Changing the interface bandwidth statement, alone, shouldn't be disruptive except other (optional) features that use it might be disruptive by it being changed.
Hmm, the policy shape average statement, might be disruptive when you change it as congestion levels, due to the shaper, will likely change too. Going "slower" is likely to create more shaper queuing possibly increasing shaper queuing drops. Going "faster' may cause a TCP flow, in slow start, to create much larger transmission windows, that when they exceed shaper queue sizes, may create a huge drop incident, with the TCP issues that can cause. (The latter, I believe, you're less likely to encounter than the former.)
My experience has been, if there is any disruption, it's often just a "blip", often unnoticed. But if you want to "guarantee" no (negative) surprises during normal operations, best to make any config changes during scheduled maintenance.
