This is a complete misuse of
Martin Quinn scores. To quote the publishers, they measure "the
relative location of U.S. Supreme Court justices on an ideological continuum." The center is always 0; doesn't matter if it is Lochner court or Warren Court, the center justice is always "neutral." They do not measure objective "conservative" or "liberal" voting. They only tell you contextually where someone falls on the court at any given time. Of course Kennedy was in the "middle" of the court; but that doesn't mean he's not conservative. The whole court has been conservative for decades!
Kennedy moving "towards the center" is of course true to some extent, but his appearance on the "left side" of the court at the end of his career is explained mostly by the appointment of much more conservative justices over the course of his career. (For example Thomas -> Marshall creates a big "leftward bump" for everyone; likewise, O'Connor -> Alito only shifts
everyone else noticeably left). Appointing ACB is going give
everyone but probably Alito/Thomas a "leftward" boost in their Martin Quinn scores, but that doesn't mean they've actually changed any of their positions.
I mean look at this graph:
Scalia magically becomes "much more liberal" during the Bush years after the appointment of Alito and Roberts. Scalia!
There is no reasonable measure by which Kennedy is a "liberal" based on his actual overall voting record.
This is why I seriously hate Martin Quinn scores. They are literally always misused by econ folks because no one bothers to read the methodology or understand the context of the court and the publishers are happy for people to think they are more important than they really are.