There was a book written by O.C. Seevers with this condition listed in it. The title was "unique power systems problems solved". I wish I had a copy handy. It is about midway through the book.
Problem #6. It was a 69 kV line. The difference was 10 kV. The separation was 100 miles.
The speed of light yields one degree per 8.61 miles because (186,000 miles/s) / (60 cycles/s) / (360 degrees/cycle) = 8.611111 miles/degree
100 miles / 8.61 miles per degree = 11.6? = 8.04 kV.
The 2 kV in the 14.4 kV case gives us 7.964? or a distance of 68.6 miles. So if we are talking about a loop difference of about 70 miles, that would be about right assuming a vaccum. Probably 180,000 miles per second is closer for the wave propagation and yields about 66 miles for our problem.
The problem is, the OP is not going to have a 60-70 mile 25 kV line. Even allowing for some voltage drop would still be too long for this voltage level (around 52 miles if we allowed a 3% drop and have a remaining 6.24?)
I would not tie it.
My first guess was measurement error, but we need additional information on the circuit configuration and loading to figure out where the 6? to 8? is coming from. Any regulators or capacitors in the line?