You raise some valid concerns about the potential for misuse by the proposed new agency. Government power has been misused may times in the past, with the FBI and CIA being but two examples. There are quite a few examples where other agencies have arguably abused their power, such as the IRS, EPA, and OSHA.

Still, I believe there needs to be some serious thought to consolidating government agencies by mission. Part of the problem with counter-terrorism is that literally dozens of agencies all want to play, so there is a lot of duplicated effort while at the same time other things get overlooked. It's kind of like the so-called "war on drugs" where you have so many competing agencies all wanting to justify their budgets by playing their part. Do we really need the DEA, FBI, (probably) the CIA, and who knows who else to fight the "war on drugs"? (setting aside the different question of whether we should be fighting that particular "war" at all?). You can find similar duplication in just about every section of the government, such as literally scores of adult literacy programs within the Department of Education.

Duplication of effort is wasteful of taxpayer money and ultimately leads to breakdowns like we had on Sept 11th. I read an article several months ago that discussed the individual pieces of information that different agencies had before Sept 11th. There were many "dots" known before the attacks but they weren't connected because the different agencies weren't sharing what they knew with one another. Preventing that sort of thing from happening again is the central focus of Bush's consolidation proposal. We'll just have to make sure that doesn't lead to abuse of power.

As an example of duplication of effort, within the military, do we really need 5 air forces? Each service currently has its own air arm. Is that really necessary, wise, or efficient? How well do the separate air arms work together to protect the country? (Actually, they work together a lot better than it used to but there's still much room for improvement).

There's another systemic problem with government agencies that I'm currently encountering in my job. I'm working on a team charged to develop a whole new way of knowing what is going on in space, with the goal being true space situational awareness. Currently, there is a host of players involved including USSPACECOM, AFSPACECOM, ARSPACECOM (Army), Navy Space Command, NRO, NSA, CIA, DIA, NIMA, NASA, and probably some others who I'm forgetting at the moment. Each of these agencies and commands has their own way of doing things, typically using non-integrated ("stovepipe") systems that don't communicate well together. For example, people in Cheyenne Mountain typically have 4 computer monitors on their consoles, each connected to a different "stovepipe" system. It's very difficult to determine how data on one system relates to the other systems due to this failure in architecture.

In addition, I believe there is excessive "compartmentalization" of classified information that hinders informed decision-making by those in charge. Beyond TS/SCI, there are a host of security compartments that are used to limit access to information. This makes integration extremely difficult if not impossible. Both within the military and other agencies, I believe there needs to be a reevaluation of the compartments, perhaps with some of the walls broken down under a more consolidated system. However, that goes against everything the government holds near and dear regarding security.

I'm pretty sure this problem extends far beyond the space arena. When different agencies have duplicated missions, incompatible computer systems, and compartmentalized critical information, the result is a far less than optimal solution, IMO.