The 53db. A weighted noise figure was provided by the factoryAssuming it is based on one of the commonly used test conditions and reference levels, a 53db A-weighted s/n ratio is in a ballpark that I would expect to result in the audible hiss levels that you seem to be describing.
In addition to trying the shorting plugs, it might also be worthwhile assessing the hiss levels with the cartridge and cable connected and with loading set to 100 ohms or thereabouts, while you are not playing a record but with the volume control set to where you presently use it.
Perhaps the reason for the better noise performance that Tom reported is that the heavier loading (lower resistance value) that he was probably using in conjunction with a LOMC improved the s/n performance of the PH-1P, relative to what it would be with a 47K input impedance. That possibility would seem consistent with Lloyd's observation, as well. 47K input loading would also seem likely to increase rfi susceptibility, compared to a low value such as 100 ohms.
Also, I note that the Grado's inductance is 2mH, which is considerably higher than that of many and I believe most LOMC's, and would not come close at high frequencies to being the near-short across the phono stage's input that those LOMC's would be. The near-short would presumably reduce both self-generated noise in the front end of the phono stage and rfi susceptibility.
Just some thoughts. Regards,
-- Al