My suggestion is that we would need to list all the variables and nail down as much of it as possible before trying to make an estimate.
How many pages per site? How much dynamic content per page? How complex is the dynamic content? (simple text ... CSS, scripts, data-driven content ... etc) How frequently are changes required? ... etc.
And even: how much will it hurt (and who will wear the hurt) if it costs twice what I estimate it to cost?
When somebody asks you for an estimate, it is generally wise to consider that they are asking you how long should be the piece of rope with which we will hang you?
Which is to say that you should be cautious. Ask as many questions as you can to determine exactly what it is that you're estimating. Assign values to as many of the variables as possible.
If you can't get answers and there are still many variables then you CAN NOT produce a reliable (i.e. useful) estimate.