5 ECONOMIC RENT ANALYSIS

This section describes the data collected about production costs and product prices in Fiji and the methodology used to calculate economic rent. It finishes by presenting the results of the economic rent analysis and describing the implications of these results for the level of forest charges.

5.1 Data collection

In order to calculate the level of economic rent from roundwood production in Fiji, it was necessary to collect detailed information about the costs of forest operations. In addition, because there is no log export market and it is difficult to obtain reliable information about domestic log prices, the costs of forest processing were also estimated in order to subtract all of these costs from forest product prices to arrive at the economic rent.

5.1.1 Roundwood production and processing costs

Production costs were calculated “from the bottom up”. In other words, information about the costs of individual factors of production (e.g. labour, fuel, tools and machines) was collected and used, along with information collected about the consumption of materials and productivity rates, in order to calculate the unit costs of production for each activity (e.g. felling, road transport, kiln drying). A simplified picture of the process used to estimate the unit costs of production is given in Figure 21.

Figure 21 A simplified picture of the process used to collect cost data and produce unit costs of production for various forestry and forest processing activities in Fiji

The above information was collected from a variety of sources from both within Fiji and abroad.¹³ The local information was collected as part of the two-week visit in October 2003, during interviews with Forestry Department staff and local producers. In total, interviews were held with 11 individuals from the forest industry, three Forestry Department staff, three landowners’ representatives and a representative of the Fiji Sawmillers Association. In addition, eight forest processing facilities were visited and one forest harvesting site.

Information on wage rates and import tariffs was obtained from official statistics from the Bureau of Statistics (2000), Fiji Trade and Investment Bureau (FTIB, 2003) and US Department of Commerce (USDC, 2000). Other documents consulted included a submission on forest charges from the Fiji Sawmillers Association (FSA, 2003) and calculations by the NLTB (2003). Discrepancies between these documents, the interviews with individuals from the forest industry and the economic rent calculations were discussed with Forestry Department and, where appropriate, the calculations presented here were modified accordingly.

It should be noted that, throughout this process, figures were constantly checked, re-checked and updated after discussions with different stakeholders and taking into account experiences from other countries. Although this is never a precise process, it is believed that the figures presented here present a reasonable and unbiased picture of the cost and price structure of forest operations in Fiji.

Other types of information from international data sources were also used to check and validate the statements given during interviews. For example, information about materials consumption and productivity rates was collected from reference sources such as machine operator’s handbooks and compared to statements from the industry. For a variety of reasons, the level of productivity and capital utilisation in Fiji is less than the maximum that could possibly be achieved, but it is probably higher than most other countries at a similar level of development. The analysis presented here is based on current harvesting methods and levels of productivity, although there may be some room for improvement.

5.1.2 Roundwood and product prices

Price information was collected from the main sources listed above. In addition, the Forestry Department supplied sawnwood export price data by species and grade for the period 1999 to 2002.

Most forest processors in Fiji harvest roundwood from their own licence areas, so it was not possible to gather much information about market prices for delivered roundwood. Table 10 presents the range of figures obtained from the data sources, along with estimates of product recovery rates. The individual number of observations is very small, so these figures should be treated with a great deal of caution. However, although the range for Class 1 species is quite wide, there was a remarkable convergence between the different data sources about the prices paid for roundwood in the other classes.

Table 10 Current (2003) prices for delivered (mill-gate) roundwood in Fiji

Royalty class	Mill-gate price (in FJD per m3)		Product recovery rate (in percent)
	Low	High	Low	High
Class 1 species	210	280	50	55
Class 2 species	130	150	48	51
Class 3 species	110	120	48
Class 4 species	90		46

Source: Field survey.

It is also interesting to note that the quoted mill-gate prices are much higher than the delivered roundwood production cost (including royalties and all other charges) calculated below. Where the majority of a commodity is processed in an integrated operation, is not unusual for “spot” prices to be much higher than production costs. However, these figures demonstrate the importance to forest processors of having a stable and secure source of supply. If all roundwood was harvested by independent loggers, the mill-gate price would probably be much lower, as it is unlikely that the industry could pay these prices for all of their raw materials. This divergence between mill-gate prices and delivered roundwood costs has profound implications for the Government’s policy towards awarding harvesting licences.

Table 11 presents information about sawnwood prices by royalty class and grade and the average grade distribution. This information was obtained from the Fiji Sawmillers Association (FSA, 2003) and NLTB (2003). These figures are quite surprising, in that the FSA are quoting higher product prices than NLTB. The figures from the FSA also seem to contradict some of the other statements in their submission (e.g. where they highlight the fact that export prices for some Class 1 species have fallen dramatically in 2003 due to competition from other countries in major export markets).

Table 11 Current (2003) sawnwood prices in Fiji

Data source and grade	Grade distribution	Price by royalty class (in FJD per cubic metre)
		Class 1	Class 2	Class 3	Class 4
FSA: FF Select and better	70%	1,200 [export]	900 [export]	480 [wholesale]	480 [wholesale]
FSA: FF Standard	25%	550 [wholesale]	480 [wholesale]	480 [wholesale]	480 [wholesale]
FSA: Commons	5%	250 [wholesale]	250 [wholesale]	250 [wholesale]	250 [wholesale]
FSA: Average of above	100%	990	763	469	469
NLTB: Average of all grades	n.a.	700	500	400	320

Source: FSA (2003) and NLTB (2003).

