The problem with this type of study is that it can by definition only cover the known externalities, but most of the price of externalities is far into the future. This means that the real margin of error is huge (and unknowable at this time).