Reinventing the Wheel--A Defense

I often find myself ambivalent (reminds me of the Truman joke about wanting a one-armed economist--the economists he had always said: "on the one hand..., on the other hand...).

Having attacked the "creation fallacy" yesterday, today I should defend the idea of reinventing the wheel. The defense is simple: people learn by doing, often better than by any other way. Every baby born has to reinvent the way to walk. So too there are times when it's better to reinvent the wheel just to educate the policy makers. I'm using "educate" here both in the sense of learning facts but also investing or committing to the venture.

Having put this market out, I'll try to keep alert for situations in which people are accused of reinventing the wheel to see if the above is correct.

The Creation Fallacy

No, I'm not talking creationism, but a fallacy that's related to NIH--"not invented here". The fallacy clouds the minds of politicians new to office, though like many fallacies there's some truth in it.

The fallacy takes the form of:

• politicians have objective X. They may legislate a program to achieve X, or they may just want the executive branch to achieve X.
  
• there is a bureaucracy that has existed for a while. It has staff, budgeting and accounting procedures, offices, computers, photocopiers, an organization and a culture.
• the politicians look at the bureaucracy and say (to themselves): "No, we can't entrust our baby, our precious, to this existing bureaucracy. They wouldn't understand it, they don't share our view of its importance, they won't work hard to achieve the objective. What we really need is a new organization. We can pick the people who run it, getting those who believe in the objective, we can dispense with the old bureaucratic rules, we can get something done.
• and they do so, except the new organization has no cohesion, no procedures, no memory. Everything, and I mean everything, is new to it. What does any organization do when it faces something new--it has a meeting, to elicit ideas, to get everyone on board, to see if anyone has an answer to the question of how to turn on the lights. Net result--the new organization staggers along.
  

This post is prompted by the resignation of the head of the Millenium Challenge Corporation. It was proposed by Bush in March 2002. Three years and 2 months later it's just starting to make grants. I don't know when the actual money will hit the ground in the receiving countries. (See this pdf review.Also see the Corporations home page for a more optimistic view.)

The post is also based on my experience at USDA, where we spent years and millions trying to integrate operations of the agencies that service farmers. And when we get the books on the post-war Iraq, I suspect we'll find the creation fallacy operating both at DOD and State.

Bureaucratic Blinkers and Learning

A letter to the editor in today's Washington Post notes that it's been a year since the World War II memorial was opened on the Mall, but Metro's signage has not been updated. From the comfort of my chair, without researching, it shows the problem with bureaucracy. Presumably either the Metro person responsible for signs and maps in the subway or a PR type in the Parks Service might have acted. But each person each day likely focused on immediate concerns within their sphere of responsibility, not on a broader focus. (Popular office sign: It's hard to remember your goal was to drain the swamp when you're up to your ass in alligators.)

It's like commuting--you drive a route a hundred times and you put on blinkers and turn off your mind. You take the route for granted, it fades to be part of the environment you ignore. For the Metro person, even in DC new tourist attractions don't open that often, it's not like you go into the office each day and ask what's new. Similarly for the National Parks PR person--she/he would seldom have occasion to ask Metro to update its maps.

Bureaucrats are capable of learning. Consider the heads of NASA. Daniel Goldin (1992-2001) learned that his motto (something like: "better, faster, cheaper") had problems, notably when NASA lost a couple Mars missions, one when the unit of measure of botched between a European contractor and the US one (another instance of environments taken for granted) Sean O'Keefe (2001-5) learned from Challenger. In both cases I'd expect the learning to be bone deep--you don't fail so drastically and publically without it sinking deep. Now we have a new administrator of NASA, Michael Griffin, with new plans, firing O'Keefe's people and putting in his own. One thing we can be sure of, Griffin surely knows of Goldin and O'Keefe's mistakes, but it's book knowledge, not bone deep. Griffin will try to avoid his predecessor's mistakes, but will make his own. I hope they won't be traumatic.

The problem with Metro signage, though, is the relative rarity of the opening of a new tourist site. Bureaucrats change position and don't learn from rare untraumatic events. (Of course, I could be wrong. It could be that Metro delayed changing its signs because of the expense. When Congress changed the name of National Airport, a particularly vociferous Representative forced Metro to spend a few hundred thousand dollars to update its signs with the new name. What Metro hasn't learned is to replace its signs with LCD screens.)

Feinberg, Bureaucrat Extraordinaire

Kenneth Feinberg has a book out. He's the master of the special fund to compensate 9/11 victims, meaning he had both the bureaucrat's creative role, that of building a system interpreting Congressional intent in a law, and an operational role, that of applying his system to all applicants. The book should be interesting.

Data Mining in Medicare Data?

The LA Times editorial page thinks it would be a good idea to mine the health care data accumulated by Medicare for research purposes:

Medicare's Bright Idea: "Somehow it always comes as a surprise when a huge government bureaucracy proposes something sensible, efficient and geared to the public good. This time the happy shock comes from Medicare.

Smart people at the agency have put forth the idea of using its massive databank of patient information to spot potential problems with drugs more quickly. So why is the Food and Drug Administration, which oversees drug approval and safety, dragging its feet?"

I'm almost always for more knowledge but... Even though it's probably possible to separate data on medical problems and treatments from personal information, my perception is that it's a tough sell in this country. We aren't like Finland, which is more homogenous and more "regimented" (in the sense of having national ID cards and national databases). I wait with interest to see if the FDA agrees to this and gets it by Congress. It only takes one person in Congress to stick a provision in the appropriations bill to kill it.

