So you have both red and yellow to 'live' connection, somewhere. 
If you really want to rule out capacitive coupling, then you generally need a test device that will draw some current. I have a bulb holder on a short length of flex that occasionally gets pressed into service to sort out low resistance connections to live from the ghost voltages caused by 'runs parallel to live for a bit and couples into it' the key is that modern voltmeters, and neon lights draw so little current they have difficulty telling if there is a real connection or not.
However, usually,as the induced core is equidistant from a live one and a ground, if it were capacitive coupling, I'd expect to see 120 odd volts not 250, so it sounds like in this case a real connection. check what happens at both ends of the cable, and if yellow is tied well out of the way as unused cores should be then look for cable damage.
If your meter has an ohms range then with the power OFF, you could measure the red to yellow resistance at both ends - the connection is near the end with the lower reading. (check by shorting the probes before you start that a low or zero resistance is recorded.)
AC (alternating and all that) does not care about meter polarity - it is like connecting half a cycle earlier or later so red/black makes no difference to the meter.
If you really want to rule out capacitive coupling, then you generally need a test device that will draw some current. I have a bulb holder on a short length of flex that occasionally gets pressed into service to sort out low resistance connections to live from the ghost voltages caused by 'runs parallel to live for a bit and couples into it' the key is that modern voltmeters, and neon lights draw so little current they have difficulty telling if there is a real connection or not.
However, usually,as the induced core is equidistant from a live one and a ground, if it were capacitive coupling, I'd expect to see 120 odd volts not 250, so it sounds like in this case a real connection. check what happens at both ends of the cable, and if yellow is tied well out of the way as unused cores should be then look for cable damage.
If your meter has an ohms range then with the power OFF, you could measure the red to yellow resistance at both ends - the connection is near the end with the lower reading. (check by shorting the probes before you start that a low or zero resistance is recorded.)
AC (alternating and all that) does not care about meter polarity - it is like connecting half a cycle earlier or later so red/black makes no difference to the meter.