Table 12 Average sawnwood export prices 1999 - 2002 (at 2003 prices)

Species	Average price by grade (FJD per cubic metre at 2003 prices, FOB)
	Select and better	Standard	Ungraded	All grades
Buabua	1,047	948	704	954
Dakua Makadre	1,053	n.a.	1,019	1,043
Dakua Salusalu	1,020	n.a.	1,220	1,105
Rosawa	874	727	773	825
Vesi	872	n.a.	849	862
Yaka	1,188	975	1,092	1,103
Average for Class 1 species	1,009	883	943	1,007
Bauvudi	803	n.a.	771	790
Damanu	792	865	673	771
Kaudamu	837	934	686	861
Kauvula	756	798	796	785
Average for Class 2 species	797	866	732	805
Rosarosa	1,407	509	570	749
Sacau	1,143	704	917	965
Yasiyasi	1,206	704	628	791
Average for Class 3 species	1,252	639	705	845

Source: Forestry Department.

Table 12 presents information about average sawnwood export prices by species and grade over the period 1999 to 2002 (more detailed statistics are available in Table 32 in Appendix 1). These figures have all been updated to 2003 prices using the GDP deflator. Unfortunately, the overall amount of exports from Fiji is so small that individual export consignments of unusually high or low value confound any attempt to draw conclusions from these figures about trends in prices or the relative value of different species and sawnwood product grades. It is, therefore, extremely difficult to identify any clear patterns in these figures or to comment on whether individual species should be moved from one royalty class to another. However, the general levels of prices recorded in these statistics would seem to confirm that the export prices given in the FSA submission seem a little on the high side.

5.2 Calculation methodology

The economic rent from forest harvesting was calculated using two simple spreadsheet models. Copies of these models, along with their associated manuals, were left with the Forestry Department and can also be found at: www.fao.org/forestry/finance.

First, the total delivered roundwood production cost was calculated using a model developed for an FAO Project in Suriname (see Whiteman, 1999a for further details of the roundwood production cost model used in this analysis). The model uses standard forestry costing techniques, such as those described in FAO (1977). The scale of harvesting operations in Fiji is similar to Suriname, so it was very easy to input the data collected in Fiji into the existing modelling framework. An example of the data used in the model and the outputs from the model is given in Appendix 2.

The results from this model were then entered into a forest industry cost model. This model was first developed for a DFID Project in Indonesia, but has been subsequently used in a number of other countries. Again, the model follows internationally accepted investment appraisal and accounting procedures used to measure profit and loss, deprecation and tax (see Scotland and Whiteman (1997), for details of how to use the model). Although the scale of operations in Fiji is relatively small compared with Indonesia, it was quite easy to input the data from Fiji into the model and to use it to examine the profitability of forest processing operations. An example of the data used in this model and the outputs produced by the model is given in Appendix 3.

It should be noted that this analysis uses a discounted cash-flow methodology to calculate the profitability of investments in the forestry sector (i.e. the ROC), or the level of “normal profit” required by processors and the residual economic rent from production. The submissions from the FSA and NLTB used the simpler approach of a required profit margin (or mark-up on non-capital operational expenditure) as a measure of the “normal profit” required by processors. While this is much simpler, it is not a very accurate or reliable indicator of the level of profits required to retain investment (or encourage new investment) in the sector. For harvesting operations, the two approaches give quite similar results.¹⁴ However, for the analysis of investments in the processing industry, the amount of profit required for a 20 percent ROC will be very different (i.e. much higher) than a 20 percent profit margin, because of the capital intensive nature of production. This explains why a rate of 20 percent ROC may seem a little low in comparison with what forest processors might expect as an acceptable profit margin.

5.3 Delivered roundwood production cost

As already noted, roundwood production costs can vary due to a range of factors outside and within the forest managers' control. Therefore, as a starting point for this analysis, the production cost for a "typical" or "representative" forest operation was constructed. A sensitivity analysis then examined variations in the total roundwood production cost, due to variations in the two factors that are likely to have the greatest impact on costs: transport distance and harvesting intensity.

5.3.1 The total roundwood production cost for a "typical" forest operation

The four main factors that influence roundwood production costs are: the area being cut each year (i.e. scale of operations); harvesting intensity (and, thus, annual production from the area); harvesting, extraction and transport methods currently used to get timber from the forest to the mill; and the average transport distance. Information about each of these variables was collected during the field visits and discussions with forest managers.

Cutting area. The majority of forest operations in Fiji are small-scale logging operations managed by small to medium-sized sawmills under Annual Licences. Discussions with the industry and the Forestry Department revealed that the average licence area includes about 500 acres (or 200 ha) of forest. Therefore, an annual cutting area of 200 ha was used in the analysis as representative of a "typical" forest operation currently working in Fiji.

Harvesting intensity. Very little information is available about the sustainable yield from the natural forest in Fiji. According to the Forestry Department, first harvests may produce as much as 30 m³/ha or more. Following this, it is believed that a cutting cycle of 20 years gives sufficient time for regeneration. However, the Forestry Department did express some concern that, after a second harvest, it may take a much longer time for the forest to reach a commercially acceptable stocking level. Assuming an annual harvest area of one-twentieth of the total productive area, the figures presented for Fiji’s long-established forest concessions in Table 24 suggest an average harvesting intensity in secondary forest of 10 m³/ha to 15 m³/ha. The transition from harvesting in primary forest to secondary forest is only gradually starting to take place in Fiji, so the higher figure of 30 m³/ha was used in this analysis. However, it must be recognised that this figure will fall in the future. By multiplying this figure by 200 ha, this would suggest that the current level of production in a "typical" forest operation in Fiji might be around 3,000 m³/year. This figure would also be in line with the log input of most of the medium-sized sawmills in the country.

