Previous Page Table of Contents Next Page

14. Financial assessment of reduced impact logging techniques in Sabah, Malaysia - John Tay*, John Healey and Colin Price*

* Innoprise Corporation Sdn Bhd, P.O.Box 11622, Kota Kinabalu, Sabah, Malaysia, Tel: ++(60 88) 24 4130/24 4100 Ext.106, Fax: ++(60 88) 24 3244, E-mail:

** University of North Wales, Bangor, United Kingdom, E-mail: and


Reduced impact logging (RIL) has been given due recognition in recent years in the Asia-Pacific region because of its relevance to sustainable tropical forest management. The primary reason for this interest is that RIL can minimize on- and off-site impacts compared with “business-as-usual” or conventional logging (CL). Case studies have shown unambigouously that RIL can reduce logging disturbance by as much as 50 percent compared with conventional logging (Marsh et al., 1996). Also, the incremental benefit from RIL in reducing logging damage for future crop trees reduces carbon volatilization during and after logging (Pinard and Putz, 1996; Healey et al., 2000). Thus, the adoption of RIL is a right step towards achieving a sustainable supply of goods and services from tropical forests. It is also an attractive option for utility companies that emit greenhouse gases to offset their liability in forest sinks through changes in forest practices. These initiatives provide an alternative source of funds for forest enterprises to invest in the rehabilitation of logged-over forests.

Although RIL is an emerging harvesting system, its adoption beyond pilot or research-scale experiments has been modest. One factor that appears to impede the uptake of RIL by loggers is the lack of information on the costs and benefits of applying RIL. This paper attempts to bridge the information gap with respect to the financial aspects of RIL.

In this paper, the financial costs and benefits of RIL are based on a case study carried out in Sabah, Malaysia. The paper draws upon a larger study on the economics of RIL covering timber and non-timber values (Tay, 2000). Its aims are essentially to:


Study site

The study was conducted as part of a project located in the Sabah Foundation forest concession area in Sabah, Malaysia. The pilot phase of the project comprised 1 400 ha and lasted three years from 1992 to 1995. In 1996, the project area was expanded by another 1 000 ha.

The topography of the project site is hilly and characterized by broken ridges with altitudes ranging from 100 to 1 200 m above mean sea level. Numerous creeks and streams dissect the area. Mean annual rainfall recorded near the project site at Danum Valley Field Centre is 2 500 mm. There are generally two slightly wetter periods in the year with the influence of the wetter northeast monsoon that lasts from October until February, and the southwestern monsoon between May and September. The maximum and minimum annual temperatures recorded were of 31o C and 23o C.

The forest is dominated by trees of the Dipterocarpaceae family. It has an unevenage structure of five distinct canopy layers corresponding to seedlings, saplings, poles, mature and emergent trees. In one study of an 8 ha primary forest near the study site, 511 species of dipterocarps in 59 families and 164 genera were counted (Newberry et al., 1992). The mean density of trees greater than 1 cm diameter at breast hight (DBH) was 2 248/ha. The mean density of lianas over 2 cm DBH was estimated at 881/ha (Campbell, 1990).

Logging techniques

Conventional logging in Sabah utilizes chainsaws (Stihl 070) and bulldozers (D7F Caterpillar). Prior to harvesting, the fellers and bulldozer operators cruise the area to determine the approximate locations of trees, roads and skid trails (map scale 1: 50 000 with contour intervals of 50 m).

During harvesting, a team of one or two fellers locates commercial trees. All trees above 60 cm DBH are felled except those with visible defects. Fruit trees and trees of less than 60 cm DBH are not cut. Generally, timber fellers have complete freedom over the direction and method of felling. Consequently, felling damage can be excessive in uncontrolled logging and as much as 62 percent of stems suffer total damage (Fox, 1968). However, the severe felling damage in Sabah is also caused by high felling intensity of 8-15 trees/ha above 60 cm DBH (i.e. 80-150 m3/ha). In addition, the abundance of climbers or lianas (500-900 m3/ha) significantly influences felling damage by pulling down and uprooting neighbouring trees. After a tree is felled, it is trimmed to a log length of about 6 m. Logs are then extracted to the roadside by bulldozers. The bulldozer team determines the layout of the skid trail and leaves a pattern of extravagant skid trails.

