trains and a double-decker city bus, killing 52 commuters. The four bombs exploded within a minute of each other, an operationally and technically challenging feat that is a hallmark of al-Qaeda attacks. A “martyrdom video” proclaiming allegiance to al-Qaeda and taped months earlier by one of the bombers, Muhammad Sidique Khan, soon surfaced on al-Jazeera. Khan apparently made the video during a visit to Pakistan, and investigators concluded that an earlier trip to Pakistan in July 2003 also had something to do with the attack. If so, then the planning of the London attack required two years, possibly less.
Again, it may seem unfair to compare a government bureaucracy, American or British, with a network of loosely organized, small terrorist cells. But unfairness is the point: Terrorists will always make the conflict between us as “unfair” as possible, avoiding our strengths and exploiting our vulnerabilities however they can. So will insurgency leaders and rogue dictators, who also happen to be surreptitious WMD proliferators; narco-traffickers and money launderers, who aid terrorists either wittingly or inadvertently. The U.S. government and similarly arrayed allies will simply lose battle after battle if our adversaries absorb information, make decisions, change tactics and act faster than we can.
Reading the 9/11 Commission Report one cannot help but be struck by how often simple delay and chronic slowness led to disaster on September 11. President Clinton told the Commission that he had asked for military options to get rid of bin Laden in late 1999. But General Hugh Shelton, Chairman of the Joint Chiefs of Staff, was reluctant to provide them. Secretary of Defense William Cohen thought the President was speaking only hypothetically. The one person who could have given a direct order to cut through the resistance and ambiguity, President Clinton himself, did not do so. He thought that raising his temper wouldn’t accomplish anything, so he allowed himself to be slow-rolled, and the issue went essentially unaddressed.
The problem wasn’t just at the top, however. Down below in the bureaucracy, things were just as bad—case in point, the Predator. The now-famous robotic aircraft was originally built for battlefield reconnaissance and was later modified to carry missiles. The U.S. Air Force had flown Predators in the Balkans since 1996, but Afghanistan was trickier. The aircraft had a limited range and thus needed a remote base and data uplinks to get the information back to Washington. It took until July 2000 to work out these details, and two more months to deploy the Predator over Afghanistan.
Predator operators thought they spotted bin Laden in September 2000, but U.S. officials disagreed over rules of engagement. National Security Advisor Samuel Berger wanted greater confidence in bin Laden’s location before approving a strike, and he worried about civilian casualties. At the same time, Air Force leaders were reluctant to carry out what looked to them, not unreasonably, like a covert operation, and the CIA was reluctant to undertake a direct combat operation—or to violate the Executive Order prohibiting assassination.
These disagreements dragged into 2001 as the Bush Administration took office. Then President Bush put everything on hold while National Security Advisor Condoleezza Rice directed a comprehensive plan to eliminate al-Qaeda. George Tenet, the Director of Central Intelligence, deferred the legal issues over whether the CIA could take part in an attack until the Administration had prepared its new strategy. So it went, until the clock ran out and the terrorists killed nearly 3,000 people.
Or take the inability of the Immigration and Naturalization Service (INS), as it existed on September 11, 2001, to track the whereabouts of known terror suspects and to report relevant information about their attempts to enter the country to other Federal agencies. The INS failed to meet its homeland security responsibilities partly because Congress systematically underfunded it. But even worse, the INS had failed to disentangle its different functions: keeping some people out of the country while letting others in. Meanwhile, everyone—the White House, Congress, the bureaucracy—failed to agree on a solution that both dealt with illegal immigration while also allowing entry to laborers essential to the American economy. The security problem flowing from this failure is obvious: As long as underfunded bureaucrats are unable to regulate the enormous flow of illegal immigrants seeking work, they will never be able to detect and track the few truly dangerous people trying to enter the country.
Of course, the story of the run-up to 9/11 is an oft-told one. Yet almost everyone seems to miss the core problem from which all others followed: There was always time for another meeting, another study, another round of coordination. Virtually no one was worrying about the clock—about whether time itself mattered. It’s not that every concern raised didn’t have some legitimate rationale (at least within the legal-bureaucratic culture that characterizes the U.S. government). It’s the fact that, while we were working out legal issues, al-Qaeda was developing and executing its plan.