Production techniques. Harvesting in Fiji is quite expensive due to the terrain and the fact that logging only takes place for part of the year in most operations. In terms of their impact on production costs, the three most important aspects of current harvesting practices are the amount of road building required, skidding distances in the forest and the quality and age of equipment used. Most forest harvesting areas in Fiji require some road building in the forest and usually require the construction of access roads as well. Based on discussions with the Forestry Department and interviews with producers, it was assumed that an average of 2 km of access road would be built in a typical harvesting area, plus another 2 km of road in the forest, giving a roading density of 10 m/ha in the forest. From the roading density, it is possible to calculate the average skidding distance, assuming an indirectness factor (in this case, this was assumed to be 20 percent). This resulted in an average skidding distance of 300 m, which was confirmed during the field visit and interviews. This is probably close to the optimal roading density in forests with a stocking level of 30 m³/ha. In terms of the equipment used in forest operations, most operators use bulldozers for road construction and skidding and medium-sized logging trucks. Most machinery is quite old, so the depreciated value of equipment is quite low but repair costs are relatively high.

Transport distances. Fiji is small, so most harvesting operations are quite close to mills. Discussions with sawmillers revealed that most of them can make one to two round trips from the forest to the mill in a day, suggesting a haulage distance of 50 km each way.

Table 13 shows the estimated cost of harvesting, extraction, loading and transport for a “typical” forest operation in Fiji, calculated by the model. Added to this was another small cost of FJD 2.00/m³ to cover the cost of planning and surveying, to give a total delivered roundwood production cost (excluding forest charges) of FJD 64.24/m³ (USD 32.12/m³). This cost is broadly comparable with the total delivered roundwood production cost in many other countries with moist tropical forests, but is relatively high in view of the quite short transport distance. In part, this is due to the small scale of operations, quite low harvesting intensity and low level of machinery utilisation.

Table 13 Estimated cost of harvesting, extraction, loading and transport for a “typical” forest operation in Fiji

Type of cost	Cost by activity (in FJD per cubic metre)
	Felling	Skidding with dozer	Skidding with skidder	Loading	Road transport	Unloading & reloading	Water transport	Road building	Total
Labour	3.57	2.50	NA	1.19	3.57	2.38	NA	1.07	14.26
Consumables	2.47	3.93	NA	1.04	6.96	2.07	NA	4.73	21.19
Capital	0.12	4.30	NA	1.14	4.11	2.29	NA	2.62	14.59
Total (excluding profit)	6.15	10.72	NA	3.37	14.64	6.74	NA	8.42	50.04
Return on capital	0.07	2.50	NA	1.86	2.99	3.71	NA	1.07	12.19
Total (including ROC)	6.22	13.22	NA	5.22	17.63	10.45	NA	9.49	62.24

Figure 22 Total delivered roundwood production cost in Fiji (by activity and type of cost)

Figure 22 shows the composition of total delivered roundwood cost by production activity and type of cost. The figure shows that felling and extraction to the roadside accounts for about half of the total cost and loading, transport and unloading accounts for the other half. The largest individual component of production cost is road transport, which accounts for 27 percent of the total cost (FJD 17.63/m³). Skidding is the next largest component, accounting for 21 percent of the total (FJD 13.22/m³).

Labour costs account for about 25 percent of the total cost (FJD 16.26/m³). This is relatively high compared with many other tropical developing countries and reflects the comparatively high labour costs in Fiji. Consumables (fuel, tools and spare parts) account for the largest share of production costs at 33 percent (FJD 21.19/m³). This is quite high, but reflects the age of machinery (which requires a lot of expenditure on maintenance and repairs). Conversely, the cost of capital is quite low (FJD 14.59/m³ or 23 percent of the total). The cost of capital is the cost of depreciation of the machinery used in forest operations. As much of the machinery used in Fiji is quite old, it is already heavily depreciated and further depreciation costs are therefore quite low. The allowance for normal profit is the amount required to earn a 20 percent ROC on the money invested in machinery. This is also relatively low due to the age (and value) of machinery used in operations, accounting for about 19 percent of the total cost (FJD 12.19/m³).

5.3.2 The impact of transport distance and harvesting intensity on the total roundwood production cost

The impact of transport distance and harvesting intensity on the total delivered roundwood production cost in Fiji is shown in Figure 23. The figures presented here have been calculated by varying the transport distance and harvesting intensity in the roundwood production cost model, while keeping the other variables the same except the cutting area, which has been increased to maintain a production level of 6,000 m³ per year.

Figure 23 Impact of transport distance and harvesting intensity on the total delivered roundwood production cost in Fiji

The slope of the lines in Figure 23 shows that, as transport distances increase, the total delivered roundwood production cost becomes greater at a rate of roughly FJD 0.17/m³ (or USD 0.08/m³) per kilometre. Transport costs increase with distance not only because more fuel is consumed transporting roundwood over longer distances but also because, as distances extend, more trucks have to be used to keep-up with production. Only one truck would be required to deliver the volume of timber produced by the "typical" forest operation up to a distance of around 60 km. Beyond this point, overtime or contractors would have to be used to supplement the forest managers equipment up until a transport distance of around 100 km, at which point purchasing a second truck could probably be justified. Another option would be to stockpile logs in the forest and transport them all year round. The model assumes that harvesting only takes place for eight months of the year. However, it is likely that loading and transporting logs could take place for longer and a number of operators said that they stockpiled logs in the forest in order to maintain an even flow of production. For this reason, the cost of a secondary loading and unloading is also incorporated into the calculations.

