Critics of the response to Hurricane Katrina in New Orleans tend to focus on the need to formulate and implement better plans. I suspect that a more sober assessment might instead identify poor improvisation as the main problem. That is, if officials close to the scene had assumed more responsibility and been granted more leeway to focus on results, rather than waiting for instructions or assistance, then some of them would have taken more initiative and averted some of the worst problems.

I think that people have a tendency to put too much faith in centralized planning, and they do not have sufficient regard for decentralized improvisation. The more ambiguity that exists in a situation--because of its novelty, uncertainty, and the absence of critical information--the more that it favors improvisation over planning.

In my previous essay, I pointed out that large organizations necessarily must lean on planning, while small organizations necessarily must rely on improvisation. I also pointed out that there are many problems which require both planning and improvisation, and that such challenges make it quite difficult to come up with an optimal organizational design.

War

As someone who reads military history but lacks combat experience, it seems to me that war is a dramatic illustration of the need both for planning and for improvisation. In World War II, D-Day or the Manhattan Project stand out as examples of extensive planning. However, other famous exploits, such as the successful evacuation of British troops at Dunkirk or breaking the German Enigma code, owed more to improvisation.

In his book Bureaucracy, James Q. Wilson wrote,

"The key difference between the German army in 1940 and its French opponents was not in grand strategy, but in tactics and organizational arrangements well-suited to implementing those tactics...

An army that could probe enemy defenses, infiltrate weak points, and rapidly exploit breakthroughs with deep encircling moves could not be an army that was centrally directed or dependent on detailed plans worked out in advance. It had to...permit independent action by its smallest units -- squads, platoons, and companies."

This analysis boggles the mind, because it suggests that Hitler's Wermacht was a model of decentralization and improvisation.

Robert Kaplan, in his new book Imperial Grunts, paints a similar picture of the front-line American troops confronting terrorists in the world's lawless regions. It is true that the United States has the ability to project awesome conventional military power using planning and logistics, which enabled our forces to drive Saddam Hussein from Baghdad. However, terrorists in Columbia, the Philippines, or post-invasion Iraq will not be overwhelmed by planning and logistics. Instead, Kaplan is impressed by the tactical improvisation of smaller military units. Our special forces operate best when given one-sentence missions with flexible ROE's (rules of engagement).

(Kaplan's field-level reporting has impressed me since I first read Eastward to Tartary, his account of travels through the broken pieces of the former Soviet empire. In contrast to Jeffrey Sachs, who comes across as someone who fights his righteous battles against world poverty from the comfort of 5-star hotels, Kaplan strikes me as someone who has been close enough to the day-to-day reality of failed states to provide truly useful insight and guidance on the issues of underdevelopment. If Sachs would like to understand why he failed the Russians as an economic adviser, he could learn from Kaplan.)

Covering Iraq in 2004, Kaplan writes,

"Nothing angered me more than to enter a vast chow hall at Camp Victory, the headquarters of the military coalition at Baghdad airport, and see it teeming with troops choosing different kinds of fine cakes for dessert, and then to travel in the countryside and see barely a U.S. presence at all. The U.S. military was everywhere hardened by a top-heavy bureaucracy, with too many layers of staff that needed pampering...[it] had set up structures it was historically comfortable with, not those particularly suited for the challenge at hand."

The reality of improvisation in wartime may reduce the significance of central planners, but it does not eliminate the need for professionalism. In order to improvise effectively, soldiers must study, drill, and train. Unskilled improvisation is simply another term for ineptitude.

Bureaucracies are famous for valuing organizational comfort over results. I do not know whether Able Danger really had drawn a bead on some of the September 11 terrorists prior to the attacks. However, it seems to me highly plausible that before September 11 only a small, independent operation could have developed this sort of intelligence, and that the rest of the bureaucracy would have discounted the views of such a "rogue" agency.

Disaster Mitigation

Disasters produce the unexpected. Communication channels disappear or become overloaded. Unanticipated problems and opportunities present themselves. On-the-spot improvisation is called for.

Disaster mitigation, which is what was expected following Hurricane Katrina, is likely to require improvisation. That means giving small, well-trained units a mandate to accomplish their missions with a minimum of centralized hurdles and constraints to overcome. Contrary to all bureaucratic instincts, they must operate with an aversion to errors of omission that is almost as great as the fear of errors of commission. They must resist the standard tendency of organizations to make it easier to fail conventionally than to succeed unconventionally.

My reading of FEMA, the Federal agency currently blamed for the awful scenes in New Orleans, is that up until very recently, it acted as a sort of insurance adjuster. FEMA officials would arrive days or weeks after a disaster, assess damages, and help process funds for compensating victims. This may or may not have been the mission in theory, but it appears to be the way that FEMA operated in practice.

Now, the public appears to expect FEMA to operate as a sort of domestic Green Beret outfit, able to parachute in to a crisis area and solve humanitarian and engineering problems in real time. In theory, such an organization would be highly valuable. In practice, it strikes me as implausible that FEMA would turn out to be that organization.

When the mission is described as real-time disaster mitigation, the value of planning is likely to be limited. Instead, training, drill, and improvisation are likely to be of more practical use.

Counterproductive Planning

When something goes wrong, there is a natural desire to blame a lack of planning. In fact, with hindsight, it is always possible to come up with a plan that would have worked better. I would refer to this as the planning illusion. This illusion causes a number of problems.

First, the planning illusion leads to the syndrome known as "planning for the last war." Organizations develop a set of operating strategies that are based on theories that are outdated, or just completely misguided.

Second, faith in planning causes organizations to become overly centralized. Information from peripheral sources is ignored. Flexibility for field-level decisionmaking is denied.

Finally, faith in planning leads people to believe that government has a solution for every problem. In many cases, better approaches emerge from decentralized improvisations.

After Hurricane Katrina, the United States faced a shortage of gasoline, caused by the disruption of refining capability. My worst fears were that panic would break out, and Americans would rush to fill their tanks, making it difficult for people with critical needs to find gas. There was a need to conserve gasoline until production could be restored. It turns out that the markets produced exactly such a response. The rise in gas prices, which in my view was relatively small considering the sudden severity of the crisis, was sufficient to induce people to live more fuel-efficiently for a while. The law of demand worked much better than any centralized rationing scheme.

I know that in the aftermath of Katrina, many pundits are saying that we need more planning and more centralization. I only wish that people would keep an open mind an examine the evidence before arriving at such a conclusion.

My sense is that we live in an age where ambiguity is on the rise, because technology is changing rapidly and globalization has increased the speed at which opportunities and threats materialize. This suggests a relatively greater need for improvisation and adaptation, with somewhat less value in bureaucratic planning. Unfortunately, the planning illusion seems to cause many people to long for a government approach that is more centralized rather than less.

To see more of the extensive coverage of Hurricane Katrina from TCS, click here.