If Trump actually pardoned Flynn and Manafort, he would have to do so publicly and accept the political consequences of this profound act. As Jack Goldsmith suggests in the New York Times story, for those who believe that the pardon power is absolute and cannot be scrutinized by courts, the remedy for a corrupt pardon is in the political arena: elections or impeachment. What’s more, if Trump actually pardoned Flynn and Manafort, then the two men could no longer assert their Fifth Amendment privilege against self-incrimination because their pardons would erase their federal criminal liability, and therefore Mueller could call both to testify in the Grand Jury and in any subsequent trial. If they continued to assert their Fifth Amendment privilege on the basis of state criminal exposure, Mueller could obtain an order granting them so-called “use immunity” which would ensure that their testimony could not be used against them in any way in state court either. Manafort and Flynn would then be compelled to testify, or risk jail for contempt of court.

The pardon dangle works completely differently—and in important respects has the opposite effects. First, this kind of dangle is not a public act. Therefore, as long as it remained secret, it could be done without incurring any of the political downstream consequences that come with actually pardoning someone. It hides the President from scrutiny rather than exposes him to it as a potential check on the use of the power. Second, the objective of the dangle appears to have been to foreclose the prospect of Flynn and Manfort’s cooperating or testifying. Once again, this is the opposite effect of an actual exercise of the pardon. The message of the dangle was sufficiently clear: hang in there and keep fighting (do not cut a deal with the special counsel) because you will be pardoned before you spend a day in jail.