This section of the report presents the calculation of roundwood cost for a "typical" or "representative" forest harvesting operation in Suriname. The section starts by describing the likely scale and level of productivity of the "typical" forest operation, based on information collected during the study. It then goes on to describe how roundwood production cost can be calculated using the harvesting cost model developed for this study. It finishes by summarising the results and showing the relative importance of the various components of the total production cost.
Most of the former forest concessions in Suriname are located in the "forest belt". This strip starts between 5 km and 20 km from the coast and reaches up to 80 km to 100 km inland. A few sawmills are located within the forest or part of the way up some of the larger rivers reaching into the interior of Suriname. But, the majority of them are located in the coastal towns of Paramaribo, Nickerie, and Moengo. Based on the location of most sawmills and former forest concessions therefore, it has been assumed that the average haulage distance from forest to sawmill is 85 km and that all roundwood is transported entirely by road.
The above assumptions are very general. Many forest concessions are currently working at far less than potential productive capacity (as are many of the sawmills). Some forest concessions also reload roundwood onto barges or pontoons and transport it by river to sawmills. However, the above assumptions have been used as a reasonable basecase scenario to show what the costs of a "typical" forest operation might look like and demonstrate how the harvesting cost model works.
Sheet 2: Miscellaneous data
The second sheet in the harvesting cost model is used to enter general information about the forest concession such as: information about forest area and stocking; haulage distances; skidding distances; the number of times roundwood is unloaded and reloaded; and miscellaneous variables such as the exchange rate, interest rates and the required rate of return. A copy of the second sheet for the basecase scenario is shown in Figure 1.
Figure 1 A copy of the second sheet of the harvesting cost model showing miscellaneous data entered into the model
The level of expected annual production (discussed above) is calculated in cell C6 from the annual cutting area and harvesting intensity (entered into cells C4 and C5). The market exchange rate used for the whole of this analysis is Sf 650 to US$ 1.00 and this is entered into cell H4. The market interest rate in Suriname is currently around 45% and this is entered into cell H6. The expected level of profit is the required rate of return on capital or markup on other expenditure which has been used in the analysis. In other words, this is the level of normal profit which forest concessionaires should be allowed to retain. Based on discussions with LBB, SBB, forest managers and sawmillers, this has been set at 20% and is entered into cell H5.
Information about transportation is entered into rows 10 to 12. Distances travelled on each of the different types of road: public road; main forest road; and branch forest road, are entered into cells C10, D10 and F10 respectively. In the base case, these distances are assumed to be 80 km, 2 km and 3km respectively (giving the overall total distance of 85 km as discussed above). Travelling speeds are entered for each type of road below the distance. The speed when loaded is entered into row 11 and the return speed when empty is entered into row 12. It has been assumed that timber trucks can travel at 35 km/h, 30 km/h and 20 km/h on public, main and branch roads respectively, when loaded and 10 km/h faster when unloaded. These assumptions about travelling speeds were made on the basis of field observations and discussions with forest managers. The total duration of one trip (from forest to sawmill and back) is calculated in cell F13. Similar information can be entered for barges or pontoons in column G if water transport is also necessary (assumed not to be the case in the base case scenario).
Information about road and skid trail construction is entered into rows 17 and 18 of the model. The existing length of forest roads in the forest concession is entered into cells C17 and D17. The figures are entered in terms of the roading density or the length or roads per hectare of forest (in m/ha). The length of forest roads in Suriname is currently unknown, but believed to be relatively small in comparison with the area of forest concessions. Therefore, a relatively low density of 5 m/ha in total (2 m/ha main roads and 3 m/ha branch roads) has been assumed in the analysis (cells C17 and D17). Also, very few forest managers are currently building forest roads, so it has been assumed that new road building is zero (cells F17 and G17). Cells F18 and G18 show the construction rate for new road building in terms of length of road construction per hour (in m/hour). These figures would determine the cost of road building later on in the model. However, because cells F17 and G17 are set to zero, the part of the model that calculates the cost of road building is not activated.
The lengths of skid trails constructed by bulldozer and/or skidder are shown in cells H17 and I17. The model will work with skid trails constructed by bulldozer (e.g. Caterpillar D6) or skidder or a combination of both machines. The base case scenario assumes that a typical forest concession only uses skidders for skid trail construction and skidding.
The length of skid trails constructed will depend upon several factors including: roading density; distance between skid trails; forest stocking; the degree of skid trail planning; and site factors such as terrain. A model was constructed to estimate skid trail construction and skidding lengths for a range of combinations of these variables. Based on current forest harvesting practices, this model determined that the typical amount of skid trail construction Suriname is currently about 245 m/ha. Discussions with forest managers and site visits suggested that skid trails could be constructed at a rate of 100 m/hour and this is entered into cell I18.
The average skidding distance is calculated in cell C24. This is calculated from the average skidding distance (which is itself calculated from the roading density  see Box 2), adjusted by the indirectness factor entered into cell C22. A figure of 82% was entered into the model based on the skidding model results. The resulting average skidding distance of 910m is very long, but was confirmed to be probably about right in discussions with forest managers and staff of LBB and SBB. To reduce skid trail construction costs and site damage, it is possible to winch felled roundwood to the skid trail and the average winching distance can be entered into cell C23. This is not, however, currently common practice in Suriname, so a value of zero was entered into this cell in the base case scenario.
The speed of winching and skidding should be entered into cells H22 to I24. Skidders are used for skidding in the base case scenario and the winch isn't used, so the only figures which are used in later calculations are the skidder's speed when loaded (cell I23) and speed when unloaded (I24). Figures of 6 km/h and 10 km/h were entered into these two cells respectively, based on the machine performance figures quoted in the Caterpillar Handbook (Caterpillar, 1996). The proportions of skidding and winching performed by skidder and bulldozer are entered into cells H25 to I26. Zero is entered into cells H25 and I25 because winching is not used and zero is entered into cell H26 and 100 into cell I26, because only skidders are used for all skidding operations.
The total average duration of each skidding trip (including the average time to construct necessary extensions to the length of skid trails) is shown in cells H27 and I27. This is calculated from: the total skidding and winching distance; skidding and winching speeds; and the length and speed of skid trail construction discussed above. This result is only calculated for a skidder because only skidders are used in the base case scenario.
The last piece of information required on this sheet is the number of times roundwood has to be unloaded and reloaded in cells L4 and L5. This is used to calculate secondary loading costs later on in the model. For example, if roundwood has to be unloaded at a riverside and loaded again onto barges, a figure of one has to be entered into each of these cells. It is assumed in the base case scenario, that timber is transported directly by road to the sawmill, so a figure of zero is entered into both of these cells.