Under the RIL system, the emphasis is on planning and controlled harvesting operations with respect to felling and skidding activities. The Sabah RIL guidelines are based on best management practices of the Queensland Forest Services, Australia. They comprise the following components:

These harvesting guidelines are additional to those prescribed for CL practices. They were conceptualized and tested in the field for application under Sabah conditions prior to their full adoption.


The objective of the financial assessment is to compare RIL with CL in terms of costs and benefits. The assumption is that observed post-logging differences between RIL and CL can be attributed to the different logging methods rather than to variations between the sites.

Source of data

The data were generated directly in the study area through the establishment of a network of growth and yield plots. Eight forest management units totalling 406 ha were divided into four pairs. In each pair, one unit was subjected to RIL and the other to CL during 1993 and 1994. The following parameters were measured: area logged; area converted to bare soil under skid trails, log landings and roads; timber volume extracted; composition and damage to the residual stands. A time and motion study of tree felling and log extraction was carried out in another equivalent pair of forest management units. Log prices, log grades and logging costs were obtained from the forest enterprise concerned.

Timeframe of analysis

The analysis covered two cutting cycles at year 0 (t0) when the first harvest was undertaken in the primary forest, and at year 60 (t60) when the second harvest will be made. The reason for taking two harvesting cycles was that timber yields after 60 years are assumed to be markedly different from the initial yields.

In order to investigate the financial impacts of the second harvest scheduled at t60, the post-logging inventory data were projected over 60 years using the process-based DIPSIM forest growth model developed by Ong and Kleine (1995). Model outputs were validated against the independent results of Ong and Kleine (1995) and those of Bossel and Krieger (1994) for logged forests in Sabah. The t60 model outputs for tree density and size were used to predict the yields under RIL and CL. For the second harvest, the study assumed that previous alignments and locations of roads, log landings and skid trails would be re-used using the same harvesting techniques in the respective RIL and CL logged units.

Levels of analysis

Three different levels of analysis are used (Figure 1):

Figure 1. Schematic representations of the three levels of analysis

Financial indicators

The relative advantage of RIL and CL is expressed by the net present value (NPV) criterion. The NPV provides a means of evaluating and comparing a stream of benefits and costs of RIL and CL over time by applying discount rates. The NPV of benefits B and costs C at time t, given discount rate i, is calculated as follows:

For the financial analysis at t60 discount rates from 2 to 10 percent were chosen to enable decision-makers in both the private and public sectors to assess the viability of the project.


Physical impacts

Harvest area

Of the 176 ha allocated to the CL units for logging, almost all of the area was logged. In RIL units, only 129 ha (56 percent) of the 230 ha was logged (Table 1). Forty-four percent remained unlogged as, according to Forestry Department Rules, slopes steeper than 35° are excluded from logging, although usually this rule is not enforced strictly throughout Sabah. Within the RIL and CL units, these areas were logged but the RIL harvesting guidelines permit logging in these areas only when incidental damage is minimal by having the tree crown falling out of these zones. Approximately 36 percent out of 47 ha with slopes over 35° were logged in the RIL units.

Harvest volume

The mean volume of timber extracted in the RIL and CL units of 106 and 136 m3/ha, respectively, was within the extraction intensity of 40-160 m3/ha reported in other parts of the study area. Such extraction intensities occur also in other locations in Sabah (Fox, 1968; Chai and Udarbe, 1977; Marns and Jonkers, 1981).

Harvesting timber using RIL techniques resulted in a substantial volume of timber foregone on steep slopes, in buffer zones and areas where felling is unsafe and causes environmental problems. The net timber volume foregone in RIL amounted to approximately 35 m3/ha. This has serious financial implications.

