Predicting Timber Sale Costs 

 From Sale Characteristics in 

 the Intermountain West 



Ervin G. Schuster 

 Michael J. Niccolucci 



INTRODUCTION 



Economic analysis of timber sale design often requires 

 estimates of logging and roading costs, or how these costs 

 might change given a proposed sale modification. This 

 cost information is also important when assessing eco- 

 nomic efficiency for groups of timber sales and associated 

 road networks, where questions of timing and cost savings 

 are central to the analysis. Timber sale costs are impor- 

 tant, because as costs per unit of timber harvested go up, 

 timber value per unit volume goes down. 



Logging and roading costs are not directly incurred by 

 the stumpage seller. Neither can they be directly ob- 

 served. They are borne by and known to the timber pur- 

 chaser only. The stumpage seller can only surmise what 

 these costs are. In the Forest Service, U.S. Department of 

 Agriculture, special studies are conducted of logging and 

 roading costs. Purchaser records are inspected; time and 

 motion analyses are performed. Findings are presented 

 through a complex system of tables and charts in manu- 

 als, handbooks, and supplements. This information is 

 widely used by the Forest Service and other organizations 

 for various purposes, but primarily as the basis for cost 

 allowances when timber stumpage is appraised by the 

 "residual value" method (see Combes 1980). 



Unfortunately, this cost information can be very cum- 

 bersome or time-consuming to access and can result in far 

 more detailed data than are really necessary. In this 

 paper we present an alternative approach to developing 

 cost allowances. The following presents a set of equations 

 that can be used to easily estimate logging and roading 

 cost allowances. These estimates are suitable for use in 

 economic analyses of individual or groups of timber sales 

 wherever traditional cost allowances are used. 



METHODS 



The kinds of timber sale cost information needed to 

 appraise timber (stumpage) value with the residual value 

 (RV) method provide the framework for our modeling of 

 cost allowances. A simplified depiction of RV-appraised 

 stumpage value, treating the costs of permanent (speci- 

 fied) roads as a timber sale cost, is: 



Total value (of the products made from the logs) 



- Manufacturing cost allowance 



- Logging cost allowance 



- Roading cost allowance 



- Profit and risk allowance 



= Stumpage value (appraised) 



In the case of the Forest Service, agency policy indicates 

 that the appraised stumpage value will be based on an 

 operator of average efficiency (USDA FS 1977). This 

 means that cost allowances, product value specification, 

 and so on are all geared to the "average" operator. But 

 the highest (or winning) bidder may be of above-average 

 efficiency. Depending on a number of factors (USDA FS 

 1987), the highest bid on the sale may exceed the ap- 

 praised value by a "bid premium": 



Stumpage value (appraised) 



- Stumpage value (highest bid) 



= Bid premium 



Conceptually, bid premium can be related to the amount 

 of competition for the sale and/or incorrect specification of 

 total value or cost allowances, all relative to the winning 

 bidder. If cost allowances are excessive compared to the 

 winning bidder's actual costs, bid premium will be larger 

 than when cost allowances are inadequate. 



We developed five equations to predict timber sale cost 

 allowances and another equation to predict bid premium: 



Logging costs 



Stump-to-truck (fell, buck, skid, load) 1 



Transportation (haul, road maintenance) 2 



Slash 3 



Temporary roads 4 



Roading costs (permanent roads) 5 



Bid premium 6 



Normally, we would use the traditional multiple linear 

 regression to estimate each cost allowance equation. But 

 costs in one phase of the timber sale can affect costs in 

 another phase. For example, the method of felling and 

 bucking can affect the ease of slash removal and/or the 

 need for temporary roads. Hence, the costs and cost equa- 

 tions associated with these processes are not independent 

 of each other. Under this circumstance, conventional 

 estimates of the regression coefficients would be biased 

 and inefficient (Kmenta 1971). This problem was over- 

 come by using the technique of Seemingly Unrelated 

 Regression (see Kmenta 1971), a technique wherein all co- 

 efficients in all equations are estimated simultaneously. 

 Statistical tests were conducted at the 10 percent level. 



Data were obtained from records of a random sample of 

 224 timber sales completed between 1983 and 1985 on 

 National Forests in the Northern and Intermountain 

 Regions of the Forest Service (fig. 1). These were large- 

 volume sales, each containing 2 million bd ft or more. 



1 