This same problem surfaced again a year later. Just about everyone agrees now that the United States was unprepared for the insurgency in Iraq, but most overlook that someone else was also unprepared: the insurgents. U.S. analysts who interviewed captured Iraqi officials and military officers for the Defense Department have concluded that Iraqi leaders had not prepared a “stay-behind” or “rope-a-dope” strategy. They had never planned to forfeit the conventional war in order to win a guerrilla war later on. Iraqi military leaders believed they would lose the war and just wanted to get it over with quickly. Saddam Hussein’s security services and core Ba‘ath Party operatives kept the lid on the various sects, tribes and ethnic groups so that they could not plan a guerrilla war either. The result was that no one was prepared for an insurgency. The United States, its coalition partners, Ba‘athis who had escaped capture, tribal leaders, religious authorities, foreign fighters—everyone was starting from scratch. So when Saddam’s statue came down in Firdos Square on April 9, 2003, the question that mattered most was who could organize and execute faster, the would-be insurgents or the U.S. government?
Alas, we were left in the starting blocks. The insurgents organized much faster than U.S. officials could recognize and respond. We were playing catch-up from the beginning, which is another way of saying we were losing.
Things would perhaps not be so bad if the war on al-Qaeda and the war in Iraq were exceptional. In truth, the problem is pervasive and getting worse. “Organizational agility” sounds abstract, but it really boils down to specific questions: How long does it take to deliver a critical weapon or information system? How fast can an agency bring new people on board? How fast can it change its mix of people if it needs to? In short, how fast can government agencies act—and is this fast enough to stay ahead of the competition?
The U.S. government is not always woefully slow. The response to the December 2004 Southeast Asian tsunami, for example, was admirably quick and reasonably effective under the circumstances. So was the relief mission that the United States effectively led following the massive earthquake that rocked northern Pakistan in October 2005. However, these few exceptions aside, the U.S. government has become an increasingly ponderous beast, unable to act quickly or even to understand how its various parts fit together to act at all.
Once, When We Were Fast
It was not always so. After the surprise attack at Pearl Harbor, one of the most heavily damaged ships was the battleship USS West Virginia. Most of its port side had been blown away. The ship sank rapidly, but on an even keel on the bottom of the harbor. The Navy needed every 16-inch gun it could muster, so Navy leaders decided to repair the ship. It was not easy, but the USS West Virginia steamed into Puget Sound in April 1943 to be refitted and modernized. It rejoined the fleet in June 1944, thirty months after it was sunk, took part in several operations and was present for the surrender ceremonies in Tokyo Bay in September 1945. By comparison, after al-Qaeda agents in Yemen damaged the USS Cole far less severely with a single improvised bomb in October 2000, it took 16 months to retrieve the still-floating destroyer and complete repairs in Pascagoula, Mississippi. The ship did not then leave its home port in Norfolk, Virginia for its first deployment until November 2003—37 months later.
World War II offers many examples like the recovery of the West Virginia in which organizations worked with remarkable alacrity. Take the effort to build the first atomic bomb. Albert Einstein wrote to Franklin D. Roosevelt on August 2, 1939, alerting him to the possibilities of nuclear weapons. He met with FDR about a month later, which led Roosevelt to establish the Uranium Committee to research military applications of nuclear fission. Vannevar Bush, Roosevelt’s science adviser, persuaded the President to accelerate the project in October 1941, as war with Germany and Japan seemed likely. On September 14, 1942, Brigadier General Leslie Groves was appointed director of the new Manhattan Project, marking the formal start of the project to build the atomic bomb. The Trinity test, the world’s first nuclear explosion, took place on July 16, 1945, and Hiroshima was bombed on August 6, less than a month later. The entire effort, costing $21 billion in today’s dollars, developed three different means of producing fissile material, two bomb designs and three devices.
Or consider the Office of Strategic Services, the predecessor of today’s CIA. President Roosevelt appointed William Donovan as his “Coordinator of Information” in July 1941, and the OSS was itself established in June 1942. Harry Truman disbanded it in September 1945. In other words, the entire history of the OSS—what many consider the Golden Age of American intelligence—spanned just 37 months. In that short time it recruited, trained and deployed a workforce of about 13,000 people. William Casey, directing OSS espionage in Europe, stood up his entire network in about 18 months. By comparison, after 9/11 Tenet said on several occasions that it would require five years to rebuild the CIA’s clandestine service.
Or recall the war in the Pacific. The Battle of the Coral Sea was fought in May 1942, the Battle of Midway a month later. Within six months of Pearl Harbor, the U.S. Navy had destroyed five Japanese carriers, along with most of Japan’s naval aircraft and aviators. It has taken us longer just to get organized for the so-called War on Terror (to the extent that we are organized for it) than it did to fight and win World War II.