The difference between the curves in Figure 23 shows the impact of harvesting intensity on production cost. As harvesting intensity falls, production cost increases exponentially. Thus, with a 50 km transport distance, a fall in the harvesting intensity from 30 m³/ha to 25 m³/ha increases the production cost by FJD 4.20/m³. A fall from 25 m³/ha to 20 m³/ha would increase the production cost by FJD 6.60/m³ and a fall from 15 m³/ha to 10 m³/ha would increase the production cost by FJD 15.40/m³

A lower harvesting intensity increases the production cost because the costs of road and skid trail construction are spread over lower production volumes as harvesting intensity falls. One bulldozer would be almost fully employed on road construction and skidding in the “typical” forest operation. If harvesting intensity falls and the cutting area is increased to compensate for this, more bulldozer time would be required to construct roads and skid trails in the larger area (i.e. with a harvesting intensity of only 15 m³/ha, the cutting area would have to be twice as large to maintain the same level of total production and two bulldozers would be required).

This analysis has assumed that the cutting area each year can be increased in order to maintain production levels. If the cutting area was kept constant, the different curves would be even further apart (i.e. production cost would increase more for each reduction in harvesting intensity). This is because, in addition to the reason already noted above, the utilisation of the bulldozer would fall (i.e. the bulldozer would be used gradually less and less for skidding operations and would be idle for more of the time).

The above analysis suggests that production costs might increase significantly as harvesting moves from well-stocked primary forest to secondary forest areas. However, this will depend upon how much road building would be required on re-entry to a previously harvested area. For example, if harvesting in secondary forest did not require any new road building, a harvesting intensity of only 15 m³/ha would result in a total production cost that is roughly the same as harvesting at an intensity of 30 m³/ha in primary forest where road building is required.

5.4 Processing costs and profitability

In the forest industry model, the delivered roundwood cost is added to the cost of all forest charges, plus other fixed and variable operating costs, to arrive at a total operating cost at the current rate of capacity utilisation. This is subtracted from the value of product sales in the model to give net income. These figures are then projected forwards over the expected life of the investment (in this case, assumed to be 20 years) to form part of the projected cash-flow from the investment.

Capital costs include the cost of all plant and machinery, replacement capital and working capital (e.g. stocks of products, raw materials and accounts due or receivable) and these are also entered into the cash-flow. Adjustments are made to the cash-flow to take into account allowances for depreciation and the calculation of interest payments and loan repayments (if applicable) and corporation tax payments.

The model calculates the net present value (or NPV) of the investment at three different discount rates (i.e. three different levels of ROC). As in the forest harvesting cost model, a central discount rate of 20 percent was chosen as the required ROC or level of “normal profit” expected from the investment. If the NPV from the investment is positive at the 20 percent discount rate, this would indicate that the rate of return on the investment is more than 20 percent (i.e. the required ROC) and that “excess profit” is being earned in the processing operation.

The model presents the NPV of the investment in three different ways:

• before financing and tax (i.e. this is a measure of the overall profitability of the investment);

• after financing but before tax (i.e. this is a measure of the return on the investors equity in the investment); and

• after financing and tax (i.e. this is the investors return after taxes have been paid).

For the purpose of this analysis, it is the last of these measures that is most important, as this takes into account the effect of borrowing and taxation on the investor’s profits.

Using all of these calculations, the model also expresses the results in terms of the shares of product value that can be attributed to production costs (including capital costs and “normal profit”), fees and charges, taxes and “excess profit” (i.e. the NPV of the investment divided by the amount of production). Income and expenditure varies throughout the life of the investment due to requirements to replace machinery, so these figures represent the average shares of the product value over the life of the investment rather than the shares in any particular year.

5.4.1 Description of an “average” sawmill

The forest industry model is flexible and can be used to analyse the profitability of any type of investment in new or existing processing facilities in a country. However, as most indigenous roundwood in Fiji is converted into sawnwood, the calculations were restricted to analysing the costs of sawmilling. In addition, the analysis only examined the profitability of an existing sawmill rather than a new investment in the sector.

As in the case of forest harvesting, there is variation in the management, location and scale of operations in the processing sector in Fiji. However, for the purpose of this analysis, it was assumed that an “average” sawmill has the capacity to produce about 5,000 m³/year, but is currently only operating at about 60 percent capacity (i.e. with an output of about 3,000 m³/year). Most sawmills claim to be running at well-below full capacity and field visits confirmed that they have the capacity to produce more than they are, so this level of capacity utilisation seems appropriate. This level of production is representative of the medium-sized sawmills operating in the sector, which seems to be the most durable part of the forest processing industry.

The model used a product recovery rate of 50 percent across all product types, which is a reasonable working assumption. This gives a required log input of 6,000 m³/year, which matches the output level assumed in the forest harvesting model. It was also assumed that all log inputs are obtained from the sawmiller’s own licence area.

