But a 20% discrepancy isn't bad, and pretty common.
This is gospel.
I wouldn't send the rig in over this.
The real facts are that wattmeters aren't all that accurate to begin with, some, the spec is around +15/-5 percent of full scale.
And, the "missing" 20W wouldn't make enough difference in your signal to be noticed, if it were missing, which it most probably isn't.
Another way to find out what's really going on is to measure the Current draw at full power CW driving a known good 50 ohm dummy load. Then multiply the DC Current times the Regulated DC Voltage to obtain the input Watts. Then multiply by the Efficiency factor to obtain the realworld PA result (minus a small amount of current to run the other things in the radio, which are actually rather negligible in xmit mode anyway).
But, really, I think the odds are very good that your rig is functioning just fine at full output, it is the meter that is likely the culprit reporting low.