Predicting Salability of 

 Timber Saie Offerings in the 

 Forest Service Northiern Region 



Michael J. Niccolucci 



INTRODUCTION 



Timber sales play an important role in achieving forest 

 management goals and objectives — providing wood raw 

 materials, nontimber outputs, and government revenue. 

 The sale is planned by foresters, engineers, and other 

 specialists. It takes approximately 2 to 10 years to 

 develop a timber sale, depending upon sale size and 

 complexity. Once the timber sale plan is finalized, it is 

 offered to the public for purchase, fully expecting the sale 

 to sell at its initial offering. 



But sometimes a timber sale offering receives no bids. 

 The sale remains unsold. During 1980 through 1985, 

 approximately 20 percent of volume offered for sale re- 

 ceived no bid. An offering that fails to sell, given all of the 

 hours of planning needed, can be viewed as a waste of 

 dollars and human resources. Even if revised, reoffered, 

 and later sold, these unsold offerings may be obviously 

 seen as evidence of poor planning, professional incompe- 

 tence, or timber supply exceeding demand. Revisions 

 require added expense and still do not guarantee a more 

 appealing offering. 



Forest Sei^ce managers need reliable, defensible tools 

 to better predict salability. Also, given the years needed 

 to plan a sale, the sooner salability is known, the better. 

 Timber sale specialists can then either modify the sale 

 design or simply offer the timber sale without further 

 modification in light of expected salability. 



Currently, Forest Service managers in the Northern 

 Region have been using Transaction Evidence Equations 

 (Merzenich 1985), Timber Sale Feasibility Analysis 

 (Peterson 1980), and DLOGPRICE Economic Model 

 (Artley 1986) to quantify salability. Transaction evidence 

 is a multiple regression approach used to predict stump- 

 age value. Because these equations are tjqpically based on 

 sold sales only, they cannot be used to make reliable 

 statistical predictions regarding unsold sales. The sale 

 feasibility and DLOGPRICE models both rely on the 

 calculation of a value-cost ratio. Because the statistical 

 significance of these ratios cannot be tested, these proce- 

 dures cannot be rigorously defended. 



The research being reported here was designed to de- 

 velop and compare two statistical approaches to predict- 

 ing timber salabiUty at various points in the sale planning 

 process. Several questions were addressed: 



1. Is one approach consistently better than the other? 



2. Does our ability to predict salability improve as the 

 sale gets closer to the auction date? 



3. Does geographical zone variation affect salability 

 prediction? 



METHODS 



Methods used in this study fundamentally reflect 

 research choices about statistical classification methods 

 and how the timber sale planning process is envisaged. 

 The planning process provided this study a timeframe 

 within which to analyze timber sales. Fortunately, the 

 Forest Service currently uses a planning process that 

 contains the desired timeframe. Because a timber sale 

 selling or not selling is a dichotomous event, the desired 

 statistical technique should be capable of classifying 

 events into those classes. Logistic regression and dis- 

 criminant analysis were the classification methods 

 selected. 



Timber Sale Preparation — The Gates 

 Process 



The Forest Service currently uses a planning and deci- 

 sion making process called "Gates" to design timber sales 

 (USDA FS 1985, 2431.2). This process encompasses a 

 series of activities that begins with the identification of 

 a sale area and ends with a sale award. Sale planning 

 activities must pass through six reporting points, called 

 "gates." 



Gate 1: Sale preparation begins. This entails identify- 

 ing the purpose and the need for the sale; identifying 

 public issues; identifying the resource opportunities in 

 the sale area, and so on. Preliminary sale volume and 

 sale area estimates are also produced. 



Gate 2: Alternative sale area designs are developed. 

 Environmental effects are analyzed and a preliminary 

 economic analysis is completed. Gate 2 results in selec- 

 tion of a preferred alternative. 



Gate 3: The activities leading to sale plan implementa- 

 tion are performed. Preparation of the contract, data 

 gathering, and the necessary outline to support the ap- 

 praisal are examples of these activities. The sale passes 

 through gate 3 when the fieldwork and the timber sale 

 report are completed. 



1 



