So to imagine that the courts actually confront a self-pardon, rather than a self-pardon that is essentially a press release, you have to imagine that the Justice Department has either built or is building a serious case against Trump. You have to imagine that this case is, or at least could be, immensely compelling, because the Justice Department is not going to bring a marginal case against Trump. There are enough prudential factors that weigh against indicting a former president that only a very strong case would even warrant serious consideration.

The first point here is that given the possibility of such a case, the self-pardon would function almost as a taunt to the Justice Department. A pardon given to anyone else would abort an incipient investigation immediately; all the investigative subject would have to do is plead the pardon if indicted, and a court would dismiss the matter. But a self-pardon is different. It presents one of the great open questions of constitutional law, and if the Justice Department backed down from investigating or indicting because a subject had pardoned himself, it would effectively be acknowledging the former president’s power to issue a pardon over the current president’s insistence on the traditional executive-branch position.

The Justice Department—assuming it had or thought it could build a powerful case—would be rightly wary of allowing the self-pardon to function as a complete negation of the justice system in the hands of an investigative target. Its only means of vindicating its law-enforcement interests in the specific case, its ability to conduct investigations of the president at all, and its traditional understanding of the law would be to go forward with the investigation or prosecution. This actually could encourage the department to contemplate cases against Trump that are weaker than those it would have contemplated had no self-pardon been attempted.