The model was constructed for a sawmilling operation on Viti Levu, but there is little reason to expect sawmilling costs to differ significantly between the two main islands. The only difference incorporated into the model due to this was the payment of premiums to NLTB.

The model assumes that the sawmill produces a range of products by grade and species group, with half of production for the export market and half for the domestic market (see below). It also assumes that 20 percent of the capital invested in the sawmill is financed by borrowing (at a real interest rate of 12 percent), with the remaining 80 percent financed by the sawmiller’s own equity. If investment in a sawmill can earn an ROC of 20 percent or more, it is generally in the sawmiller’s financial interest to borrow as much as possible if the interest rate is lower than this. However, interviews revealed that commercial lenders are wary of over-exposing themselves to the forest processing sector. Therefore, an assumption of a debt-equity ratio of 20:80 seems reasonable.

5.4.2 Operating costs

The total operating cost of a sawmill comprises the log input cost plus the fixed and variable processing costs such as fuel, labour, and consumables. The total log input cost calculated in the forest industry model is shown in Table 14. The first set of costs is the delivered roundwood production costs and these are taken directly from the results of the forest harvesting model (described above).

The second set of costs in the model is the forest charges paid to landowners and government. Payments made directly to landowners are the first component of these charges and have been calculated on the basis of average “commission” and “goodwill” payments to landowners, determined in interviews with the industry and landowners. The “commissions” paid vary according to the species group and the input mix shown here (by species class) is based on the submission on forest charges from the FSA (FSA, 2003) and discussions with sawmillers. The majority of Class 1 and Class 2 species are used for export while production for the domestic market tends to be more heavily weighted towards the lower value species classes. The second component of forest charges are the royalties and fees paid to NLTB and the last component of these charges are the fees paid to the Forestry Department for scaling and maps. Lump sum payments, such as the map fee and NLTB licence application fee, have been converted to charges per cubic metre, based on the assumed level of roundwood production of 6,000 m³/year.

The total level of all of these forest charges amounts to around FJD 50/m³ for roundwood used to produce sawnwood for the domestic market and FJD 75/m³ for roundwood used to produce export grade sawnwood. These differences are due to the different mixture of species used to produce sawnwood for these two different markets and the different levels of charges paid for the different species.

Table 14 Total log input cost for an “average” sawmill in Fiji

Log production costs (FJD per m3 of roundwood)	Sawnwood for export market		Sawnwood for domestic market
Planning and surveying		2.00			2.00
Felling		6.22			6.22
Skidding		13.22			13.22
Loading, unloading and reloading		15.67			15.67
Transport		17.63			17.63
Road building		9.49			9.49
Total		64.24			64.24
Royalties, fees and other charges (FJD per m3 of roundwood)	Sawnwood for export market		Sawnwood for domestic market
	Proportion (%)	Charge	Proportion (%)	Charge
Class 1 Landowners commission	60	25.00	20	25.00
Class 2 Landowners commission	40	20.00	10	20.00
Class 3 Landowners commission	0	15.00	50	15.00
Class 4 Landowners commission	0	10.00	20	10.00
Landowners goodwill		5.00		5.00
NLTB Licence application fees		0.20		0.20
Class 1 Royalty	60	40.00	20	40.00
Class 2 Royalty	40	30.00	10	30.00
Class 3 Royalty	0	10.00	50	10.00
Class 4 Royalty	0	6.50	20	6.50
Premium		6.00		6.00
FD Scaling fee		3.50		3.50
FD Map		0.03		0.03
Total		73.73		48.53
Other roundwood variables
Total log cost from licence area (FJD per m3)		137.97		112.77
Proportion of logs from licence area (%)		100		100
Cost of logs from elsewhere (FJD per m3)		150.00		130.00
Conversion factor (m3 logs per m3 of product)		2.00		2.00
First year of production		1		1
Last year of production		20		20
Total log cost (FJD per m3 of product)		275.94		225.54

The last set of figures in Table 14 is used to account for purchases of logs from elsewhere (not used in this case) and the product recovery rate or conversion factor. With an average conversion factor of two to one, the total log input cost is simply the delivered roundwood cost plus total forest charges, all multiplied by two. This amounts to about FJD 225/m³ for sawnwood sold in the domestic market and FJD 275/m³ for exported sawnwood.

The other fixed and variable operating costs used in the model are shown in Table 15. Estimates of these costs were obtained from discussions with sawmillers, site visits and from the FSA submission. For example, labour costs were based on discussions about the numbers of employees in the mills visited, average wage rates and social costs and the cost of specialised labour such as sawdoctors. Differences in these costs can be attributed to differences in treatment, drying and packaging used when producing sawnwood for the domestic and export markets. However, the log cost accounts for over half of all variable costs and most of the difference in production costs between sawnwood destined for the two different markets.

Table 15 Fixed and variable operating costs for an “average” sawmill in Fiji

Variable costs (FJD per m3 of production)	Sawnwood for export market		Sawnwood for domestic market
Log cost		275.94		225.54
Labour		78.96		78.96
Resins/chemicals/treatment		15.00		15.00
Fuel, energy and consumables		60.00		45.00
Packaging		32.00		6.00
Delivery		15.00		15.00
Total		476.90		385.50
Fixed costs (FJD per m3 of capacity)
Personnel: administration		12.00		12.00
Personnel: support staff		7.00		7.00
Personnel: maintenance staff		9.73		9.73
Sales cost		15.00		15.00
General overhead		30.00		30.00
Total		73.73		73.73
Total cost (FJD per m3 of production)
Total calculated at final level of utilisation		599.79		508.39

The last line in the table shows the total operating cost per cubic metre of sawnwood produced, which is estimated at around FJD 600/m³ for export grade sawnwood and FJD 510/m³ for domestic sawnwood. It should be noted that this figure could be reduced by up to 20 percent with higher levels of capacity utilisation (i.e. the fixed operating costs could be spread across a greater volume of production).