Timber stocking

Prior to harvest, total stem densities for trees greater than 1 cm DBH in the CL and RIL units were 4 382 and 3 798 trees/ha (Table 1). The original stand structure did not differ significantly for the six DBH classes (1-5, 5-10, 10-20, 20-40, 40-60, >60 cm DBH) except for trees with 1-5 cm DBH. Following harvest, tree density in four of the six diameter classes (5-10, 10-20, 20-40 and >60 cm DBH) was significantly greater in the RIL than the CL units (Figure 2).

An assumed benefit of RIL is that it leaves a more valuable forest for the second harvest due to lower damage compared with CL. Based on the simulation results, the number of trees greater than 10 cm DBH in the RIL units was about 10 percent higher than in the CL units (RIL = 624 trees/ha; CL = 567 trees/ha, Figure 3). Growing stock was about 31 percent higher (RIL = 343 m3/ha; CL = 260 m3/ha). The biggest difference was for trees with 20-40 cm DBH. They totalled 41 trees/ha (31 percent difference between treatments) and 44 m3/ha (35 percent between treatments).

Although the growing stock benefited from RIL, the potential gains were not realized fully in the second harvest. It appears that trees in these diameter classes had not reached the harvestable size by the end of the 60-year simulation. Trees of 10 cm DBH would have to grow at a mean diameter increment of 1 cm/yr to reach a harvestable size of 60 cm DBH in 60 years. However, growth rates for dipterocarps or non-dipterocarp species of 10 cm DBH are commonly much lower. For example, Abdul Rahman et al. (1992) reported that the periodic diameter annual increment of dipterocarp and non-dipterocarp species greater than 10 cm DBH in a tractor-logged forest in Pahang, Malaysia was 0.52 cm/yr and 0.30 cm/yr, respectively. Chiew and Garcia (1998) reported a 0.56 cm/yr mean diameter growth rate of all species greater than 10 cm DBH in a tractor-logged forest in Sabah. In Peninsular Malaysia, the diameter growth rates of dipterocarps and non-dipterocarps greater than 10 cm DBH were 0.60-0.69 cm/yr (Thang and Yong, 1994). The figures for a selection of dipterocarps in Sarawak were 0.30-0.43 cm/yr (Chai et al., 1994).

Logging damage

In extracting 9 to 13 trees of above 40 cm DBH, the overall damage inflicted on the residual forests averaged 60 percent and 30 percent in the CL and RIL units, respectively (Table 1). The considerably lower damage in the RIL units was due to several factors. First, climber-cutting before felling reduced the overall proportion of trees with crowns snapped off (Appanah and Putz, 1984).

Second, the benefit of directional felling was most evident for trees in the 5-40 cm DBH range where the remaining stem density was higher in the RIL units compared with the CL units. In the RIL units, trees were marked for directional felling with due consideration to existing regeneration to minimize felling damage. Furthermore, trees were felled to facilitate skidding and minimize the bulldozer movements, thereby reducing incidental damages (Fox, 1968). Directional felling did not reduce the incidence of snap-off in the RIL units. It is most likely that the harvested trees had large crowns and stems that caused unavoidable damage to smaller trees.

Table 1. Summary of physical impacts of RIL and CL techniques on different forest values

Forest values



RIL-CL difference (%)

A. Timber

  Logging area

     · Before logging

· 230 ha

· 176 ha

· Area logged in

     · After logging

· 128 ha

· 175 ha

RIL = 56 %, CL = 99%

Stand structure (>1 cm DBH)

     · Before logging

· 3,798 ±101 trees ha-1

· 4,382 ± 212 trees ha-1

· Density after logging

     · After logging

· 3,001 ± 131 trees ha-1

· 2,463 ± 212 trees ha-1

RIL =79%, CL = 56%

Species (density of dipterocarps)

     · Before logging

· 522 ± 69 trees ha-1

· 742 +100 trees ha-1

· Density after logging

     · After logging

· 388 ± 46 trees ha-1

· 435 ± 49 trees ha-1

RIL = 74 %, CL = 57 %

Removals (extracted and killed)

