The correct answer is that you would select a bushing that left the necks at 0.307"
after sizing. That would generate the requisite .003" increase in neck tension (i.e.
interference fit) once the bullets are seated.
However, there are caveats to that in terms of the
exact bushing diameter needed. As noted above by
@Webster, bushings don't always yield the exact diametral value in the brass as the numerical value with which they are stamped. It is not uncommon to observe two or three different bushings, all stamped with the same diametral value to yield neck diameters of +/- .0005", or even .001", in actual use. Thus, you'll want to find one that leaves close to 0.307" neck diameter in the sized necks. That may mean a .307" bushing, or it could mean a .306" bushing that was slightly on the large side. Find one that gets you close and call it a day; for example, I wouldn't be concerned if the average neck tension turned out to be .0025" instead of the full .003". You can always test and determine optimal neck tension using precision as the readout, if desired, but worrying about it past half a thousandth not necessary for most applications.
In addition, springback of the brass can also have an impact as mentioned above by
@Coyotefurharvester. In my hands, springback is usually worth about .0005", but I generally try not to move the brass more than about .001" to .002" farther than necessary. For example, I generally first re-size brass using a bushing die with a bushing that will leave the necks approximately .001" to .002" smaller than I want them to be at the finish. Then, I will use a mandrel to open up the necks that will leave them with the desired neck tension/interference fit. In my hands, .002" neck tension is usually the target figure. Because the mandrel will increase neck diameter (i.e. the bushing die leaves the neck diameter smaller than the mandrel), when the neck is opened up by the mandrel it will close slightly due to springback as the mandrel is withdrawn. I typically find that a mandrel of approximately .0015" under bullet diameter will leave the neck at approximately .002" under bullet diameter due to the extra .0005" or so of springback. In general, the total amount of springback can depend on how far past the desired dimension one has moved the brass in an earlier step, but as a working figure, .0005" seems to be reasonable for me. None of this is really rocket science in terms of the mechanism(s) by which it occurs, it's merely an experience thing...the more one has worked with various different brass, bushings, mandrels, etc., the better they will know approximately the correct bushing diameter they need to try first. If that bushing doesn't yield the desired neck diameter, some testing may be in order. For that reason, I generally experiment with steel bushings; they're cheaper than nitrided bushings and so a little less painful to find the correct size.