5.4.3 Capital costs

It is difficult to determine the value of the capital utilised in an “average” sawmill. As a starting point, the likely cost of plant and machinery seen during sawmill visits was estimated from international data sources, adjusted to cost and price levels in Fiji (e.g. allowing for import tariffs and shipping costs). Most forest processing machinery is imported, so this provides a useful starting point for the analysis. These estimates were then adjusted for depreciation, to take into account the age of machinery seen during sawmill visits (see Whiteman (1999b) for a description of how to do this). The results obtained were then compared with the results of discussions with sawmillers about machinery that they had just purchased or installed in their mills. Generally, the figures obtained by these two routes were broadly comparable and resulted in the estimates used in the forest industry model shown in Table 16 below.

The total capital investment in an “average” sawmill was estimated at about FJD 1.4 million, with a further FJD 573,000 invested in working capital. The figures for investment in stationary equipment (e.g. bandsaws and kiln drying units) and mobile equipment (e.g. loaders, trucks, etc.) are close to the figures quoted by sawmillers (adjusted for the different capacity levels of the different mills visited). The estimate of total investment in production plant capital is also close to what would be expected for a mill of this size in other similar countries. It is slightly below some of the figures quoted by some sawmillers, but is believed to be a more realistic estimate of the true capital value of the typical mill of this size currently operating in Fiji.

The figure for investment in working capital would be considered very high by international standards. This is due to the generally high levels of stock seen during sawmill visits (particularly stocks of unsold products) and the payment terms that seem to be common within the industry in Fiji.

All industry in Fiji benefits from generous allowances for depreciation (“accelerated depreciation”), such that most items of capital expenditure can be depreciated over five years. Although accelerated depreciation is more of a tax benefit than a measure of the true consumption of capital, it is only used in the model to calculate corporate tax liability, so a depreciation period of five years was used throughout the analysis for all depreciable assets. Various assumptions were also made about the replacement or overhauling of some capital items after three to five years (see Figure 34 in Appendix 3 for further details), based on typical replacement and renewal schedules for the types of sawmilling equipment used in Fiji.

Table 16 Capital cost of the investment in an “average” sawmill in Fiji

Production plant capital (FJD '000)	Amount	Depreciation period
		Total	Remaining
Site and services	120	5	2
Structures	300	5	2
Stationary equipment	650	5	5
Mobile equipment	300	5	5
Mill stores and miscellaneous supplies	30	5	5
Owner's construction overhead	0	5	5
Engineering	0	5	5
Total	1,400
Other capital (FJD '000)
Pre-operating expense	0
Working capital	573
Capitalised interest	0
Total	573
Total capital (FJD '000)
Sum of plant and other capital requirements	1,973

Table 17 Product sales by market and grade of production for an “average” sawmill in Fiji

Product	Grade	Export tax (percent)	Output (percent)	Mill net price (FJD per m3)		Real price change (percent/year)
Sawnwood for export market
Class 1	FF Select	0	50.0		1,200		0.0
Class 1	FF Standard	0	7.5		900		0.0
Class 1	Ungraded	0	2.5		900		0.0
Class 2	FF Select	0	30.0		800		0.0
Class 2	FF Standard	0	7.5		800		0.0
Class 2	Ungraded	0	2.5		750		0.0
Total/average			100		1,009		0.0
Sawnwood for domestic market
Class 1	FF Select	0	15.0		750		0.0
Class 1	FF Standard	0	5.0		550		0.0
Class 2	FF Select	0	5.0		580		0.0
Class 2	FF Standard	0	5.0		480		0.0
Class 3	FF Select	0	35.0		480		0.0
Class 3	FF Standard	0	12.0		480		0.0
Class 4	FF Select	0	15.0		480		0.0
Class 4	FF Standard	0	3.0		480		0.0
Class 3 and 4	Ungraded	0	5.0		250		0.0
Total/average			100		518		0.0

5.4.4 Product mix and selling price

The composition of sawnwood produced by an “average” sawmill (by species class and grade) and the selling prices of each of the different products are shown in Table 17. The output mix was based on the FSA submission, discussions with sawmillers and discussions with the Forestry Department. The prices were based on the same sources, but were reduced slightly in some cases in light of the export statistics produced by the Forestry Department (see Table 12).

The above figures resulted in an average selling price of FJD 763/m³, which is in the region of the average product prices revealed in discussions with sawmillers.

5.5 Estimated economic rent

Based on the cost and price information presented above, the complete cost structure and the level of economic rent generated from forest processing in Fiji was calculated and is presented below.