· 797 trees ha-1

· 1,920 trees ha-1

· 58 %

· Volume extracted

· Att 0

· 9 trees ha-1 or 106 m3 ha-1

· 13 trees ha-1 or 136 m3 ha-1

· 31 %

· Att 60*

· 111 m3 ha-1

· 85 m3 ha-1

· 31 %

· Trees damaged (destroyed)

· 21 % of original stems

· 44 % of original stems

· 52 %

· Forgone timber

· 50 % by area (35 m3 ha-1)

· None

· 100 %

B. Soil*

· Length (m)

· Skid trails

· 8.17 ± 0.37 m

· 34.83 ± 0.99 m

· 77 %

· Roads

· 3.76 ± 0.53 m

· 4.67 ± 0.15 m

· 20 %

· Area (ha)

· Skid trails

· 4.3 ± 0.2 ha (4 % of logged)

· 21.0 + 0.8 ha (12 %)

· 80 %

· Log landings

· 0.7 ± 0.1 ha (<1 %)

· 1.7 ± 0.2 ha (1 % of logged)

· 59 %

· Roads

· 4.3 ± 0.6 ha (4 %)

· 6.5 ± 0.4 ha (4 %)

· 34 %

· Percent

· Bladed surface

· 38 ± 5

· 87 ± 3

· 56 %

· Churned soil

· 50 ± 4

· 11 ±3

· 354 %

· Intact topsoil

· 9 ± 5

· 2 ± 0.41

· 78 %

Figure 2. Density of trees over 1 cm DBH before and after logging in units logged by CL (first column) and RIL (second column) techniques. The original stand structure in the experimental units did not differ significantly for the six DBH classes except for trees in the 1-5 cm DBH P 0.06 (t=2.291). After logging, there was a higher number of standing trees in the RIL units across all DBH classes. Tree densities in four of the DBH classes (5-10; 10-20; 20-40 and >60 cm DBH) were significantly greater in the RIL than the CL units

Third, better planning of skid trails in RIL made an important contribution. As skid trails were marked on the ground, bulldozer operators had a better sense of direction and purpose, hence, avoiding unnecessary movements. Marking skid trails also prevented the bulldozer operator from driving the machine to the stump, and in the process, causing a higher incidence of stem damage. With planning, there are fewer skid trails and log landings in RIL units, which also reduces the loss of trees.

Open spaces

The total area of skid trails, log landings and roads in the RIL units was only 40 percent of that in the CL units (Table 1). They represented approximately 7 and 17 percent of the total area logged in the RIL (129 ha) and CL units (175 ha), respectively. All three categories of openings (skid trails, log landings and roads) in the RIL units were smaller than in the CL units, but only skid trails showed a significant difference in area between the treatments.

Figure 3. Simulated diameter distribution (logarithmic-transformed) in units logged with CL (top frame) and RIL over 20-year time intervals

RIL units had a smaller area of skid trails occupying 4 percent of the total area logged compared with 12 percent in the CL units. Skid trails in the CL units, however, occupy a far larger proportion of the total forest area (Figure 4). The average length of skid trails in the RIL units was significantly shorter than in the CL units. Skid trail widths were the same.

Figure 4. Extent and intensity of skid trails (light line), log landings (dark lumps) and roads (dark lines) in CL and RIL units. The shaded areas in the RIL units represents unlogged areas, where slope is greater than 35°.

Skid trails in the RIL units with a bladed surface were only half the proportions encountered in the CL units (Table 1). The proportion of skid trail area with churned soil in the RIL units was higher than in the CL. The proportion of skid trail area with topsoil intact was also higher in RIL units than in CL.

Log landings occupied less than 1 percent of the total area logged in both units. Their average size in the RIL unit was less than half that in the CL units (Table 1).

The total area of roads in RIL units was slightly more than half that in the CL units (Table 1). Roads represent approximately the same proportion in both RIL and CL units; about 4 percent of the total area logged.

