How was this accomplished?

I know how to do most of this stuff, but how was the dancers tracking applied to the the bearded actor. It appears to me they were only tracking the the dancers head since his neck and body were very stiff.

tracking and gridwarps -- you could do this all in fusion. no need to use mocha.

My guess is
Track Dancer's head for the XY translation
apply that to the bearded actor only for the XY tranlastion
Track the static trackers on both side as a scale reference.
apply that scale to the bearded actor

note: also maybe use the center of the neck to be your pivot point for the scaling

I'll track the dancer with mocha and import the tracking comp. Then copy and paste that same tracking to bearded actor. That's easy enough, but how do I get things centered. Is there an easy way, or do I have to use manual offsetting?

If you are on a budget you don't need to have a "stand-in" or a tracking-dummy just practice your X-Y manual keyframe skills. I would assume that the old guys is divided into three parts for the sake of simplicity: head, body and legs. Since the dancing legs are grounded you can slap the body on top of them and mimic the X-Y motion based on the knees and hip flex to make it as real or cartoony as you want. Once you've done that then just simply track it and paste it so the two parts stay glued together. Last thing you might wanna do to add some additional quirky or realistic motion is to study movement relation between the head and the body a little bit and adjust X-Y keyframes for the head as well.

Since the whole spot is more fun and comedy oriented you cant really go wrong with manual X-Y adjustments. Have fun and dont forget that this could be either be a simple fun project or a very tedious but still fun project, it all depends how much you want to push it.

Piece of Cake with Fusion, you can make it as simple or complex as you want.