Box 3 Composition of sawnwood production costs for an “average” sawmill in Fiji

Delivered roundwood cost (excl. charges, incl. ROC of 20% on harvesting operation) 128.47 17% Forest charges 122.27 16% Other fixed and variable operating costs 303.35 40% Total: operational costs 554.09 73% Capital costs 85.17 11% “Normal profit” (i.e. ROC of 20% on sawmill investment) 82.35 11% Corporation tax 35.25 5% “Excess profit” 6.27 1% Total: capital costs, taxes and profit 209.04 27% Average product value (FJD per cubic metre of sawnwood produced) 763.13 100%

Box 3 shows how the average (sawnwood) product value is distributed between operational costs, capital costs, taxes and profit. The delivered roundwood cost includes an allowance for “normal profit” in the harvesting operation, but excludes the payment of forest charges. Forest charges include the payment of all official and unofficial charges (i.e. “goodwill” and “commissions”). Capital costs represents the consumption of existing and replacement capital throughout the life of the sawmill investment (assuming zero residual value), while “normal profit” represents the amount required for the sawmill operator to earn a 20 percent ROC (after tax) on their equity investment in the sawmill. Corporation tax is the estimated average annual tax liability divided by annual sawnwood production.

The economic rent from production is equal to forest charges plus “excess profit” and amounts to FJD 128.54 per cubic metre of sawnwood produced, or about 17 percent of the average product value. With the assumed product recovery rate of 50 percent, this implies that the economic rent is equal to FJD 64.27 per cubic metre of roundwood consumed by the sawmill.

Figure 24 shows the distribution of economic rent from production (per cubic metre of roundwood) between landowners, the Government and industry. The figure shows that total forest charges currently capture 95 percent or almost all of the economic rent from production. Fees paid to the Forestry Department and NLTB each account for a further 5 percent of the economic rent. Landowners receive the remaining 85 percent of the economic rent. Of this, slightly more than half of this comes through the official system of royalty payments as opposed to the unofficial system of “commissions” and payments for “goodwill”.

It should also be noted that, in addition to the payments of charges to the Government, the “average” sawmill should be paying corporation taxes to the Government of around FJD 35.25 per cubic metre of sawnwood produced (equal to about FJD 17.63 per cubic metre of roundwood).

Figure 24 Distribution of the economic rent from production in Fiji between landowners, government and industry per cubic metre of roundwood (CUM)

5.5.1 Comparison with the FSA submission

Although this analysis has used some of the information from the submission on forest charges by the FSA (FSA, 2003), it has also incorporated a lot of other data and information from experiences in other countries, information from international data sources and statistics held by the Forestry Department in Fiji. In addition, the calculation methodology used differs significantly from the approach taken by the FSA.

The estimate of delivered roundwood costs produced by the FSA is somewhat higher than the figures presented here. Their figure of FJD 94/m³ (excluding forest charges) is higher than the estimate presented here of FJD 64/m³. This could be due to the amount for “cartage” in their calculations (which is not explained in their submission). Another difference between the two approaches is that their figures include an amount for overheads in the harvesting operation, while the calculations here have assumed that all overheads would be covered in the forest processing operation.

On the processing side, the estimate of processing costs calculated here is somewhat higher than the FSA estimate. The differences in the calculation methodologies are so large that it is not possible to identify what might be the cause of these differences.

Although there are some differences between these figures and those from the FSA, when both parts of the production operation are combined, the figures calculated here are remarkably close to the figures from the FSA. For example, they quote figures for total sawnwood production costs (excluding a profit margin) in the range of FJD 570/m³ to FJD 720/m³. The figure calculated here for total operational costs plus capital costs amounts to FJD 640/m³ (see Box 3), which is roughly in the middle of this range.

As noted earlier, it is always difficult to estimate production costs and economic rent with any degree of accuracy, because of the difficulty of collecting reliable information about costs and prices in the industry. This is particularly difficult when the time available for interviews, field visits and data collection is very limited. This analysis has tried to bring some international experiences and a rigorous analytical approach to the forest charging issue, in order to provide the Forestry Department with an independent estimate of the economic rent from production. As the figures calculated here are not markedly different from those provided by the FSA, it is suggested that this might be taken as an indicator of their reliability.

5.5.2 Species variation

The level of economic rent from production undoubtedly varies by species, as higher valued species can be sold for more profit in the market place. However, when considering the adjustment of forest charges by species group it is important to take into consideration the different grades of product quality that can be manufactured and the acceptability of each species in local and export markets. For example, Class 1 species produce a mixture of high, medium and lower quality sawnwood grades, which are sold in both the local and export markets. Furthermore, forest operators are required to take the lower valued species (and to pay for them) in their licence areas. Added to this, there is the problem already noted of collecting information about the selling prices of different species and product grades in different markets. This makes it very difficult to come to any conclusion about what changes, if any, should be made to the relative levels of forest charges between different species groups.

It is not proposed here to present any firm recommendations about revising the relative levels of the different royalty classes or changes in the species included in each of the classes. Rather, the analysis below presents a simple methodology that could be used to explore this issue further (if desired).

Figure 25 The relationship between the selling price of sawnwood and the economic rent from roundwood production in Fiji

Figure 25 presents the relationship between the selling price of sawnwood and the economic rent from roundwood production, based on the cost data collected and used in the forest harvesting and forest industry models. The relationship for exported sawnwood was derived by assuming that all sawnwood production would be sold in this market and assuming that only one species would be used for production. For any given sawnwood price, the royalty payment for this species was then adjusted until the NPV of the sawmill equalled zero. The economic rent from roundwood production was then taken from the summary table presented in the forest industry model. This process was repeated for different levels of sawnwood prices and under an alternative assumption that all production would be sold in the domestic market, to arrive at the two curves and equations shown in the figure.