Delivering the Product
Everything else today is moving faster, thanks to jet airliners, interstate highways, the computer and the Internet. But government, including the parts responsible for national security, is moving slower, and it’s getting worse.
Everyone knows, for example, that weapons have been getting more expensive per unit, but few realize that it now also takes much longer to get a weapon into the hands of the warfighter. In the early 1940s, it took 25 months to get a new fighter like the P-47 Thunderbolt into action from the time the government signed a contract for a prototype. In the late 1940s, this delay had grown to about 43 months for an early jet fighter like the F-86 Sabre. By the 1960s, the F-4 Phantom required 66 months, and its 1970s replacement, the F-15 Eagle, 82 months. The latest fighter to enter service, the F-22 Raptor, traces its development to a prototype built under a contract signed in October 1986. The prototype first flew in September 1990, and the production model entered service in December 2005—a total of 230 months, or about 19 years. Put another way, that comes to slightly longer than the typical career of an officer in the U.S. Air Force. (The new F-35 Lightning II, which will replace the F-16, is slated to require “just” 15 years from signing the contract for the prototype to when it enters service. We’ll see.)
One might think the problem with jet aircraft is a result of the growing technical complexity of modern fighter aircraft, but that argument does not hold up. No rule says that the more complex a technology is, the longer it takes to deliver. Government aircraft of all kinds take longer to develop, and longer than their commercial counterparts. Compare a military transport, like the C-17 Globemaster III with the new Boeing 787 Dreamliner. The C-17 required 12 years to enter service, while the 787—more complex than the C-17 in many respects—will take just four. And the 787, for example, will require a little less time to develop than its predecessor from the early 1990s, the 777.1
The problem holds for most weapons other than airplanes, too—ships, tanks, electronic systems and so on. Threats are changing much faster than we can develop the means to counter them. This is why some officials occasionally say we have to anticipate requirements further into the future. But that’s simply unrealistic. When you try to forecast two decades ahead because your weapon takes twenty years to develop, it isn’t analysis; it’s fortune telling.
The ever-slowing pace of government appears in other ways, as well. Simply getting a presidential administration into place is a stellar example. According a 2005 National Academy of Sciences study, every Administration since Kennedy’s has taken longer than its predecessor to fill the top 500 jobs in government. In the 1960s it took just under three months; today it is three times as long. A new administration isn’t up and running until almost a year after the election that put it in office. How can a team possibly win the Big Game if half the players don’t show up until the end of the first quarter?
That is more or less what happened in 2001 as al-Qaeda was preparing 9/11. The Bush Administration’s Cabinet Secretaries were confirmed and ready to go when the new President was sworn in on January 20, 2001, but that was about it. The Administration didn’t nominate Paul Wolfowitz to be Deputy Secretary of Defense until February 5, and he had to wait until March 2 to be confirmed and sworn in. Wolfowitz’s wait was comparatively short; most positions took longer to fill. Richard Armitage, nominated for Deputy Secretary of State, waited until March 23, 2001. Six months passed before the top Defense Department leadership was in place. Douglas Feith, the Under Secretary of Defense for Policy—as in “policy for combating terrorists”—was last to be sworn in, in July 2001.
What is so depressing about the National Academy of Science study is that the problem just keeps getting worse. If top officials have to wait two or three months at the beginning of an administration, candidates for positions at the assistant secretary level in the middle of a term can often wait six months or more. Further down the food chain, bringing on new staff is paced largely by how long it takes to obtain a security clearance. For civil servants, this can take almost a year; for government contractors, the average is about 450 days.
What explains this bureaucratic torpor? In part, government is slowing down because more people insist on getting involved. Ever more congressional committees, lobbyists and oversight organizations vie to get their prerogatives enacted in a law, regulation or procedure. As the participants multiply, workloads expand and everything slows down.
At another level, it’s because there is more obligatory paperwork to handle—financial disclosure in the case of officials, cost justification in the case of contracts, quality assurance documentation in the case of hardware. At yet another level, it’s because all organizations have standard procedures that never seem to get shorter or more flexible; quite the reverse. New procedures are almost always cumulative, accreting in ever thicker layers of bureaucratic hoariness. Indeed, we may be seeing a classic case of “organizational aging”, a phenomenon perhaps first defined by economist Anthony Downs back in 1967.
