Instead of observable prices of recreational visits, travel cost method (TCM) researchers are obliged to substitute researcher-assigned visitation cost estimates. I argue that visitation costs are inherently subjective, but are ordinally measurable so long as the cost increases with distance travelled. It follows that traditional TCM yields only ordinally measurable welfare estimates. The household production function formulation of TCM "resolves" this problem only by imposing severe and untestable analytical restrictions. TCM cannot serve as a stand-alone technique for estimating recreation benefits; rather, it must be calibrated using information generated with fundamentally different methods.
