It has to do with strain at the alkene and therefore the rest of the ring. Consider a 180 degree dihedral between trans olefin substituents and the 120 degree sp2 bond angle. When you try to connect the trans substituents together by four atoms, the alkene gets twisted, pi orbitals don't align properly, the four atom linker gets pulled into a disfavored conformation, etc. Not so much of a problem with cis. Do you see why? Try drawing it.