Logging efficiency


The total time taken to fell a tree differed significantly between RIL and CL (Figure 5). RIL took approximately 32 minutes and CL took 19 minutes. The main reason for this difference was attributed to differences in work organization and the manner of executing the individual work elements.

Figure 5. Felling time for CL and RIL. The top frame shows the time taken to fell trees, the middle frame the time for bucking, and the lower frame non-productive time

In RIL, fellers cut only two to three trees at a time in one area to minimize damage inflicted by trees falling on trees that had been cut down previously but not extracted. This procedure also eases skidding jams where fallen trees have been stacked on each other amidst logging debris. Bucking was done immediately after felling to facilitate skidding. In CL practices, fellers cut trees without interruption from 0700 to 1130 hours, and with no apparent regard to the facilitation of skidding. In most cases, bucking was left unfinished and had to be postponed until skidding commenced because of the congestion created by too many trees in one location.

On average, RIL and CL fellers spent 66 percent and 72 percent respectively of their time productively. Time taken to direct the fall of a tree was significantly higher in RIL than CL. This was the most time-consuming activity. RIL’s higher unproductive time was due to fellers having to wait for fallen trees to be extracted to prevent log jams. Idling and personal times in RIL were also much longer compared with CL but were related to the longer working hours in RIL which made them more apparent compared with CL.


The total time required to extract a log using RIL was significantly longer than in CL (RIL = 39 minutes; CL = 26 minutes, Figure 6). With RIL, 72 percent of the overall time for skidding operations was spent on productive work elements compared with 86 percent with CL. The time taken to winch logs was the most important variable in RIL because logs were skidded uphill to disperse skid trail runoff.

Figure 6. Skidding times for CL and RIL. The top frame indicates productive times and the bottom frame unproductive times.

The corresponding non-productive times for RIL and CL were 28 percent and 14 percent of total skidding times. More frequent stoppage of work due to mechanical failures occurred in RIL, possibly due to uphill skidding. Personal and delay times were also higher in RIL, but this could be associated with the longer working hours that made these times more apparent to the study.

Cost-benefit analysis

Per volume (m3) logged

The net contribution of RIL from the first harvest (t0) was less than half of CL (RIL = RM27/m3; CL=RM57/m3, Table 2). The lower NPVRIL was due to high extraction costs, and a lower yield. RIL costs 18 percent more than CL. The bulk of the additional cost comprised extraction costs at RM18/m3 or 12 percent of CL.

Table 2. Summary of financial analysis of RIL and CL on a per cubic metre/logged/management hectare

Year of Harvest


Per m3


Per logged ha


Per representative ha
















20 776

27 200


11 653

27 045






17 864

19 485


10 019

19 374






2 912



1 634

7 671







73 260

55 133


41 089

54 820






48 559

30 579


27 235

30 405






24 701

24 554


13 854

24 415







27 613

32 269


15 488

32 086






10 440

15 199


5 855

15 112






5 260

10 049


2 949

9 992






3 661

8 459


2 053

8 411






3 156

7 957


1 770

7 912






2 993

7 796


1 678

7 752



Price/cost increase at 2 % p.a.

Harvest yield at t0... RIL=106 m3/ha CL=136 m3/ha

Harvest yield at t60... RIL=111 m3/ha CL=85 m3/ha

For the second harvest (t60), RIL and CL yielded RM223/m3 and RM289/m3 respectively, without discounting. The reduction in profit due to the adoption of RIL was 23 percent of CL.

Total t0 and t60 NPV for RIL and CL without discounting was RM250/m3 and RM346/m3, respectively. At a 2 percent discount rate, the total t0 and t60 net NPVRIL dropped to RM95/m3 or 62 percent of the undiscounted value compared with a drop of 58 percent for CL. Higher discount rates favour CL even more. This phenomenon is associated with the fact that high interest rates favour high yields in the short term (Price, 1989; Leslie, 1987).

