I am trying to capture economic contributions from existing visitors to a lake in Oklahoma using commodity change. I am curious what do you think about handling the potential issue of double counting (similar to multi-industry contribution analysis) in case of multi-commodity contribution analysis. How to handle it? Can I follow the appended link? If yes, I would love to hear theoretical justification from Phil.
Also, I have a question on Phil Cheney's post concerning 'considerations of industry contribution analysis'. As Phil mentioned in the link "contribution Analysis (Method) is a unique method which affects a constraint upon the Model by "removing" feedback linkages or buybacks to the Industry being analyzed. Typically, this method is used in conjunction with the IMPLAN Study Area Data because you are no longer looking at an individual firm, or a group of firms, but rather an entire Industry Sector. This method can also be used with single firms, but when it is the results of this method should be considered conservative".
So my question to Phil is:
Assume that there are two sawmills (sector 134) and three pulp mills (sector 146) in a given county and we are interested to know the economic contribution of just one sawmill and one pulp mill, what would be the proper method for economic contribution analysis. For simplicity, let's assume that we got expenditure data directly from these mills.