For information, the total roundwood production cost in Sf and US$ (including an allowance for normal profit) is also shown in cells C27 and C28 on this sheet.
Box 2 The calculation of average skidding distance from roading density and the indirectness factor
The calculation of average skidding distance from roading density and the skidding indirectness factor is explained below using the example of the base case scenario. The average theoretical skidding distance The easiest way to visualise the calculation of the average theoretical skidding distance from roading density is to think of strips of forest served by skid trails leading on to forest roads. Thus, for example, assuming that main skid trails are 200m apart and that roading density is 5 m/ha, two main skid trails, one off each side of a 200m segment of forest road, must serve 40 ha (200 m divided by 5 m/ha). The length of road is the length of one side of the strip of forest served by the road so, from the total area (ie. 40 ha), the length of the strip can be determined. One hectare equals 10,000 m^{2}, so 40 ha equals 400,000 m^{2} and, if the road side is 200 m long, the length of the strip of forest served by the road must be 400,000 m^{2} divided by 200 m, which equals 2,000 m. This is shown in the figure below: The total length of the strip on each side of the road is 1,000 m, so the centre of the strip on each side of the road and, hence, the average theoretical skidding distance is 1,000 m divided by two, which equals 500 m. More generally, the theoretical skidding distance can be calculated using the formula below: Theoretical skidding = 2,500 . distance (in metres) roading density (in m/ha) The skidding indirectness factor In reality of course, it is very difficult to achieve the average theoretical skidding distance because site factors such as steep areas and streams make it necessary to deviate from the most efficient route to each tree. Poor skid trail layout can also lead to a large amount of deviation from the most efficient route. Therefore, the above calculation is usually multiplied by an indirecness factor to take into account these considerations. A model was constructed for this study to examine the impacts of stocking, skid trail layout and roading density on skidding length, which suggested that, in the base case scenario, a high level of indirectness (82%) would probably be typical in forest concessions in Suriname. The final average skidding distance shown in sheet 2 of the harvesting cost model is therefore, equal to the theoretical average skidding distance plus this factor which, with a roading density of 5 m/ha, equals 500 m x (100% + 82%), which equals 910 m. 
Sheet 3: Productivity
The third sheet in the harvesting cost model contains information about rates of machine productivity for each of the various forest harvesting activities. The purpose of this sheet is to estimate the amount of capital equipment needed to produce the annual volume of roundwood production specified in sheet 2. A copy of the sheet, showing the information about productivity entered into the model for the basecase scenario, is shown in Figure 2.
Figure 2 A copy of the third sheet of the harvesting cost model showing information about productivity entered into the model
The model is designed with the assumption that most forest managers will use their own equipment for production. Therefore, a certain amount of flexibility has been built into the model to reflect different circumstances. For example, the amount of skidding by bulldozer can be set to zero on sheet 2 (the previous sheet), in cases where all skidding is done by skidder (as in the base case scenario). If this is done, all further cost calculations under the activity: "Skidding by bulldozer" are set to "NA" (i.e. not applicable) irrespectively of the cost data entered into cells under this activity. Alternatively, skidding by bulldozer could be set to 100% on sheet 2 and skidding by skidder could be set to 0%, in which case all calculations under the activity: "Skidding by skidder" would show "NA". Similarly, transport distances by road or by water can be set to zero on sheet 2, depending on how roundwood is transported from the forest to the sawmill or harbour. In the base case scenario it is assumed that all haulage is by road, so costs results under the activity: "Water transport" are set to "NA". The same is true of "Unloading and reloading" and "Road building" under the base case scenario.
The only activities that absolutely must be specified in this part of the model are "Felling" and "Loading" (or else the model will not calculate results). In cases where these activities are carriedout by contractors, the costs of these services can simply be substituted for the costs of these activities calculated by the model, when the final results are calculated. Similarly, the cost of other contracted services can also be used to calculate the total cost in the final analysis, where these activities are contractedout.
Machine productivity is a combination of several factors, including: the number of hours worked each year; the amount of time machinery is unavailable for work due to repairs; the amount of time it takes to perform essential functions such as refuelling; and the rate of production when the machine is actually working. This spreadsheet calculates productivity and machine utilisation based on information entered into the model about each of these factors.
Information required to calculate the number of hours a machine is available to do work each year is entered into rows 5 to 7 of the productivity sheet. The number of working days is entered into row 5 of the model for each activity and the number of working hours per day is entered into row 6. Discussions with forest managers and staff from LBB and SBB suggested that 200 working days per year and an 8 hour working day were normal practice in Suriname and these figures have been used in all the analysis presented in this study. Average machine availability is the amount of time that machines are not brokendown or away from the forest for servicing and repairs. Although most of the harvesting machinery in Suriname is quite old, the scarcity of machinery ensures that forest managers try to reduce work stoppages due to breakdowns, servicing and repairs. Therefore, a relatively high value of 90% has been used in this analysis. The number of available working hours each year is calculated in row 8 of the sheet, using the following formula:
available working = number of working x number of working x average machine
hours each year days per year hours per year availability
The number of available working hours represents the amount of time each year that each machine is fully functioning and ready to perform harvesting activities.
In addition to major breakdowns, servicing and repairs, there are other time losses i n the operation of machinery due to factors such as: breaks and rest periods for staff; refuelling and minor onsite maintenance operations; and time spent waiting for other parts of the harvesting operation to complete their activities. For example, timber trucks have to wait to be loaded every time they make a trip. The time that a piece of machinery is capable of working after such losses have been taken into consideration is called its effective working hours.
Information required to calculate effective working hours is entered into rows 9 and 10 of the sheet. For some machinery, it is easier to think of unproductive time in terms of unproductive time per trip (e.g. the time taken to load a timber lorry or the time taken to attach logs to the back of a skidder) and this is entered into row 9. For other pieces of machinery, it is simpler to think in terms of unproductive time per hour (e.g. the average amount of time per hour which tree fellers take to walk between trees, take breaks and refuel and maintain their chainsaws). Where this is the case, this information is entered into row 10 of the model. For activities where unproductive time is entered in terms of time per trip, this is converted to an average time per hour in row 10, based on the average duration of trips shown on sheet 2.
The figures entered into these rows for each of the harvesting activities in the base case scenario, were based on discussions with forest managers, the staff of LBB and SBB and the author's previous experience in this area. They have also been kept the same throughout the analysis.
