Is government secrecy “presumptively illegitimate”?

Yes, says Jack Goldsmith in a comment on Kinsley’s review of Greenwald’s book on Snowden:

I think Kinsley is also wrong about the normative question of who should decide.  The government should not have the final say about which of its secrets is published. Government action undisclosable to the American public is presumptively illegitimate. We tolerate secrecy to some degree because it is necessary for national security.

He continues:

But in a constitutional democracy where the People ultimately rule, secrecy from the People is aberrant (for they cannot govern what they do not know), and it demands special justification (thus the presumption against its legitimacy).

All of this sounds like it must be right until you think about it for more than a few seconds. In fact, government secrecy is pervasive, and it goes well beyond national security.

For example, the government’s records about everyone’s finances, collected by the IRS and other agencies, are secret. So are medical records for people in Medicaid, Medicare, and the VA hospital system. Employment files for millions of federal employees including military personnel–secret. Public school teachers’ evaluations of our children–secret. Social workers’ judgments about clients–secret. Deliberations of government officials–regulators, legislators, judges–secret. Evidence gathered in criminal investigations never used at trial–secret. Trade secrets and other private information collected by courts and regulators–secret. Algorithms and policies used to decide who to audit and investigate, where to set up roadblocks or hide plain-clothes police officers–secret. Security arrangements in government facilities–secret.

All of this secrecy requires a “special justification,” of course. But so does disclosure. If the government discloses my financial records to The New York Times, it will be asked to provide a justification. Everything the government does needs a (“special”) justification.

Now, to be sure, there are varying levels of secrecy. Sometimes secrets can be discovered through litigation; sometimes they are disclosed for policy purposes; sometimes they are disclosed as a matter of course after the passage of time. Indeed, government secrecy is almost always temporary–sometimes lasting only weeks or months, sometimes lasting years. But as a matter of practice as well as formal law, most of what the government does is secret for the period of time during which it matters, and we seem quite comfortable with that fact.

How can the people “rule” when so much of what the government does is secret? There are several answers. One is that the people can make rough judgments about overall outcomes–the healthiness of the economy, the existence of war, etc.–and cast their votes on the basis of those judgments, without knowing the details of government action that produced them. Another is that because the government is itself composed of people, and those people are autonomous, secret actions that are highly controversial or objectionable are likely to be leaked, or even blocked before they are put into place. A third is that review systems are set up within government. Whether or not the right balance has been struck is a difficult empirical question, and not one that is answered by the slogans of the founders.