What you're talking about is counterfactual inference, which is a non-trivial problem even in well-controlled settings. The effect of large-scale protests would probably be very difficult to quantify with the rigour you're alluding to. I'd love to be proven wrong, though.