The number of effective working hours each year for every harvesting activity is calculated in row 11 of the sheet, using the following formula:
effective working = available working x 60  unproductive time (in minutes per hour)
hours each year hours each year 60
The number of effective working hours represents the amount of time each year that each machine is fully functioning and capable of performing harvesting activities.
The last pieces of productivity information are entered into rows 12 to 14 of the sheet. Rows 12 and 13 are used to enter information about the volume of timber that can be handled by each piece of machinery either by the trip (row 12) or by the hour (row 13). Again, information entered on a trip basis is also converted into an hourly figure, based on average trip duration. It is important to note that this information should be entered in terms of productivity assuming no stoppages for breaks, maintenance and refuelling etc. Thus, for example, the productivity of felling has been entered as 8 m^{3}/hour on the basis that an average tree is 2 m^{3} in volume and that it takes 15 minutes to fell, delimb and crosscut a tree. The time taken to walk between trees, refuel and maintain the chainsaw and breaks, is accounted for in the calculation of effective working hours.
The figures used in the base case scenario are based on discussions with forest managers, staff of LBB and SBB and field observations at the Bruynzeel forest concession. Felling productivity was explained above. The productivity of skidders and bulldozers is based on an average tree volume of 2 m^{3} and two trees per skidding journey. The productivity of loaders is based on observations of the loader employed at Bruynzeel. Productivity of road transport and water transport is based on the average capacity of timber trucks, barges and pontoons currently used in Suriname. The average capacity of most timber trucks used in Suriname is quite low and this will increase the roundwood production cost compared to what could be achieved with more efficient forest operations. The productivity of road building machinery is not expressed in terms of m^{3} production, so this information is not presented here. Rather, in the case of road building machinery, capacity utilisation is a function of the time taken to build roads and the planned roading density (shown on sheet 2) and this is used to calculate machine utilisation on this sheet.
The number of machines used in the forest concession is entered into row 14. This variable is changed until average machine utilisation is reasonable. Annual production per unit is shown in row 15 and is calculated by dividing the annual production shown on sheet 1 by the number of machines used (above).
Actual working hours, shown in row 16, is calculated by dividing the annual production per unit (in row 15) by the hourly output rate (in row 13). This represents the amount of time that each machine must operate in order to produce the output shown in row 15. This figure is used later on to calculate consumption of consumable items such as fuel and lubricants.
Average machine utilisation is shown in row 17 and is calculated as:
Average machine = Actual working hours .
utilisation Effective working hours
This represents the proportion of effective machine time that is actually used in the forest operation. A utilisation rate of over 100% would indicate that the amount of work required is greater than the time available. Therefore, the number of machines specified in row 14 must be adjusted until the utilisation rate is reasonable (i.e. in most cases  less than 100%, although it may be possible to have machine utilisation of somewhat more than 100% if operators work overtime). For example, one skidder in the basecase scenario would have to work for 2,052 hours to extract 9,600 m^{3} of timber in a year, but the number of effective working hours is only 1,205 hours (which would give a utilisation rate of 170%). Therefore, two skidders must be used in the forest operation.
Sheet 4: Labour
The fourth sheet in the harvesting cost model contains information about the cost of labour inputs used in each of the various forest harvesting activities. A copy of the sheet, showing the information about labour costs entered for the basecase scenario, is shown in Figure 3.
Figure 3 A copy of the fourth sheet of the harvesting cost model showing information about labour costs entered into the model
Most forest harvesting machinery is operated by a machine operator with one or more assistants. The number of operators and assistants used per machine is entered in rows 5 and 8 of this sheet, under each harvesting activity. Generally, most machinery in Suriname is operated by a crew of one machine operator and one assistant. The exception to this is barges, which may use more than one assistant to help operate the vessel.
Rates of pay for these workers can be entered in two ways. Monthly pay rates can be entered in row 6 (for the machine operators) and row 9 (for assistants). Alternatively, if workers are paid on a piecerate system, the payment per m3 produced can be entered into rows 7 and 10. If figures are entered for both alternatives by mistake, the piecerate figures are used in the calculation of total labour cost shown at the bottom of the sheet.
The pay rates entered into this sheet for the basecase scenario were based on information collected from forest managers (see Whiteman, 1999). Many forest managers use local villagers to cut timber and pay them Sf 750/m^{3} in total, so this figure was used in the basecase scenario. Most other operations are performed by salaried staff and the figures shown in Figure 3 are believed to reasonably reflect current rates of pay in the forestry sector in Suriname.
In addition to the direct labour costs in forest operations, there are also two other costs that must be considered. They can be described as follows:
Social cost. The social cost of employment is the additional cost of: insurance; sickness benefits; medical benefit; and (sometimes) housing benefits and food allowances, given to employees. Some of these benefits (e.g. medical benefits) may extend to the whole family of the employee.
Overhead. The overhead on top of direct labour costs includes other costs of employment such as the cost of camping equipment and transport to and from the forest, where the employer provides these facilities. It can also include such items as the cost of employing additional staff not directly working in the forest, such as: the exploitation manager; security staff; clerical staff; drivers and other assistants.
A figure of 40% was commonly quoted as the social cost of employment and this figure was applied to the cost of all harvesting activities in the basecase scenario except felling (where it is doubtful whether such benefits are extended to local villagers). These figures were entered into row 11 of the model.
Two main overhead costs were considered in the basecase scenario: the cost of providing camp equipment for workers who have to stay overnight in the forest; and the cost of transporting workers to and from the forest. Rough calculations indicated that the former cost might add around 5% onto the direct cost of employment and the latter might add a further 20%. Therefore, an overhead figure of 5% was included in the basecase scenario for felling activities (local villagers camp in the forest, but tend to live fairly close, such that transportation is not likely to be very expensive). A figure of 25%, to cover camping and transportation, was included for all other workers who have to stay in the forest. These figures were entered into row 12 of the sheet.
Other overhead costs such as additional staff (e.g. exploitation manger and clerical staff) were not considered, because only the largest forest operations tend to employ such workers. Where these functions have to be performed in small to mediumsized forest operations, they are usually carriedout by staff of the sawmill.
