That's the wrong way to look at it. Populations are almost never homogeneous. The Proto-Italo-Celts from the Pontic Steppes or Danube region almost certainly were an admixture comprising various clades of R1b-L11, including R1b-L11*, R1b-S21, R1b-S116*, R1b-L21, R1b-S28, R1b-Z196, + many of their subclades + other minor side lineages. Even subclades as deep as L20 (subclade of L2, itself a subclade of S28) or S68 (subclade of L176.2 and Z196) probably existed before the invasion of Western Europe by the Indo-Europeans. It's just that they were limited to a few individuals, and only expanded a few centuries or millennia later, either because this clan fared better than the others (more food, less war casualties, less diseases, more children) or because a member of that clan got into a position of power (new dynasty) that allowed him and his descendants to have more children than other people.
Don't think of the PIE as a single group, but more of a confederation of tribes, with plenty of chieftains and minor kings, a bit like in ancient Gaul, ancient Britain/Ireland and among ancient Germanic people. We can deduct from the analysis of PIE languages that PIE society was strongly hierarchical and patrilinear, but also clannish, just like the ancient Celtic and Germanic people. In this context, the dominant male (chieftain/king)'s lineage can easily expand exponentially.
My point is that S116* in Iberia is not the ancestral population of the subclades of S116 elsewhere in Europe. S116 is found throughout Europe, in Anatolia and even in Russia and Central Asia. Iberia has more S116 simply because the Indo-Europeans who migrated there just happened to have more of this lineages. Actually it is almost certainly not S116*, but a subclade that hasn't been discovered yet (or one of the numerous subclades that was discovered this year, like Z196, S182 or DF19, but wasn't tested in older Iberian studies).