In his classic book, Inside Bureaucracy, Downs observed that when organizations are first established, they have few rules, written or unwritten, and because new organizations tend to be small, they have a flat, short chain of command with little hierarchy. As time goes by, alas, organizations add personnel. Since managers can oversee only a limited number of people, they develop a reporting hierarchy, which adds to the time and difficulty of making a decision. More members are in a position to say “no”, and the joint probability of “yes” diminishes. This translates into the well-known bureaucratic adage, “Where there’s a will, there’s a won’t.” The fact that people expect promotion to positions with greater responsibility (and pay) also encourages the establishment of more management slots with the selective power to say “no”, or just to kibbitz. Either way, the process takes more time.
Also, as organizations mature, they develop dogma—sometimes written, sometimes simply part of the organization’s culture. This, of course, is exactly what bureaucracies are supposed to do: simplify decisions and improve efficiency by adopting rules. This is fine, until the rules become cumbersome or no longer appropriate to the situation—which is exactly what is happening today.
But the most insidious problem of all is that as organizations mature their character changes. New organizations with few rules offer lots of challenge and risk, so they tend to attract risk-takers who want to make their own rules. Mature organizations with well-defined rules and missions, on the other hand, attract the “Organization Man”—the sort who wants to plug himself in and carry out tasks as set forth in an official, approved job description.
This is why it is somewhere between ironic and pointless to hear critics complain that this or that long-established government organization needs to become less risk-averse and more innovative. Inevitably, they are speaking to people who, by self-selection, are where they are precisely because they are risk-averse. They like the way things are; they would not otherwise have joined the organization and stayed with it. Organization Men are no less patriotic, dedicated or capable than risk-takers; they’re just temperamentally opposite.
If we are serious about gaining agility, we will clearly have to break some china. Improving agility means more than just rearranging boxes on an organization chart, though that is mostly what we have tried to do. There have been countless studies on how to streamline contracting, speed up background investigations, shorten the process of nominating and confirming appointees, and so on. None of these recommendations will ever amount to anything unless we find a way to produce a new mix of people who can develop new ways of doing things, and attract the kinds of recruits who thrive on doing just that.
It’s easy to get lost in the day-to-day specifics of why it takes so long to get anything done in the American national security community today. It is far more important to recognize that the underlying theme connecting all the sources of our sloth is that we are trying to balance risk with speed, and there is rarely a champion for speed. The risks that concern people take many forms—that some group will be underrepresented in a decision, that a design or work task will be flawed, that a secret will be compromised, that someone will cheat the government, that an official will have a conflict of interest. Whatever the specifics, we lose agility every time we manage risk by adding a step to reduce the probability of something bad happening. Rarely does anyone with responsibility, opportunity or power say that we should be accept more risks so that we can act faster.
It is easy to argue for doing something to avoid some hypothetical bad thing happening. It is much harder to argue that one can take so many precautions against some kinds of risk that other kinds of risk actually increase due to an organization’s diminished capacity to act in a timely fashion. The real question is, or ought to be, how much speed do we want to sacrifice in order to reduce certain kinds of risk? There is no single, objective answer to such a question, but without advocates and mechanisms for greater speed, we will be protected against risk so well that arguably our most dangerous adversaries will beat us every time.
Examples of Speed and Success
Lest we be too pessimistic, there are cases—including a few fairly recent ones—in which government organizations moved out smartly on national security missions. These cases show us what we need to do if we want organizations to move fast. Consider, for example:
• The U-2 aircraft: In the 1950s, the United States needed a higher-flying airplane to take pictures of Soviet military facilities. The CIA gave Lockheed authority to proceed in December 1954; the aircraft flew its first reconnaissance mission over the Soviet Union in July 1956. Total time required: 18 months.
• The Explorer I satellite: Desperate to match the Soviet Sputnik I launched in October 1957, the Defense Department authorized the Army Ballistic Missile Agency to prepare a satellite for launch on November 8, 1957. Werner von Braun’s team launched it three months later, on January 31, 1958.
• The GBU-28: At the start of Operation Desert Storm in 1991, the Air Force discovered it did not have a bomb that could penetrate Iraq’s deepest underground shelters. To pack enough kinetic energy, the bomb had to be long, streamlined and heavy. The Air Force Research Laboratory took surplus gun barrels from eight-inch howitzers as a casing, filled them with explosive, bolted an existing laser guidance system to the front end, and—after assigning it an official Air Force designation— delivered a bomb in 27 days.