The profitability of RIL was sensitive to log price changes using data from the second harvest only (Table 3). Without discounting, a 10 percent increase in log prices across all species for the second harvest would improve the base NPVRIL by approximately 30 percent from RM223 to RM290/m3. The increase in NPVRIL was almost on par with the base NPVCL. At a log price increase of 30 percent, the NPVRIL nearly doubled against the base case. A similar level of improvement was found with discounting, although the NPVRIL value was lower. For example, at a 2 percent discount rate, the base NPVRIL of RM68/m3 rose to RM88/m3, an increase of about 23 percent with a 10 percent change in log prices. Log price increases of such a level or even higher have occurred in Asian countries before. For example, real prices of Asian tropical logs more than tripled in the first half of 1993 (ITTO, 1996). However, the improvement in RIL over CL with log price changes assumes that CL prices will not increase. This may or may not be the case depending on the supply and demand of logs.

Table 3. Sensitivity analysis on RIL for harvest at t60 on a per cubic metre basis

Discount rate (%) ==>

Net Present Value














1 Price increased by






















2 Log grade increased by






















3 Cost decreased by






















Timber prices might also increase because of certification, which may provide access to niche markets with a price premium. Certification requires the application of environmentally sound logging practices. A number of tropical countries have recognized the growing interest in certification and have responded with the development of national certification programs. Log price changes associated with timber certification would only affect timber from well-managed forests and not from conventionally logged forests.

The NPVRIL (t60) was also sensitive to changes in timber grade. Without discounting, a 10 percent improvement in SQ grades resulted in a 16 percent higher NPVRIL. With a 50 percent improvement in SQ grades, the NPVRIL was 78 percent higher. When benefits were discounted, identical levels of NPVRIL increases were obtained since all effects are at year 60. Although log grade improvement has a smaller effect on NPVRIL than price change, it can be influenced directly during harvesting operations. For example, directional felling, winching and avoiding log jams reduce damage to logs. These factors affect especially the t60 revenues because the quality of the timber from the next cut depends on the degree and extent of damage inflicted on future crop trees during the first cut. For the t0 revenues, the difference in timber quality between the RIL and CL units was less critical because timber was harvested in old growth forests.

Reducing the cost of extraction and operating overheads also significantly increases the NPVRIL. Without discounting, a 10 percent reduction in these costs raises the NPVRIL by 17 percent. At 50 percent reduction in costs, the NPVRIL nearly doubles. With discounting, identical levels of improvements in NPVRIL were found.

There are a number of opportunities to reduce the extraction costs. First, log harvest planning could be confined to loggable areas only, which can be only slightly more than half of the total area. Hence a 100 percent stock mapping incurs unnecessary costs. Stock mapping (including the production of field maps) costs RM1.35/m3 or RM143/ha, and is the most expensive activity of the additional activities required for RIL. There is a need to find an alternative means to determine what, in a given area, is loggable and unloggable. Second, the non-operational costs include other costs such as consultancy and training costs, which are expected to decrease over time when RIL becomes an integral part of forest management and staff is trained sufficiently. Third, there could be potential savings in RIL through reduction in machine wear and tear from better-planned skid trails and harvesting techniques. The potential savings in extraction costs could be at least 10 percent.

Per logged hectare

For the first harvest, the profits generated from using RIL and CL were RM2 912/ha and RM7 715/ha, respectively (Table 2). The 62 percent lower profits for RIL were due to lower harvested volumes. However, if only extraction activities were considered, RIL costs RM344 more than CL (RIL = RM9 573 and CL = RM9 917). This peculiarity was due to the opportunity cost of foregone timber in RIL not being included in the analysis, as some costs that were paid piecemeal on the basis of area covered (e.g. stock inventory, climber cutting etc) did not consider the foregone area.

For the second harvest, the NPVRIL and NPVCL without discounting were RM24 701/ha and RM24 554/ha, respectively. The relatively small difference favouring CL was due to improvement in RIL’s t60 yields over CL. At 2 percent discount rate, the total t0 and t60 net NPVRIL were lower than NPVCL by 71 percent. With increasing discount rates, the financial returns of both RIL and CL decline drastically.

