It seems to be implicit in your model that funding for political parties is a negative-sum arms race.

It probably is on the margin. I'd guess that, while if both parties received 99% less donations there might be some kind of adverse effect, if both parties received epsilon less donations the effect would be of order epsilon squared or smaller.