• JAWBREAKER: President Bush asked for options to respond to the September 11 attacks. The CIA presented its plan two days later to use Northern Alliance forces as a surrogate army. CIA units, called “JAWBREAKER”, arrived in weeks, and Kabul was taken on November 14, 2001.
These programs are all related to national security, but they are as different from one another as one can imagine. One is an aircraft development program, one a space research mission, one a weapon system and one a covert paramilitary operation. The Army, Air Force and CIA are all represented. Two were in wartime, two in peacetime. Yet they share some common features, the most important of which seems to be that someone was willing to bend rules and take responsibility for getting things done. This is a logical—even a necessary—condition for speed.
Every organization has a “natural” maximum speed defined by its standard procedures, which are designed to reduce risk. Some are formal, others implicit. Together they establish the organization’s operations—who has to confer with whom, who can approve, what materials have to be prepared and so on. Organizations usually operate well below this optimum speed, but in principle one could analyze any organization and then assess whether it can act faster than its competitors. It is hard to measure maximum speed precisely, but it is easy to identify most of the “hard points” that constitute it, like the one official or office lying in the critical path of workflow. Conversely, when government organizations have moved faster than their normal maximum speeds, it’s almost always because someone either bent the rules or managed to evade them. Consider the cases cited above.
In developing the U-2, the CIA avoided the constraining pace of the annual Federal budget cycle by using its special authority to spend money without a specific appropriation—the first time the CIA had used that authority to develop a major system like an aircraft. The CIA also wasn’t bound to Defense Department regulations, so rather than use the arduous military acquisition and contractor selection process, the CIA simply chose Lockheed.
Lockheed’s famous “Skunk Works”, in turn, shortened or eliminated many steps a military contractor would usually take. For example, by having all its people working in one location, an engineer could ask metal workers to adjust the design on the spot with a conversation rather than a meeting, and follow up with documentation later. This would violate normal Defense Department acquisition regulations.
The Army also broke rules in building the Explorer 1 satellite—specifically, the rule saying that the Army wasn’t supposed to build satellites. The Defense Department and White House had given the Navy that mission. Major General John Medaris, the Army Ballistic Missile Agency director, “went out on a limb”, as he put it, and set aside hardware that later gave the Army the ability to get off a quick shot after the Soviets launched Sputnik.
The U-2 and Explorer I also had something in common: They “stole” a lot of technology from other programs, using them in ways that no one had originally intended but that sped up the process. The U-2’s design was in many ways just like that of the F-104 Starfighter that Lockheed had designed earlier for the Air Force, but with longer wings and a lot of weight cut out. The rocket that launched Explorer I was based on an Army Redstone ballistic missile, which, in turn, was an updated V-2 that the Army’s German engineers had developed during World War II.
In the case of the GBU-28, the Air Force Development Test Center team compressed a development program that would ordinarily have taken two years into less than two weeks by taking engineering shortcuts and a more liberal approach to safety. For example, it tested the aerodynamics and ballistics of the weapon with a single drop, rather than the usual thirty.
Note that it required an individual with the authority and inclination to make the decision on how to interpret a contract or a standard. If a person could not legally give approval, the organization would not have followed his direction. If a person had not been willing to use his authority (and, in the process, accept responsibility), nothing would have happened, either—which brings us to JAWBREAKER.
CIA officers like Gary Schroen, who first went into Afghanistan to prepare the operation immediately after 9/11, had largely acted on their own initiative in the 1990s when they kept up personal contacts with Northern Alliance figures like Ahmed Shah Masoud. After the Soviets were defeated in the U.S.-supported guerrilla war from 1980 to 1989, the CIA had turned its interest elsewhere. Schroen’s contacts and experience in the region greased the re-establishment of the relationship when the United States decided to retaliate against al-Qaeda and the Taliban.
After the fighting started, the CIA was fortunate to have officers on hand with admitted inclinations for focusing more on results than procedures. Gary Berntsen, who took command of JAWBREAKER as the fighting began, once described himself as a “bad kid” from Long Island who graduated second from the bottom in his high school. Once in the CIA, he bragged about his “grab-’em-by-the-collar” approach.