Appropriations--Back and Forth We Go

Another item from the Justice Appropriations bill (from http://thomas.loc.gov):

" SEC. 103. None of the funds appropriated under this title shall be used to require any person to perform, or facilitate in any way the performance of, any abortion.

SEC. 104. Nothing in the preceding section shall remove the obligation of the Director of the Bureau of Prisons to provide escort services necessary for a female inmate to receive such service outside the Federal facility: Provided, That nothing in this section in any way diminishes the effect of section 103 intended to address the philosophical beliefs of individual employees of the Bureau of Prisons."

I'm curious whether Sec. 103 was added to the appropriations bill one year, then later a case came up that caused Sec. 104 to be added in the next go-round. Sec. 103 says "require", so conceivably a prison doctor could opt to perform an abortion without violating it. However, if the prison doctor does not perform abortions, then you'd have to transport a pregnant inmate whose life was being threatened by the pregnancy.

Privacy--Lessons from the Usability Lab

One of the innovations of the 20th century was the "usability lab" for testing software being developed and before release to the field. In my old agency, the lab was a setup where we'd put a person at a computer terminal, install new software on the computer, give the person a set of scenarios to execute using the new software, and videotape the operation, one camera on the terminal to see what she was doing and one on the user to see how she was doing (i.e., puzzled, stumped, flying through the exercise). Such testing was highly recommended by the software consultants and development experts, it was interesting to do, but we rarely did it. The reasons for that may bear on some concerns about the loss of privacy in the today's world.

The user had no privacy--we had 2 hours of tape for every hour she spent in the lab. That was the problem, we had too much data. Once the test was over, we could debrief the user quickly. But going through all the tapes would require spending the same amount of time watching as she had spent doing. It just wasn't a rewarding exercise.

It seems to me the same is true in many cases where people/companies collect personal data now. If there's no one with a strong motive to look at the data, and then misuse it, the risk to my privacy is negligible. Safeway's computer knows I buy cat food, and often provides a coupon at checkout to try to get me to buy a different brand, even when all I'm getting is milk and fruit. But that's all happening in bits and bytes. Even if someone at Safeway could look at my records, there's no motive to do so. The concern for movie stars and other famous people is different; they could attract stalkers and nuts who are motivated to find and use the data. (Like the IRS employee(s) who looked at Elizabeth Taylor's 1040.)

One of the things you can give thanks for is that you aren't famous.

Appropriations--The NRA Filament

The 2006 Appropriations Act for State-Justice-Commerce includes this bit, illustrating the power of the gun lovers:

"SEC. 634. Notwithstanding any other provision of law, no department, agency, or instrumentality of the United States receiving appropriated funds under this Act or any other Act shall obligate or expend in any way such funds to pay administrative expenses or the compensation of any officer or employee of the United States to deny any application submitted pursuant to 22 U.S.C. 2778(b)(1)(B) and qualified pursuant to 27 CFR Sec. 478.112 or .113, for a permit to import United States origin `curios or relics' firearms, parts, or ammunition."

Without researching further, I'd assume that an importer of say the "Brown Bess" British musket used in the 18th century still has to submit an application, but it can't be denied. It would be more logical, and accomplish the same goal, to revise the definition of a gun to exclude curios or relics. The problem with being logical is that the gun-lover who originated this (I'm blaming the NRA but it could have been an individual) has influence with the Appropriations Committee, not with the legislative committee and/or amending the basic legislation opens up a big battle over gun laws.

Appropriations--Cutting Phone Lines in the Executive

This is from the 2006 Agriculture Approppriations Act. I vaguely remember the incident that prompted it--I think the House committee had requested data from USDA. When the bureaucracy came up with the data, someone else got it at roughly the same time. Who was the "someone else"? Someone at OMB or the White House, if I remember correctly. Anyway at least one House member got on his high horse:

"SEC. 716. None of the funds appropriated or otherwise made available to the Department of Agriculture or the Food and Drug Administration shall be used to transmit or otherwise make available to any non-Department of Agriculture or non-Department of Health and Human Services employee questions or responses to questions that are a result of information requested for the appropriations hearing process."

So the result of the flap is to make such data top secret. The fungal filament creeps into the executive branch and cuts the phone lines.

Appropriations--Don't Coddle Inmates

From the Justice Appropriations Act (note, the URL's obtained when you search on the http://thomas.loc.gov site are temporary, so you have to redo the search.)

: " SEC. 112. (a) None of the funds appropriated by this Act may be used by Federal prisons to purchase cable television services, to rent or purchase videocassettes, videocassette recorders, or other audiovisual or electronic equipment used primarily for recreational purposes.

(b) The preceding sentence does not preclude the renting, maintenance, or purchase of audiovisual or electronic equipment for inmate training, religious, or educational programs."

So no cable for federal prison inmates, just the over-the-air programs. My impression is that most prisons are in rural areas so choice may be very limited. When TV came to upstate NY, the TV shop in Greene, NY set up a local cable system just because reception was so bad. Then I remember in the 70's seeing the first big cable dishes in those areas.

There's probably a home for this in the legislation governing the federal prison system, but again it's easier for the person who feels strongly about the issue to get it in the Appropriations Act.