The HP number they throw out is probably gross HP of the engine under optimal conditions. Real world is probably a little below that. Net (useable) HP is going to be less than that after everything is connected (alternator, transmission, etc)
If the engine isn't running at its peak horsepower RPM to achieve 540 RPM on the PTO shaft, the input power is lower resulting in a lower PTO hp.
ASABE calculated standards put PTO hp at around 83% of input power. 46 PTO hp puts the engine at 55.5 hp. That is a believable loss.
All dynos of the same type are going to vary amongst themselves. The variance is even greater if you throw wet brake up against an Eddy Current dyno. My guess is that they are using Eddy dynos in Japan.
Those are all my "shoot from the hip" thoughts.
I have the 540/540E PTO on my M6060 but I think it's standard on the M4. The gear reduction for 540 is 4.000:1 i.e. engine speed of 2,160 rpm. The gear reducion is less for 540E at 3.385:1 i.e. engine speed of 1,828 rpm. Dynamometers measure torque and convert to hp using the equation HP=(TQ x RPM) / 5,252. Is that engine rpm or PTO shaft rpm in the case of a tractor? If it's engine rpm then the position of a 540/540E lever would make a difference to the reported hp.
Just another variable to throw into the mix, but with such a small sample size of pre-owned tractors subjected to who knows what, and with questionable data collection and reporting based on information provided thus far, how can anyone make any sweeping statements regarding the technical specs provided by Kubota and real-world measurements.
Kubota has a reputation for playing it safe and understating performance specs, so I'll continue to trust their engineers and their literature over the incomplete and inaccurate statements of a random new forum member.