The total labour cost of production, by activity, is shown in Sf/m^{3} and US$/m^{3}, in rows 17 and 18 of the sheet. For activities paid on a piecerate basis, this is calculated as the operator's and assistant's payments (per m^{3}), multiplied in each case by the number of operators and assistants used in the activity, then multiplied again by (1 + Social cost % + Overhead %). For salaried staff, this is calculated as the annual staff cost for one machine (i.e. (number of operators x operator's monthly salary) + (number of assistants x assistant's monthly salary), all multiplied by 12), divided by annual production per machine (from sheet 3), then multiplied again by (1 + Social cost % + Overhead %).
Sheet 5: Consumables
The fifth sheet in the harvesting cost model contains information about the cost and rates of consumption of consumables used in each of the various forest harvesting activities. A copy of the sheet, showing the information about consumable costs entered for the basecase scenario, is shown in Figure 4.
The information which has to be entered into this sheet falls into broadly three categories: fuel and lubricant costs and consumption; the consumption and cost of filters, tracks or tyres and other minor parts; and the cost and consumption of other miscellaneous items such as tools, protective clothing and miscellaneous items.
Figure 4 A copy of the fifth sheet of the harvesting cost model showing information about consumable costs entered into the model
The unit cost of fuels and lubricants is entered into cells C5 to C9 (in Sf per litre). These costs were taken directly from list prices collected in Suriname (see: Whiteman, 1999). The rate of consumption of each of these inputs were taken from the Caterpillar Handbook (Caterpillar, 1996) and Bruynzeel's records (e.g. for the use of fuels and lubricants in chainsaws) and were entered into rows 5 to 9 of the sheet, under each of the harvesting activities. These costs and rates of consumption are standard throughout the whole analysis.
The cost of filters, tracks or tyres and other minor parts are entered into the model in terms of the average cost in Sf per actual working hour, in rows 14 to 19. The hourly cost of these items was calculated by dividing the cost of each item (collected in Suriname) by the estimated life of each item (estimated from the Caterpillar Handbook and Bruynzeel's records). For example, the cost of an air filter for a skidder or loader is about Sf 8,000. Skidders and loaders have two air filters: one has to be replaced every 2,000 hours and the other has to be replaced every 1,000 hours. In other words, three air filters have to be replaced every 2,000 working hours.
Thus, the total hourly cost of air filters for a skidder or loader was calculated as follows:
Hourly cost of = number of air filters changed x cost of an air filter
air filters number of working hours that the filters are used
= 3 x Sf 8,000 = Sf 12/hour
2000 hours
Using the above formula, the hourly cost of all such parts was calculated for every piece of equipment used in the forest operation.
The cost of tools, clothing and other consumable items are entered into the model in terms of the average annual cost in Sf per year, in rows 20 to 22. These were estimated from cost information collected locally and a rough approximation of annual use.
The total labour cost of production, by activity, is shown in Sf/m^{3} and US$/m^{3}, in rows 27 and 28 of the sheet. Total annual fuel and lubricant cost was calculated by multiplying the unit cost (cells C5 to C9) by hourly consumption (rows 5 to 9) and the number of actual working hours shown on sheet 2 and then adding all these items together. Total annual filters, tyres, tracks and other minor parts cost was calculated by multiplying the cost of each of these items (rows 14 to 19) by actual working hours and adding them together. The total annual cost of other consumable items was calculated by adding together the costs shown in rows 20 to 22. These three annual costs were then added together and divided by annual production per machine (on sheet 2) to give the total consumable cost per m^{3} for each activity, shown in rows 27 and 28.
Sheet 6: Capital
The sixth sheet in the harvesting cost model contains information about the cost of capital equipment used in each of the various forest harvesting activities. A copy of the sheet, showing the information about capital costs entered for the basecase scenario, is shown in Figure 5. The calculations in this sheet are far more complicated than in the other sheets in the model and some of the data entered into this sheet was derived from other calculations. These will be explained below.
The first piece of information required about each activity, is the type of machinery usually used (entered into row 5) and its purchase price (entered into row 6). The types of machinery specified throughout this analysis are shown in Figure 5. These machinery types (or very similar alternatives) are believed to be those most commonly used in forest operations in Suriname.
Most forest machinery in Suriname is quite old so, ideally, the secondhand or resale price of such machinery should be used in the analysis as the measure of machinery cost. But, very little machinery is bought or sold each year, so it was impossible to get hold of such data. However, it was possible to get list prices and rough estimates of the price of equivalent new pieces of machinery used in each of the forestry harvesting activities and these were entered into row 6 of the sheet. In order to estimate the current value of harvesting machinery, the average age of such machinery was also entered into row 8.
The depreciation rate for each piece of machinery is the amount by which its value reduces for every year older it gets, expressed as a percentage per year. Estimates of depreciation were calculated from resale price information collected from international sources (see: Whiteman, 1999). An explanation of these calculations is given in: Section 3.2.6 Calculation of the estimated depreciation rates used in the analysis. The values calculated in this way were entered into row 7 of the model.
Figure 5 A copy of the sixth sheet of the harvesting cost model showing information about capital costs entered into the model
In the absence of secondhand machinery price information, an estimate of the current value of each piece of harvesting machinery is calculated by the model in row 9. The model calculates a depreciated replacement value for each piece of machinery, using the following formula:
Depreciated = Purchase price x (1d)a
replacement value
where d is the depreciation rate and a is the current age of the machinery. For example, in the case of the Caterpillar 950 loader, current value was calculated as follows:
Current = Depreciated = Purchase price x (1d)a
value replacement value
= US$ 235,000
(1.039)15
= US$ 235,000
(1.775)
= US$ 129,395
The investment period is an estimate of the remaining machine life, or the number of years that the machine will be used before it is eventually replaced. Given the high cost of new machinery in Suriname and the general difficulty of financing new machinery purchases, it was assumed that existing machinery would be used for quite some time (despite it already being generally very old) and a figure of 10 years was assumed for all major pieces of machinery and a remaining life of five years was assumed for chainsaws. This information is entered into row 10 of the sheet.
The residual value of machinery is its resale value at the end of the investment period. This is calculated by the model in row 11 of the sheet in the same way as the current value (explained above).
Information about external financing of machinery purchases is entered into rows 12 and 13 of the model. The amount of loan financing for a new piece of machinery, or outstanding loan obligations for an existing piece of machinery is entered into row 12 of the sheet, as a proportion of the current value of the machine. The duration of the loan period, or outstanding loan period, is entered into row 13. Information collected during this study indicated that very little forestry investment in Suriname is financed by loans, so the values in row 12 were all set to zero throughout the analysis.
