That would go against the philosphy of the P-element solver, where you increase element order rather than element size.
Also, there will be no guarantee that the mesher can actually find a good mesh for a smaller element size, so you might run into issues where the analysis starts running okay and then complains in the middle of the analysis that the mesher fails. An engineer with little experience in FEA would probably have trouble understanding that, and curse at the software.
It would be nice if PTC were to explain how exactly they implement contact in the P-solver. It could also help in making better models.
But one of the drawbacks of Creo Simulate w.r.t. other FEA software is its poor documentation, especially about the underlying theory, so I don't think we'll see that anytime soon.