Ak is distance from the home of l< to the nearest existing alter- 

 native facility offering recreation opportunities similar to those at 

 site j. 



Sjk is an index of the availability of substitute recreation facilities 

 (e.g., ocean beach for reservoir beach) in the market area cen- 

 tered at k's home. 



Qj is a vector of variables describing the quality of recreation at 

 site j. 



Ij is the increment or decrement in recreation at site j specified 

 in the contingent valuation mechanism. 



(ii) This method has several desirable character- 

 istics: (A) The VjK are current WTP estimates of 

 value for increments and decrements in recreation 

 opportunity; (B) the V, are annual values of the ex- 

 istence of the recreation facilities at site j, and thus 

 replace user days and unit day values; (C) the \/,k 

 are not arbitrarily set at the same daily value for all 

 users, as are unit day values; (D) the variables in 

 vector Q, provide a systematic statistical basis for 

 estimating how Vj varies with site quality; (E) the 

 variables C^, Sjk, and A^ provide a systematic sta- 

 tistical basis for adjusting Vj to account for compet- 

 ing and substitute facilities. 



(iii) Estimating a value estimator model requires 

 the following steps: 



(A) The final bids, after any calculations neces- 

 sary to convert them to annual or daily household 

 values, serve as the observations of the dependent 

 variable. 



(B) The observations of demographic variables 

 serve as observations for the first set of independ- 

 ent variables. 



(C) Existing recreation resource inventories and 

 planning data provide the basis for specifying the 

 second set of independent variables, i.e., those de- 

 scribing the existing stock of recreation opportuni- 

 ties. The location of each respondent's home is re- 

 corded on the completed survey instrument, and, 

 together with the inventory and planning data for 

 existing resources, permits calculation of individual 

 observations of those variables that relate the ex- 

 isting stock of recreation opportunities to the loca- 

 tion of the respondent's home. To complete the 

 task of specifying these variables, some indices of 

 the availability and quality of the existing recreation 

 stock must be developed. These include indices of 

 facilities and conveniences, and of site quality, es- 

 pecially esthetic quality. 



(D) Site-specific descriptors serve as the third 

 and final set of independent observations. These 

 are the data presented to the respondent and upon 

 which he based each of his bids. The estimated es- 

 thetic score of each photograph used in the bidding 

 process serves as one of these site-specific de- 

 scriptors. Other descriptors are the information pre- 

 sented to the respondent on size, distance, etc. 



(E) Using the best available econometric tech- 

 niques, the equation is then estimated. The de- 

 pendent variable is expressed in terms of annual 

 value per household, eliminating the need for sepa- 

 rate estimation of user-days and the mean value of 

 a user-day. 



(iv) Using an existing VEM to estimate the recre- 

 ation benefits of a proposed project involves the 

 following steps: 



(A) Determine the market area for the recreation 

 services affected by the project. If the market area 

 is expected to exceed 120 miles, document the 

 reasons. 



(B) Determine from census data the demographic 

 characteristics of the market area population. 



(C) Divide the market area into groups on the 

 basis of demographic variables and distance from 

 the proposed site. One such group might be 

 "households headed by a male of (ethnic group) 

 with 10 to 12 years of education and household 

 income between $12,001 and $15,000 annually, 

 living 51 to 75 miles from the site." 



(D) Calculate separately for each market subarea 

 the values of the variables describing existing rec- 

 reation facilities obtained from inventory and plan- 

 ning data. 



(E) Obtain from project planning data the values 

 of the variables describing project-specific attri- 

 butes. 



(F) Use the specified data and the fitted model, 

 to estimate the household value for the proposed 

 increment or decrement in recreation opportunities 

 for a typical household in each group. 



(G) Multiply this value by the number of house- 

 holds in the group, and sum the group values to get 

 the aggregate benefit estimate. 



(2) Applying CVM to a specific proposed project. 

 In some circumstances, CVMs may be used to esti- 

 mate the recreation benefits of a specific proposed 

 project. Great care must be taken in the design of 

 the survey instruments and editing of the data, 

 however, because some respondents may try to in- 

 fluence the outcome of the analysis by their bidding 

 responses. The survey design and sampling re- 

 quirements of such a study are discussed under 

 "Data requirements" below. 



(3) Data requirements— {\) Survey design. For 

 contingent valuation exercises, the survey instru- 

 ment must contain two major sections: One for bid- 

 ding formats and one for collecting appropriate de- 

 mographic data; a brief final section should elicit re- 

 spondent feedback. Since there is no reason to 

 prohibit the use of additional sections, other data 

 useful for recreation planning may be gathered 



82 



