THS4531: CMRR of ths4531

In a way it does, if you drive the two input R's with  CM signal, it will generate an output differential term arising from the feedback ratio mismatch - that does simulate. Incidentally, the THS4531A is the improved DC precision version (offset voltage)

if there is a 1% mismatch between feedback ratio then what would be the reduction in CMRR.

if your mismatch is exactly +/-.5%, your CMRR will be 40dB

As I said Balwinder, this will simulate, 

Here I put +/-1% uniform tolerance on the R's. Worst case mismatch should have given -34dB CM to DM conversion. And I could have forced that in the R values for a nominal run. But, V11 allows monte carlo, here is 400 cases getting close to that worst case -34dB. If you need better, you need to use tighter tolerance R's. I think the rise at higher F is the bandwidth of the CM loop in the model.