OTVN makes an excellent point that I'd like to expand on to make sure you understand what he's saying. If this is something you already know, then forgive me..
Sometimes, the actual video may be shifted even though the raster is centered - this is the horiz phase error OTVN mentions.
By lowering the contrast and raising the brightness, you can actually see the scan of the deflection circuits on the fact of the CRT. Alternatively, you can raise the G2 of the green tube a bit (not too far, and mark where it was) to see the raster. Look down the throat of the green lens. You will see if the raster sits right in the burn pattern the tube developed from years of running. With the TV on a snow pattern, you might see a slight overscan, which is normal, or slightly underscanned on a video input with no signal, which is normal.
If the raster is sitting more or less in the middle of the burn pattern on the face of the green tube, the deflection circuits and physical centering adjustments (ring magnets) are fine. At this point, you have a horiz phase problem. I don't recall if these models had a horiz hold, but it it does, you can see if that helps.
If the raster is off center, but otherwise full, you have a DC offset on the horiz winding.
Some TVs (Sony projectors) used a slightly different DC offset on each yoke, so mixing up the yoke plugs would cause a badly misconverged picture as the three yokes were nowhere close. I don't know if the Hitachi had this or not, but if it did, make sure the yokes are connected to the right colors on the board.
I tossed all my RCA manuals a couple of years ago so I no longer have that one. Is it on-line somewhere?
John