The curve for sawnwood sold in the domestic market is higher than the curve for exported sawnwood, because the operational costs are lower for the former (e.g. due to lower packaging, drying and treatment costs). If costs are lower then, for a given sawnwood price, the economic rent will be higher. The difference between these two curves is FJD 20/m³, which is equal to the differences in these costs (see the figures in Table 15 and divide the differences by two to convert from sawnwood to roundwood). It must also be remembered that, although the relationship is higher for sawnwood sold in the domestic market, the selling price for sawnwood of a given species and grade will generally be lower in the domestic market than in the export market, leading usually to a lower level of economic rent.

Table 18 Calculation of the average level of economic rent by species class

Royalty class	Market	Grade	Proportion of total production	Selling price (FJD per m3)	Economic rent (FJD per m3)	Contribution to total economic rent (FJD per m3)	Average level of economic rent by market and royalty class (FJD per m3)
Class 1	Export	FF Select	25.00%	1,200	252.80	63.20	230.23	161.28
Class 1	Export	FF Standard	3.75%	900	117.39	4.40
Class 1	Export	Ungraded	1.25%	900	117.39	1.47
Class 1	Domestic	FF Select	15.00%	750	69.02	10.35	50.97
Class 1	Domestic	FF Standard	3.75%	550	-21.24	-0.80
Contribution to the economic rent from the whole operation						78.63
Class 2	Export	FF Select	1.25%	800	72.25	0.90	67.24	37.24
Class 2	Export	FF Standard	7.50%	800	72.25	5.42
Class 2	Export	Ungraded	2.50%	750	49.69	1.24
Class 2	Domestic	FF Select	2.50%	580	-7.70	-0.19	-30.27
Class 2	Domestic	FF Standard	2.50%	480	-52.84	-1.32
Contribution to the economic rent from the whole operation						6.05
Class 3	Domestic	FF Select	17.50%	480	-52.84	-9.25	-52.84	-52.84
Class 3	Domestic	FF Standard	6.00%	480	-52.84	-3.17
Contribution to the economic rent from the whole operation						-12.42
Class 4	Domestic	FF Select	7.50%	480	-52.84	-3.96	-75.40	-75.40
Class 4	Domestic	FF Standard	1.50%	480	-52.84	-0.79
Class 3 and 4	Domestic	Ungraded	2.50%	250	-156.64	-3.92
Contribution to the economic rent from the whole operation						-8.67
Average level of economic rent from the whole operation						63.59

Note: the sawnwood selling price is per cubic metre of sawnwood, all of the economic rent figures are expressed in terms of the economic rent per cubic metre of roundwood.

Table 18 shows how these relationships might be used to examine the relative levels of economic rent produced from harvesting different species. The figures in the first five columns in the table are the same as those presented in Table 17, re-ordered by species (Royalty class).¹⁵ The economic rent shown in the sixth column is calculated from the equations given in Figure 25. The contribution to total economic rent (column seven) is the economic rent for each individual product (i.e. row) multiplied by its contribution to total production (i.e. the figures in columns six multiplied by the percentages in column four).

As the subtotals for each species class show, at the given selling prices, Class 1 species account for the majority of economic rent, with a modest contribution from Class 2 species. Class 3 and Class 4 species have a negative economic rent and are unprofitable (i.e. production costs exceed the sawnwood prices used here, even before forest charges are taken into account). The sum of the four subtotals - FJD 63.59/m³ - is very close to the average level of economic rent presented above (FJD 64.27/m³), indicating that these two equations are quite accurate.

The figures in the last two columns of the table are the most interesting, as they show the average level of economic rent by market and species class. These are calculated as weighted averages for each market and species class, using the percentages and estimates of economic rent (columns four and six) for the respective sub-sets of species class and market. For example, the weighted average level of economic rent for Class 1 species sold in the export market is calculated as follows:

(25 x 252.80) + (3.75 x 117.39) + (1.25 x 117.39) = 230.23

(25 + 3.75 + 1.25)

These figures show that, under the given assumptions about selling prices and distribution of sales by market and grade, the economic rent for Class 1 species is about FJD 160/m³ and about FJD 40/m³ for Class 2 species. Again, Class 3 and Class 4 species have a negative level of economic rent. This would suggest that the relative levels of the royalties should be much higher for the top two species classes with, possibly, a zero royalty for the bottom two classes.

The above analysis has only served to demonstrate how this issue might be investigated further. A table such as the one above could be completed with more reliable data about prices and the composition of outputs by species (market and grade) in order to examine the relative differences in economic rent between the different species.

In the interest of maximising the potential production from the forest, it is desirable to have producers remove all commercial species from the forest. To encourage this, there will probably always be a certain amount of cross-subsidisation in forest operations, where processors are allowed to take high valued species at sub-optimal prices in return for also having to take the lower valued species. However, this simple analysis suggests that there may be scope to widen the difference in royalty payments between the lowest and highest royalty classes in order to encourage the harvesting of the lower valued species and the development of markets for them.

¹³ For example, see Whiteman (1999b) for details of internationally available sources of information about machinery and equipment costs

¹⁴ Indeed, the harvesting cost model also calculates delivered roundwood cost including a profit margin or return on expenditure as an alternative calculation methodology - see Figure 31.

¹⁵ Note: the percentage contributions to total output are half of the amounts shown in Table 17, because they take into account that half of production is sold in the export market and half is sold in the domestic market.