Well, you'd have to take that up with the failover clustering team.  It is
their explicit recommendation for ALL clusters that you will install a
failover cluster instance in.

Not having it there is going to cause several errors during the upgrade as
well as throwing additional errors during a fresh install.  I haven't tested
enough permutations to say whether it is needed or not, so at this point, I
tend to trust the people who have spent thousands of hours testing hundreds
of different configurations.  They know the code a lot better than me since
they wrote it.  If that's the recommendation, I'm sticking into in this
case, because I have no way of proving that it shouldn't be followed.