One of the improvements that was made in Process Version 2012 was a revamping of highlight and shadow clipping. I don't know that what you're seeing is really "out of whack" or if it is just the difference between process versions.
The exposure control has become more of a "brightness" control, in that it raises the values of mid-tone pixels without adding that much to the highlights. The wedding cake pixels measure under the clipping level. Like it or not, that is how Process 2012 functions.
Looks as if I've got to learn to use the Highlights and Whites sliders correctly.
The histogram is not indicating clipping in either situation.
If you want to upload that file to dropbox and post a public share link to it, maybe we'd see something else on our monitors with our newer versions of software.
What I would guess is happening is the dynamic highlight compression of PV2012 is keeping those brightest areas from actually clipping so you are able to raise the overall brightness by a significant amount.
It's also possible that ACR 8.1 that still works with Vista has an issue with highlights and the newer versions of ACR work differently, which is why it would be helpful to have your actual raw file to test with.