It generally increases defensive challenge rating by 2 to 6, so final challenge rating goes up 1, 2, or 3...
The DMG guidance for creating a custom monster with resistances states:
For example, a monster with an expected challenge rating of 6, 150 hit points, and resistance to bludgeoning, piercing, and slashing damage from nonmagical weapons effectively has 225 hit points (using the 1.5 multiplier for resistances) for the purpose of gauging its final challenge rating.
So to determine defensive challenge rating, we see:
Read down the Hit Points column of the Monster Statistics by Challenge Rating table until you find your monster’s hit points. Then look across and note the challenge rating suggested for a monster with those hit points.
Since the proposed resistances means we multiply effective hit points by 1.5, just eyeballing the defensive challenge rating table gives us a DCR increase of +2 on the lower end, up to +6 on the higher end. Since resistances don’t affect Offensive CR at all, and final CR is the average of DCR and OCR, +2 to +6 DCR yields +1 to +3 final CR. The more hit points the monster has, the more significant of an effect adding resistance will have.
Finally, it must be observed that if the whole party can ignore the particular resistance, CR doesn't change at all. Phillipp's answer gives a robust discussion of this.
...except when defensive challenge rating is really low.
Now, when the CR is really low, like less than 1, it does get a little more difficult, as is the case with the lion mentioned in the question. Going off of hit points, the Lion's defensive CR comes out to 1/8, but because it has pack tactics and can possibly make two attacks per turn, its offensive CR comes out to 2, which is where the CR 1 final calculation comes from. Adding 50% on to the lion's effective HP only brings its DCR up to 1/4, which would still average out to CR 1 in the end. The method outlined in the first section seems to work just fine when working with DCRs of one or more, but the fractional challenge ratings complicate the averages and can lead to "no change" when the resistances are applied to monsters that already had really low DCRs.