I have a question if standard resistors like LN4040B that are supposed to be in oil might drift if taken out of a bath (or put in)?
The normal room temperature variations don't affect the very long term drift, but can trigger changes which need several months to recover.
During the six years I have had them they seems to have started to drift upwards slightly, maybe 1ppm/year but I am not sure.
How do you know that your reference (whatever it is) doesn't for example drift +3 ppm/year and the resistor +4 ppm/year?
Also the temperature coefficient of these resistors is in the 10 ppm/C range making the measurements without a bath more difficult.
The LN4040B was +47ppm when I got it and now I think it is +53ppm.
Because the resistors were manufactured a long time ago, you can deduce something by comparing the current value with the assumed initial value. Which usually was very close to the nominal value, but unfortunately varies a lot between individual resistors and the time on manufacture.
I have several LN resistors that was in a cal lab in Sweden before in a bath.
Did you get any historical calibration data from the previous owner?