So this is a good point. I was just talking with a colleague about this. Some of the herd immunity models take factors like how mobile people are into account. It seems like masks could be a factor in the models of herd immunity. This might be a semantic argument. If population characteristics such as age are part of herd immunity (which is a moving target depending on population), then mask wearing could be considered a population characteristic and included in the model. This would net "drop" herd immunity levels.