Repair costs are entered into rows 14 and 15 of the sheet. The cost of repairs is entered into the sheet in terms of average cost per actual working hour. Because most parts are imported, this is split into parts (in US$/hour) and labour (in Sf/hour). The figures shown in Figure 5 were calculated from the Caterpillar Handbook and adjusted to reflect circumstances in Suriname. A full description of the simple model used to estimate repair costs is given in: Section 3.2.7 Calculation of repair costs used in the analysis.
The last piece of information entered into this sheet is the annual cost of insurance, which is entered in row 16 as a percentage of the current value of each machine. The little evidence that could be obtained suggested that a figure of 3.5% might be appropriate and this figure was used throughout the analysis.
The total capital cost of production, by activity, is shown in Sf/m^{3} and US$/m^{3}, in rows 25 and 26 of the sheet. This is calculated by adding together the four components shown in rows 21 to 24. These individual components are calculated as follows:
Financing payments (row 21) are calculated as the sum of interest payments and principle repayments on the value of the loan for each piece of machinery, divided by total production throughout the investment period. The value of the loan is calculated as the current value of the machine (row 9) multiplied by the amount of loan financing in percent (row 12). It is assumed that loan payments will be made monthly and the average monthly payment is calculated using a standard formula available in excel (based on the value of the loan, the number of monthly payments (i.e. the repayment period x 12), and the interest rate shown on sheet 2). The average monthly payment is multiplied by the number of monthly payments to give the sum of interest payments and principle repayments on the value of the loan.
Investor's capital (row 22) is calculated as the current value of the machinery less the value of any loan, all divided by total production throughout the investment period.
Residual value (row 23) is calculated as the residual value shown above (in row 11), divided by total production throughout the investment period.
Repairs and insurance (row 24) is calculated by adding together the annual repair cost and the annual insurance cost and dividing the result by total annual production. The annual repair cost is calculated as the hourly costs shown in rows 14 and 15, added together and multiplied by actual working hours per year (shown in sheet 3). The annual insurance cost is calculated as the current value of the machine (row 9) multiplied by the percentage shown in row 16.
This capital cost, however, does not include any allowance for the return on capital, which will be critically affected by the amount and timing of payments (e.g. loan payments) and the expected rate of return. Therefore, the additional allowance for return on capital (based on the expected level of profit entered into sheet 2) is also calculated in rows 31 and 32.
The additional allowance for return on capital is calculated from the NPV (at the expected level of profit) of the first three items shown above (i.e. loan payments plus investor's capital less residual value, but excluding repairs and insurance payments). This NPV is then annualised over the investment period and divided by annual production, to give the capital cost (including return on capital) per m^{3}, from which the three items in rows 21 to 23 are subtracted to get the additional allowance shown in rows 31 and 32.
Calculation of the estimated depreciation rates used in the analysis
Information about depreciation rates used in the forestry sector in Suriname could not be obtained locally, so depreciation rates were estimated from information about secondhand machinery prices in North America, collected during the cost and price survey. Resale prices were collected from a number of international sources for several major pieces of forestry machinery (see: Whiteman, 1999) and these prices were analysed using ordinary leastsquares (OLS) linear regression techniques. The following model was used in the analysis
Ln(PRICE) = a + b(AGE) + c(MODELTYPE)
where Ln(PRICE) is the natural logarithm of the resale price of each piece of machinery collected in the survey, AGE is the age of the machinery and MODELTYPE is a dummy variable set to 1, for specific model types that were believed to be more valuable than the majority of the machines in the sample (e.g. selfloading timber trucks as opposed to ordinary timber trucks). The dummy variable was only used for some of the types of machinery examined in the analysis.
An adequate amount of information for such an analysis, was collected for four types of forest machinery: timber trucks; skidders; loaders; and bulldozers. The results of the regression analyses for each of these types of machine is presented below:
Ln(TRUCK PRICE) = 11.477  0.137 (AGE) + 0.791 (SELF LOADER) n = 16
(81.68) (13.60) (3.96) R^{2 }= 93%
Ln(SKIDDER PRICE) = 11.544  0.082 (AGE) n = 40
(82.68) (7.70) R^{2 }= 60%
Ln(LOADER PRICE) = 11.150  0.039 (AGE) + 0.450 (LARGE CAPACITY) n = 25
(70.26) (4.07) (2.57) R^{2 }= 50%
Ln(BULLDOZER PRICE) = 12.221  0.069 (AGE) + 0.375 (D8 MODEL) n = 16
(70.34) (8.38) (2.71) R^{2 }= 86%
The tstatistic for each coefficient estimated by the four models is shown in parentheses under the coefficient and the number of observations (n) and adjusted rsquared value for each regression result (R^{2}) are shown to the right of each equation.
As these results show, age had a significant effect on resale price (i.e. the tstatistic on the age coefficient was outside the range ±1.96 in all cases), as did the general type of machine for sale (i.e. larger models or models with better equipment generally attracted better prices). The explanatory power of some of the equations was quite low, but this is to be expected given the range of machinery types collected in the survey and the number of different locations where machinery was being sold. For example, variants of two basic models of bulldozer (the Caterpillar D6 and D8) were covered in the survey, compared with 16 different types of skidder from eight different manufacturers. The estimated resale prices for each of these types of machinery and the goodnessoffit of the regression lines are shown in Figure 6 to Figure 9.
Figure 6 Estimated sale and resale values of timber trucks for sale in North America in 1998
Figure 7 Estimated sale and resale values of wheeled skidders for sale in North America in 1998
Figure 8 Estimated sale and resale values of wheeled loaders for sale in North America in 1998
Figure 9 Estimated sale and resale values of bulldozers for sale in North America in 1998
The results of these four models were used to estimate machinery depreciation rates in Suriname on the assumption that they would not be significantly different to the rates of machinery depreciation found in North America. This seemed reasonable, given that the types of machinery used in Suriname are mostly the same as those covered by these models. Indeed, the justification for doing this is even stronger when it is considered that much of the harvesting machinery currently used in Suriname was originally imported secondhand from North America. The only important point to note, however, is that local machinery prices seem to be much higher than the prices obtained from North America (see Table 2). Therefore, the values shown in Figure 6 to Figure 9, can not be used directly as an indication of potential resale values in Suriname. The model implicitly takes this into account by applying the depreciation rate to a local price for new machinery, which acts as a rough adjustment to local market prices.
