The transformers at my club (2 Z4000s) did this. It is now sometime ago and I don't recall much except under load the voltage readings seemed ok. This would be at the Z's themselves. We had close to 100 feet of #14 out (and the same back) to the TIU's plus maybe another 25 feet the same to the most distant track block connections.
I was more concerned about the voltage drop to the engines (I had a 4-pulmor diesel, which, at a certain point on the most distant rails of the most distant block, could move itself when restricted to a 2-pulmor consist, but not move with the addition of a single passenger car. Try 1.35 form factor (that's the heating and thus voltage drop in the Z output, relative to a pure sine wave, when set at 18 volts, compared to the same drop for the pure sine wave). That form factor takes 2-1/2 in wire size number to overcome. So #14 increased to #12 does not quite offset the form factor... not to mention trying to keep to the good practice of no more than 5% voltage drop in circuit. You see, the raw output of the main Z coil is 28-volts pure sine wave, but only the envelope of the 18-volt wave is a sine wave. The actual wave is somewhere between 300 and 600 cycles (the patent is written by lawyers)... I suppose the 400-cycle wire tables would be eye-opening enough. (That info will actually be useful.)
Anyway, faced with the cost of improving this situation or the awkwardness of moving the Z's away from the refreshments bar, spurious voltage readings were of little concern to me, and I gave them no further thought... until now. My best guess is that the meters involve some sort of voltage sensitive circuit coupled through a rather small capacitor(s), and amplified to drive the digital display by some off-the-shelf standard design. In an otherwise idle circuit, but having the output handle set at 18-volts, the active circuits (they were all somewhat bundled together, IIRC) induced voltage waves of opposite sign into the idle circuits, thus causing their meters to read less than 18-volts, or perhaps even causing them to read voltage when the handles were off... I cannot quite recall.
Look at this as being a perfect example of that old saying, "When you don't understand how something works, it will seem like magic." Sometimes life will just be easier if you leave it at that. 
(Sorry, I was in 6 hours of bumper-to-bumper 60mph to zero traffic today for 6 hours of lectures to get 5 PDH (the missing hour went to engineer jokes)).
--Frank
