A model that combines application and license data may be 

 appropriate for the complex licensing system in Montana. 

 However, development of such a model is beyond the scope of this 

 project . 



It is important to note that licensing constraints mean that 

 the conventional TCM net economic values presented in this study 

 are conservative. In addition, because the special permit system 

 influences the distribution of hunting pressure across sites, the 

 difference in net economic values across sites reflect both the 

 influence of the permit distribution system and the underlying 

 site values. 



An additional complication in modeling the demand for 

 Montana deer hunting is the substantial difference in license 

 fees between residents ($9 for deer tag or $35 for a combination 

 license) and nonresidents (combination license $300 or $102 for 

 special permit). 



The basic method used for estimating the first stage demand 

 curve (Equation 1) is ordinary least squares (OLS) regression 

 analysis. The statistical assumptions include correct 

 specification, inference, and variable measurement. 



The main specification issues from a statistical standpoint 

 are the choice of regressors and the choice of functional form. 

 If the estimated model is incomplete, there is a danger of 

 omitted variable bias. Since the under!" ying model for OLS is 

 linear, it is necessary that the variable transformations 

 (implicit in the choice of functional form) result in a linear 

 relationship. There is a considerable literature on functional 

 form for travel cost models. The analysis reported below is 



8 