Table 2 Comparison of new forest machinery prices in North America with the price of similar equipment available in Suriname
Type of machinery 
Manufacturer and model 
Average price (new) in North America (in US$) 
Quoted price (new) in Suriname (in US$) 
Price difference in Suriname 
Bulldozer 
Caterpillar D8 
295,000 
490,000 
+66% 
Bulldozer 
Caterpillar D6 
203,000 
280,000 
+38% 
Wheeled skidder 
Caterpillar 528B 
170,000 
210,000 
+23% 
Wheeled loader 
Caterpillar 950F 
150,000 
245,000 
+63% 
Source: Author's own estimates (North American prices) and Surmac (Suriname prices)
The estimated annual rate of depreciation can be directly read from the coefficient on age with the model specification used in this analysis (i.e. with the natural logarithm of prices on the lefthand side). For example, the AGE coefficient in the bulldozer model is 0.069, which equals a depreciation rate of 6.9% per year. Therefore, these figures were entered directly into sheet six of the model. No information about depreciation of chainsaws or barges could be obtained from any source, so the information used in the model is the author's best estimate of likely depreciation in comparison with the rates established for the other types of machinery.
Calculation of repair costs used in the analysis
The other area where a separate analysis was required to estimate figures to put into the forest harvesting cost model, was the cost of repairs to forest machinery. Only one forest manager had adequate information about repair costs in Suriname and he admitted that the budget he had for repairs was not really sufficient to cover all his needs. Therefore, this data could not really be used in the analysis. However, the Caterpillar Handbook contains detailed information about average repair costs for most of the major pieces of forest machinery used in Suriname and this was adjusted to produce a repair cost that reflects local circumstances.
The first task in this analysis, was to adjust the data presented in the Caterpillar Handbook to reflect local circumstances (i.e. costs of labour and spare parts). This was done using a simple spreadsheet model and a copy of the data input sheet of this model is shown in Figure 10. The Caterpillar Handbook gives average repair costs (in US$ per actual working hour) for six types of machinery commonly used in forestry operations and gives low, medium and high estimates of repair cost (which depend upon site circumstances and other operating conditions such as the skill of the machine operator). This information is entered into cells D10 to F15 of the data input sheet of this model. The Caterpillar Handbook also specifies the proportion of the total cost that is estimated to be the cost of parts and the proportion that is labour cost. The labour cost proportion is entered into cells G10 to G15 for each machine.
The costs presented in the Caterpillar Handbook are based on an average labour cost of US$ 50/hour and parts are costed at US list prices. These costs have to be adjusted to the typical levels of prices in Suriname and this is done by entering the relative cost of parts and labour in Suriname in cells I10 and I11. Assuming that the difference in the price of parts in Suriname and the USA is roughly the same as the difference between new machinery prices in North America and Suriname, the price of parts in Suriname was increased to 150% of the US list prices for parts. In terms of the labour cost, it was assumed that the labour cost of repairs in Suriname would be about 25% of the cost in the USA (i.e. 25% x US$ 50/hour = US$ 12.50/hour). This information is then used to calculate the parts and labour cost of repairs in Suriname, shown in cells D20 to I25. The parts cost is expressed in US$ per actual working hour and the labour cost is expressed in Sf per actual working hour (calculated using the exchange rate entered into cell D26).
Figure 10 A copy of the spreadsheet model used to adjust repair costs given in the Caterpillar handbook to an estimate of repair costs in Suriname
The second adjustment that must be made to these figures, is to alter them to reflect the age of machinery currently being used in Suriname. The repair costs presented in the Caterpillar Handbook are based on a machine life of 10,000 working hours. However, if a machine is used for longer than this, then the average cost of repairs over the life of the machine is likely to be higher than the figures presented above. Therefore, the cost figures shown above should be increased to reflect the greater age of machinery.
The Caterpillar Handbook presents extended life multipliers that can be used to convert the given repair costs to the repair costs which might be expected if a new machine is used for a longer period of time. These were used to create a series of annual repair cost multipliers for a range of machinery ages and expected remaining lives.
The first part of this process was to estimate an average repair cost multiplier curve for a new machine over any expected life. This was done using OLS regression (see Figure 11). However, most forest machinery in Suriname is not new, but already several years old, so the average repair cost multiplier was converted into an annual repair cost multiplier for every machine age (the marginal repair cost multiplier). This was then used to calculate an average repair cost multiplier for a range of existing machinery ages and expected lives, by adding together all the annual repair cost multipliers for every possible machine age and expected machine life and dividing the total by the expected machine life (see Table 3).
Figure 11 Average and marginal repair cost multipliers for forest machinery of varying ages
Note: Machinery life shown above is based on the assumption of 700 working hours per year
The figures shown in Figure 10 and Table 3 were then used to calculate the numbers entered into the Harvesting cost model, using the medium repair cost estimates given in the Caterpillar Handbook. So, for example, the skidder in this scenario is 12 years old and is expected to work for 1,026 hours per year. This would give an approximate current age of 12,312 hours (1,026 x 12), with an expected life of 10,260 hours (1,026 x 10). From Table 3, the row for 12,000 hours current age and the column for an expected future life of 10,000 hours, gives a the repair cost multiplier of 173%. This multiplier should then be applied to the central cost estimate of US$ 4.95/hour for parts (from Figure 10), to give the adjusted repair cost of US$ 8.71/hour shown in Figure 5 (i.e. US$ 4.95/hour x 173% = US$ 8.71/hour).
As with the depreciation rate, no information was available about the repair cost of some of the other pieces of machinery (such as timber trucks, chainsaws and barges), so estimates were made based on comparison with the rates established for these types of machinery.
