We collected annual cross-section, time series data of visits to state parks in Arizona and New Mexico, and southwestern national parks. For national parks, we also collected monthly visit data and obtained data on visit forecasts made by the National Park Service Public Use Statistics Office. That agency also provided us with the data and programs of the forecasting models the Park Service uses to forecast annual visitation to each national park. In addition, economic and geographical data was collected and used to develop indexes of market potential for each park and of spatial competition between parks. Other control variables included fuel prices and exchange rates.
Changes in visits were mapped into local economic impacts using the MGM2 (Money Generation Model Version 2). The MGM2 model (developed at Michigan State University for the National Park Service) is an input-output model specifically designed to examine impacts of changes in national park visits. The model accounted for multiplier effects of spending. It provided estimates of direct, indirect, and total effects of changes in visitor spending. Outputs included changes in local sales, personal income, value added, and jobs.