Mach-Zehnder interferometer and Power redistribution

I was trying to run the ‘Interference Example 2: A Mach-Zehnder interferometer’ simulation file from the Zemax Opticstudio directory (non-sequential mode), but I couldn’t see any power redistribution between the two output arms of interferometer. The intensity pattern distribution in one arm should be complimentary (negative) to the other one. But it is not happening in simulation. In a perfectly parallel beams with zero divergence, I would expect change in power splitting in output arms with phase difference. That is not happening in the given sample simulation. Please share your comments/explanations for this issue.