Table 3 Repair cost multipliers for a range of machinery ages and expected remaining lives
Current age 
Expected remaining life (depreciation period) in years 

of machine 
1 000 hrs 
2 000 hrs 
3 000 hrs 
4 000 hrs 
5 000 hrs 
6 000 hrs 
7 000 hrs 
8 000 hrs 
9 000 hrs 
10 000 hrs 
New 
78% 
80% 
82% 
84% 
87% 
89% 
91% 
94% 
96% 
99% 
1 000 hrs 
82% 
84% 
87% 
89% 
91% 
94% 
96% 
99% 
101% 
104% 
2 000 hrs 
86% 
89% 
91% 
93% 
96% 
98% 
101% 
103% 
106% 
109% 
3 000 hrs 
91% 
93% 
96% 
98% 
101% 
103% 
106% 
109% 
111% 
114% 
4 000 hrs 
96% 
98% 
100% 
103% 
106% 
108% 
111% 
114% 
117% 
120% 
5 000 hrs 
100% 
103% 
106% 
108% 
111% 
114% 
117% 
119% 
122% 
126% 
6 000 hrs 
105% 
108% 
111% 
113% 
116% 
119% 
122% 
125% 
128% 
132% 
7 000 hrs 
111% 
113% 
116% 
119% 
122% 
125% 
128% 
131% 
134% 
138% 
8 000 hrs 
116% 
119% 
122% 
125% 
128% 
131% 
134% 
137% 
141% 
144% 
9 000 hrs 
122% 
125% 
128% 
131% 
134% 
137% 
140% 
144% 
147% 
151% 
10 000 hrs 
128% 
131% 
134% 
137% 
140% 
144% 
147% 
151% 
154% 
158% 
11 000 hrs 
134% 
137% 
140% 
143% 
147% 
150% 
154% 
157% 
161% 
165% 
12 000 hrs 
140% 
143% 
147% 
150% 
154% 
157% 
161% 
165% 
169% 
173% 
13 000 hrs 
146% 
150% 
153% 
157% 
161% 
164% 
168% 
172% 
176% 
180% 
14 000 hrs 
< P ALIGN="RIGHT">153% 
157% 
160% 
164% 
168% 
172% 
176% 
180% 
184% 
188% 
15 000 hrs 
160% 
164% 
168% 
172% 
176% 
180% 
184% 
188% 
192% 
197% 
16 000 hrs 
168% 
171% 
175% 
179% 
183% 
188% 
192% 
196% 
201% 
206% 
17 000 hrs 
175% 
179% 
183% 
187% 
192% 
196% 
200% 
205% 
210% 
215% 
18 000 hrs 
183% 
187% 
191% 
196% 
200% 
205% 
209% 
214% 
219% 
224% 
19 000 hrs 
191% 
195% 
200% 
204% 
209% 
214% 
218% 
223% 
228% 
234% 
20 000 hrs 
200% 
204% 
209% 
213% 
218% 
223% 
228% 
233% 
238% 
244% 
Expected remaining life (depreciation period) in years 

11 000 hrs 
12 000 hrs 
13 000 hrs 
14 000 hrs 
15 000 hrs 
16 000 hrs 
17 000 hrs 
18 000 hrs 
19 000 hrs 
20 000 hrs 

New 
101% 
104% 
107% 
110% 
113% 
116% 
119% 
122% 
125% 
128% 
1 000 hrs 
106% 
109% 
112% 
115% 
118% 
121% 
124% 
128% 
131% 
135% 
2 000 hrs 
112% 
115% 
118% 
121% 
124% 
127% 
130% 
134% 
137% 
141% 
3 000 hrs 
117% 
120% 
123% 
127% 
130% 
133% 
137% 
140% 
144% 
147% 
4 000 hrs 
123% 
126% 
129% 
133% 
136% 
139% 
143% 
147% 
150% 
154% 
5 000 hrs 
129% 
132% 
135% 
139% 
142% 
146% 
150% 
153% 
157% 
161% 
6 000 hrs 
135% 
138% 
142% 
145% 
149% 
153% 
157% 
160% 
165% 
169% 
7 000 hrs 
141% 
145% 
148% 
152% 
156% 
160% 
164% 
168% 
172% 
176% 
8 000 hrs 
148% 
151% 
155% 
159% 
163% 
167% 
171% 
175% 
180% 
184% 
9 000 hrs 
155% 
158% 
162% 
166% 
170% 
175% 
179% 
183% 
188% 
192% 
10 000 hrs 
162% 
166% 
170% 
174% 
178% 
182% 
187% 
191% 
196% 
201% 
11 000 hrs 
169% 
173% 
177% 
182% 
186% 
191% 
195% 
200% 
205% 
210% 
12 000 hrs 
177% 
181% 
185% 
190% 
194% 
199% 
204% 
209% 
214% 
219% 
13 000 hrs 
185% 
189% 
194% 
198% 
203% 
208% 
213% 
218% 
223% 
228% 
14 000 hrs 
193% 
197% 
202% 
207% 
212% 
217% 
222% 
227% 
233% 
238% 
15 000 hrs 
201% 
206% 
211% 
216% 
221% 
226% 
232% 
237% 
243% 
248% 
16 000 hrs 
210% 
215% 
220% 
225% 
231% 
236% 
241% 
247% 
253% 
259% 
17 000 hrs 
219% 
225% 
230% 
235% 
240% 
246% 
252% 
258% 
264% 
270% 
18 000 hrs 
229% 
234% 
240% 
245% 
251% 
257% 
263% 
269% 
275% 
281% 
19 000 hrs 
239% 
244% 
250% 
256% 
261% 
267% 
274% 
280% 
286% 
293% 
20 000 hrs 
249% 
255% 
260% 
266% 
272% 
279% 
285% 
292% 
298% 
305% 
Note: The current age and expected remaining life of machinery is expressed in actual hours worked
Production cost is given in Sf/m^{3} in rows 9 to 16 and US$/m^{3} in rows 20 to 27 and the cost of each individual harvesting activity within the whole operation is shown in columns D to K. The first three rows (i.e. rows 9 to 11 and 20 to 22) of each of these two blocks of results show the labour, consumables and capital components of production cost (copied from sheets four to six in the model). These are added together in the row below (rows 12 and 23).
These figures do not allow for any return on capital or profit on other expenditure. The fifth row of each block (i.e. rows 13 and 24) shows the additional amount that would be required to give the forest manager the rate of return on capital specified in the second sheet of the model (in cell H5). This is added to the total (in rows 13 and 24) to give the total production cost including an allowance for return on capital (TOTAL including ROC) in rows 14 and 25.
The seventh row of each block calculates the markup or allowance for profit (from cell H5 in the second sheet) on other expenditure. This alternative way of including an allowance for normal profit in the calculation, is added to the totals in rows 12 and 23, to give the total roundwood cost including an allowance for a return on other expenditure (TOTAL including ROE) in rows 16 and 27. This markup is applied to all labour and consumable costs, plus the repair and insurance components of capital costs (which are not included in the calculation of return on capital).
Figure 12 A copy of the first sheet of the harvesting cost model showing the total roundwood production cost estimated for the basecase scenario
Discussion of results
For the purposes of updating this information, it is important to know which variables are likely to have the most impact on the total roundwood production cost if they were to change. It is also useful to know which forestry activities account for most of the costs, so that the costeffectiveness or efficiency of these activities can be examined in greater detail.
