In his 1989 book Bureaucracy, James Q. Wilson divided public agencies according to two characteristics. The first characteristic is whether their activities can be observed. The second is whether their results can be observed.
Consider a public agency responsible for responding to various types of disasters. When FEMA goes to work, it is easy for citizens and journalists (though maybe not the White House) to observe its various activities. FEMA's actual work is done out in the open, and you don't have to be an expert on disaster management to know what is going on -- to observe whether supplies are being delivered or people are being rescued. Moreover, citizens and journalists (and even the White House) can figure out what results FEMA produced: Did people live or die?
Thus, Wilson would call FEMA a "production agency." And, production agencies are relatively easy to manage. You don't have to be a specialist in potholes to manage the pothole-filling unit of a municipal public works department. You can see who is filling what potholes, and you can see whether they were filled well or not. No particular expertise is required.
The same is true for FEMA. You don't have to be an emergency management specialist to manage FEMA (though, as always, a little experience might help in your first round of planning). After all, you will quickly face a variety of emergencies -- situations in which you can observe both activities and results. Consequently, even if your agency's response to the first few emergencies contains some significant mistakes, you can quickly learn from them, and adjust your future activities.
The Army Corps of Engineers, however, is different. It's not difficult for the commander to observe the results of his organization's work. When the levees break, the top manager knows that the organization has done a poor job. But, the commander can't easily observe the activities of his organization that produced this disaster. Thus, the Corps is what Wilson calls a "craft organization."
Yes, the commander of the Army Corps of Engineers started at the bottom of the organization as a second lieutenant. Once upon a time, he possessed a lot detailed expertise about some of the engineering challenges that the Corps faces. He might even have been an authority on levees. But, as he rose though the ranks, he became less of a engineering specialist and more of a management generalist.
Sure, he could go out and look at a levee being built. And, he could ask intelligent questions. But, he would no longer be in a position to evaluate the answers. If the engineer in charge of the levee project tells him that the design is state-of-the-art, that the materials are of the highest quality, and that the workers have excellent qualifications and are doing a superior job, the commander lacks the expertise to challenge these judgments.
Consequently, fixing one of Katrina's management disasters is easier than fixing the other. Put a competent manager in charge of FEMA and he or she will -- by observing both activities and results -- be able to make significant improvements very quickly. The same, however, is not true for the Army Corps of Engineers.
The Corps, like other craft agencies, needs people who can observe the quality of organization's activities and who can report their findings to top management. If the leadership of a craft agency can hire a few outsiders -- individuals or organizations who possess the expertise to evaluate the activities of the organizations, and who can provide an independent judgment about the quality of these activities -- they have a chance to prevent the next disaster. Otherwise, they are at the mercy of those middle managers who will say, "Don't worry boss. Everything is under control. My people are doing a great job."
Max Weber, the sociologist who made so many insightful observations about bureaucracy, once noted, "The absolute monarch is powerless opposite the superior knowledge of the bureaucratic expert." To emphasize this point -- to contrast the power of the low-level expert with the power of the top-level "political master" -- Weber referred to those in the nominal position of authority as "dilettantes."
"The question is always who controls the existing bureaucratic machinery," wrote Weber. "And such control is possible only in a very limited degree to persons who are not technical specialists."
In craft agencies, top management -- regardless of whether you describe them as dilettantes or generalists -- need to hire their own experts.