From what I see, you've got no phase margin - 0deg or even negative, not quite visible from the plot.
Two comments. You have a fully differential amplifier but i don't see you using a common mode feedback circuit. So, most probably your amplifier is completely off. Can you check what is the DC voltage at the outputs of the first stage and at the outputs of the second stage?
You are not simulating the loop gain correctly. To simulate it, you have to put the amplifier in a DC closed loop and break the loop in AC. If you are using cadence, you could simply use the iprobe for that. The way you are doing it can not guarantee proper DC bias of your circuit.