Figure 13 shows the distribution of costs for the "typical" forest concession between labour, consumables, capital and profit. This shows that the total cost of production is fairly evenly split between these four items. Figure 14 shows how changes in the cost of individual items of expenditure would affect the total production cost. For example, a 25% change in labour costs would alter the total roundwood production cost by just over 5%, but a 25% change in capital costs would change the total cost by nearly 10%. Or, to put it another way, a change in capital costs would have roughly twice the effect of the same percentage change in labour costs.
Figure 13 The distribution of total delivered roundwood production cost, between labour, consumables, capital and normal profit (20% return on capital)
Figure 14 The impact of changes in individual cost items on total roundwood production cost
Figure 15 The effect of different exchange rates on total roundwood production costs in Sf and US$
Given that most capital equipment used in the forestry sector in Suriname is imported and, consequently, paid for in US$, the results shown in Figure 14 would suggest that changes in the exchange rate might also have a significant impact on total production costs. Figure 15 therefore, shows the impact of different exchange rates on the total roundwood production cost in local and foreign currency.
As the figure shows, as the exchange rate increases (i.e. the Suriname Guilder devalues), the cost of production in local currency increases (due to the higher cost of imported items used in production), while the production cost in US$ goes down (because local currency expenditure on inputs such as labour, becomes relatively cheaper in US$). However, due to the relative cost of items purchased in local and foreign currency, these changes are not exac t opposites. For every 10% increase in the exchange rate, the production cost in local currency increases by 5%. But, as the exchange rate increases, the production cost in US$ falls rapidly at first, then starts to levelout at about US$ 20/m^{3} when the exchange rate goes above Sf 1,000 to the US$. This reflects the fact that, at this exchange rate level, the cost of items purchased in local currency starts to become insignificant compared with the cost of items purchased in US$.
Figure 16 shows the distribution of total roundwood production cost across the four main forestry activities. This shows that the costs of skidding and roundwood transport account for the largest share of total production cost. This is largely due to the high cost of machinery and consumables (such as fuel) that are used in these activities. In comparison, the cost of felling is relatively insignificant.
The cost of loading calculated here only accounts for about 18% of total roundwood production cost. However, under the assumptions made about forest productivity in this calculation, the loader is fairly well utilised, which keeps the unit cost of this activity down to a relatively low level. In smaller forest operations, it may not be possible to fully employ a forest loader and costs may be consequently higher, or it may be more profitable to contractout this operation. Loading costs are also likely to be much higher in situations where timber has to be unloaded and reloaded onto barges or at the logyard.
Figure 16 The distribution of total delivered roundwood production cost, between forestry activities
Figure 17 The effect of transport distance on total roundwood production cost
Figure 17 shows the effect of transport distance on total roundwood production cost. The points in the figure are the results of the model calculation for a number of different transport distances and the line is a trend line between these points. The points do not follow a straight line because transport cost is a function of the number of vehicles required as well as the cost of running the vehicles (i.e. the running cost per km travelled). Thus, for example, the first three points require one, two and three timber trucks respectively to transport timber over these distances, but at the distance represented by the fourth point, three trucks will suffice and it is not necessary to use a fourth timber truck. This is why the fourth point appears below the trend line. At greater distances, it may be more economical to use water transport to transport roundwood, although this may also then require additional loading and reloading costs.
The effect of changing the level of normal profit on the roundwood production cost
As noted above, the cost of roundwood production in the model changes if it is assumed that forest managers require a different level of profits to remain in business. Figure 18 shows how the roundwood production cost varies with the level of profit used in the model. The roundwood production cost changes by roughly 7% for every change of 5% in the amount of profit forest managers are allowed to retain.
Figure 18 The effect on roundwood production cost of changing the assumption about the amount of profit forest managers are allowed to keep (the level of normal profit)
The crucial question that is intimately linked with the calculation of economic rent and setting of forest levies, is: how much profit should forest managers be allowed to keep (i.e. what should be the normal level of profit)? A level of 20% has been used here, based on discussions with forest managers and staff of LBB and SBB. This figure is high compared with the returns to forestry in many temperate countries, but is broadly comparable with the returns expected in other tropical countries. In part, this rate is high because of the greater risks associated with conducting business in tropical countries, but it also reflects the generally high returns that can be obtained from other investments in such fast growing economies.
Comparison of the model results with contract costs collected during the study
Another pointed noted earlier was that one sign that there might be excess profits or rent capture by private individuals operating in the forestry sector is if the contract cost of operations is higher than calculated by the model. In other words, if the government were collecting all of the economic rent from roundwood production, there would be very little scope for contractors to charge high rates for their services that allowed them to operate inefficiently or earn excess profits.
Table 4 Current costs of contractingout various forest operations compared with the costs estimated by the roundwood production cost model
Operation 
Model results (costs per m^{3}) 
Contract costs (per m^{3}) 

In local currency 
In US$ 
Roundwood for the domestic market 
Roundwood for the export market 

Felling (with own chainsaw) 
Sf 1,300 
US$ 2.00 
Sf 2,000 
n.a. 
Skidding 
Sf 6,800 
US$ 10.50 
Sf 8,000 
n.a. 
Loading 
Sf 3,100 
US$ 4.70 
Sf 1,500  Sf 3,250 
US$ 5.00 
Unloading 
Sf 3,100 
US$ 4.70 
Sf 2,000  Sf 4,000 
US$ 10.00 
Road transport 
Sf 70/m^{3}/km 
US$ 0.11/m^{3}/km 
Sf 50/m^{3}/km  Sf 180/m^{3}/km 
n.a. 
Source: Model results and Whiteman (1999)
A comparison between the current costs of contractingout various forest operations and the costs estimated by the roundwood production cost model is shown in Table 4. In most cases the costs calculated by the model are at the lower end of the range of costs quoted during interviews with forest managers and contractors. This could indicate one or more of three things: that the model assumptions are incorrect in some way; that operators are not working efficiently; or that operators are earning excess profits.
The production costs calculated by the model are comparable with those calculated by the author in other countries, suggesting that contractors may be operating inefficiently or earning excess profits. Another indication that they may be earning excess profits is that the costs quoted for handling export quality roundwood are somewhat higher than the costs quoted for handling other roundwood. There is no reason why the costs for basically the same type of operation should vary in this way, which suggests that contractors are charging higher prices for handling export quality roundwood in order to capture some of the economic rent from the production of this roundwood. They are, presumably, able to do this because of the limited amounts of capital available to many forest managers. In other words, they can charge a rent above their production cost because they own a piece of scarce equipment (i.e. forest machinery).