Robert Gellman | @Info.Policy: A new approach to classification
- By Robert Gellman
- Aug 13, 2006
It may be time for a whole new approach to an intractable problem'namely, overclassification of government documents. No one has found a reasonable way to control document classification, and the practice continues to proliferate. The federal government uses dozens of systems for marking documents, and most markings are meaningless.
We have plenty of documentation for the problem. Studies go back for decades with the same conclusion. We've had reports from commissions, the Government Accountability Office, etc. A newer report from the Congressional Research Service written by Harold Relyea goes through the long history of overclassification.
It's somewhat ironic that the CRS report, which isn't classified in any way, still isn't available directly to the public. CRS clings to the fiction that its products are only for Congress, so you have to go elsewhere to find a copy. Luckily, this isn't hard. The Federation of American Scientists maintains a lovely collection of CRS materials. (You can find this report by going to GCN.com
and entering 651 in the GCN.com/search box.)
Here's the idea. We should have a budget for every classification and administrative marking system used by an agency. Each person authorized to mark a document will have his or her own marking authority budget permitting the marking of a fixed number of documents each year. No one can mark a document without sufficient budget authority. Anyone who runs out of budget for marking documents will have to seek additional authority.
This system would not be limited to documents classified in the interest of national defense or foreign policy. Traditional national-security classification markings need stronger controls, but controls are needed even more for other less well-defined administrative marks such as For Official Use Only. The only classification that I would consider exempting from budget requirements is derivative classification for national-security documents.
Any marked document will require a control number and name of the classifier to be recorded along with the marking. The number identifies the budget to be charged. Marked documents will also have to be registered in a central system. Registration will add additional pressure against unjustified marking. Imagine a Freedom of Information Act request for all documents marked For Official Use Only during any particular period.
Today, that type of request would likely be impossible, but with a registry system, it would be simple.
By requiring the person who marked the document to be identified, we will have accountability. Today, no one is accountable for marking a document. Perhaps people will think twice before placing an Eyes Only or Top Secret mark on a newspaper clipping if there is a chance that a congressional committee might ask why someday.
The need for both limits and accountability is clear. There seems to be a deep-seated bureaucratic need to mark documents. It clearly doesn't matter what markings mean because in many cases, the markings have no clear definitions. It also doesn't matter what the consequences of markings are because often there are no consequences. A document marked official use is still subject to disclosure under the FOIA in precisely the same way that it would be if it were not marked. The current system places no constraints on document marking today. Since free markings are overused, let's use scarcity to control classification.
Budgets will only be meaningful if they actually limit classification. Budgets not used in one year should be reduced in the future. However, year-end markings should be monitored and discouraged, just like year-end spending. We will also need a method to allocate budget and provide extra authority when it can be justified.
A classification budget will cost something. Nevertheless, there is a good chance that the control system will actually cost less than what we spend today for the various classification systems. We can save lots of money if we reduce the lifecycle costs of classification by preventing classification in the first place.