As Admiral Ernest King supposedly said about wartime, “When they get in trouble, they send for the sons of bitches.” If you don’t have SOBs on staff and a way to get them to the front line, organizations will plod along at their routine pace. True, if everyone broke the rules all the time, there would be no rules. But one of the keys to a fast organization that can beat its opponent to the punch is almost always a willingness to break the rules. This is nothing new. It was said often in the 19th century that Paris sent officials into the French countryside not to enforce rules, but to decide judiciously when and how to ignore them.
How To Get Faster
If we want more speed and agility, some lessons are clear. We must: Make sure U.S. national security organizations have a legal mechanism for bending or breaking existing rules; make sure they have the means for having such rule-benders at hand; make sure these rule-benders exercise influence; and make sure they don’t get out of control. (Even unofficially designated rule-benders need some clear lines of accountability.)
Basically, we need to allow responsible senior officials to put the government into overdrive when it’s really important. With the possible exception of the operating forces of the military and their counterparts in the intelligence community, even top officials lack this ability today. This encourages other kinds of risks: workarounds. Cabinet secretaries who need to get decisions fast and begin operations expeditiously know that they cannot entrust such matters to the standing bureaucracy. But workarounds and shortcuts spite the institutional memory of an organization and court disaster from ignorance. Iran-Contra is a good example of a workaround gone wrong. The only way to avoid such dangers is to make the responsible bureaucracies faster only when they really need to be fast.
There is always a tension between orthodoxy and innovation, and between direct command and checks and balances. There is no sure-fire way to ensure the best mix. But we don’t seem to be close now, or even trying to get closer. Ultimately, our willingness to balance different sorts of risk must be a political decision, in which voters can turn incumbents out and try something else if they are dissatisfied. But if we don’t at least have the foundation for rule-bending, they will never get that choice.
What then, should we do? First, to build agility into the key parts of the U.S. government, Congress will clearly have to cooperate. That’s the system; that’s the Constitution. It is therefore folly for any administration to try to steamroll the legislators—as the then-popular Bush Administration did from about 2002 to 2004, such as when it shunted aside congressional concerns that the Iraq insurgency was gaining steam rather than entering its “last throes”, or that U.S. forces did not have the resources to deal with the worsening situation. Accepting these concerns and criticisms quickly would have both improved the situation and solidified support for the effort by getting Congress’ “buy in” on the record.
Second, we should consider establishing a small number of powerful “bottleneck breakers” in the Executive Office of the President. Senior experienced officials could be designated by the White House, formally or informally. The important thing is that officials down the line know that these bottleneck breakers are acting at the behest of the president to make sure his policies are carried out. Unlike the too-familiar “czars” that have been given responsibility for drug enforcement, energy conservation and, most recently, the war in Iraq, these officials would know how, and be given the authority, to work quickly and quietly with the Office of Management and Budget. It would take only a few examples of a sequestered budget line, a dismissed appointee or a transferred senior executive to give these bottleneck-breaker envoys the implicit power they require. The very existence of such EOP envoys, and the only occasional demonstration of their authority, would work wonders with hidebound, risk-averse bureaucrats.
Other measures that would counter the natural tendency of bureaucracies to slow down come readily to mind:
• Requiring senior civil service executives to periodically do a tour in a different Executive Branch department. This would make them more familiar with conditions in other departments, so they could anticipate what might slow down an action. It would also build social networks that could help clear these impediments.
• Create an “up or out” system of promotion for senior executives resembling the approach used in the military to create a dynamic that keeps the bureaucracy from getting too settled.
• Adopt a mandatory, congressionally approved, periodic de-layering of bureaucracies.
• Increase the number of Schedule C appointments to give new administrations a better ability to rattle cages. We need not repeal the Pendleton Act completely, but the trend in most sectors of the economy is toward “at will” employment. As an employee rises higher in the organization, it should be easier to move or remove him or her.
One could think of other measures in the same vein, and some have. The point, however, is that if we do not do something to increase the speed of government, we will be sure to fall behind future events, get beaten to the punch, and lose ground to our most ruthless competitors. Given the stakes in today’s world, that is a loss we cannot afford.
Also consider today’s automobiles, which are much more complex than earlier models. Like jet fighters, cars today go faster and handle better. They can also locate their current position and tell you how to reach your destination—all while meeting ever-tougher safety and emissions standards. Yet the time required to develop a car and get it into the showroom is getting shorter all the time. Toyota is best at about two years, and it is trying to cut this time to 12 months. Ford and GM are trying to keep up, but still take one to two years longer than Toyota—one reason they have been taking a beating in the market.