Per representative hectare

The total undiscounted net benefits per representative hectare logged by RIL and CL at t0 and t60 were RM15 488 and RM32 086, respectively (Table 2). The 52 percent difference was due entirely to the opportunity cost of foregone timber that could not be harvested in RIL. With a 2 percent discount rate, the total net benefit for RIL and CL had similar trends of declines as per logged area. Higher discount rates make RIL even less attractive.

Comparing levels of analysis

It is evident that in most cases, RIL has substantial net costs (i.e. the net benefits are much less than those of CL). This is mainly because of the lower timber volumes harvested at t0. The assessment per logged hectare shows also the effect of the lower volume harvested by RIL (106 m3/ha) compared with CL (136 m3/ha). The opportunity cost becomes still greater if the smaller proportion of area harvested by RIL (129 ha out of 230 ha) compared with CL (175 out of 176 ha) is considered, which is demonstrated in the per representative hectare calculation.

For the t60 scenario, the lower damage to the remaining stand caused by the initial logging following RIL led to a greater timber yield (111 m3/ha) than following CL (85 m3/ha). Thus per cubic metre, RIL appears to be more profitable than CL. On a per logged and per representative hectare basis, RIL is less financially attractive compared with CL because of opportunity costs due to foregone timber.

Increasing the discount rate from 2 to 6 percent shows RIL to be less favourable when calculating on a per logged hectare and per representative hectare basis. This peculiarity is explained by the harvest yields at t0 and t60. Per logged hectare, the long-term timber benefit at t60 favours RIL. Increasing the discount rate makes this benefit less significant as an offset against the opportunity cost in t0. Per representative hectare, the opportunity cost of t60 timber becomes an insignificant item.


The lessons learned from this study are:

1. RIL effectively reduces logging damage. In extracting 9-13 trees of above 40 cm DBH, logging damage decreased by 50 percent due to the removal of climbers, tree marking for retention of potential crop trees and directional felling, pre-harvest skid trail planning and RIL-imposed harvesting guidelines, e.g. restricting bulldozer movements to marked skid trails, limiting skid trail width and restricting felling in sensitive areas.

2. RIL costs money, especially when it is applied in the hilly terrain of Sabah. This study reveals that RIL is less profitable compared with CL because nearly half of the forest areas needed to be excluded from logging. However, such restrictions on timber production have environmental gains. The financial implications are borne by the timber producer, while society at large benefits. On the other hand, if a larger area needs to be exploited to make up for the shortfall in volume, the forest is used faster than planned and environmental benefits are similarly affected.

3. The “conventional” approach to costing RIL needs re-inventing to capture the true RIL costs. This study has demonstrated that the costs depend on the level of analysis (i.e. per cubic metre logged, per logged area or per representative area). Per cubic metre and per representative area are superior to account for the opportunity costs of foregone timber.

4. Discount rates significantly affect the NPV given that the t60 benefit and cost are distant in the future. Above 6 percent, RIL becomes financially infeasible.

5. The significance of RIL warrants a review of the RIL guidelines in two aspects:

(i) reducing the proportion of the area of a forest actually being allocated for logging; and

(ii) much more careful harvesting of timber from the area actually being logged.

6. If RIL is adopted at all, it has to be paid for at the forest enterprise level by paying the workers properly. At the global level, appropriate transfer mechanisms should be designed to distribute costs equitably.


Innoprise Corporation Sdn. Bhd., the British Council and the International Tropical Timber Organization provided financial support for the research. Numerous people have lent advice and assistance in this research: F.E. Putz, M. Pinard, R. Ong, D. Lee, M. Kleine, P.M. Costa, R. Nussbaum, C. Marsh, S. Williams, A.M. Rajin, and S. Gapid. We are grateful to all of them including others not mentioned here due to space limitation.


Abdul Rahman, K., Wan Razali W.M., Shahrulzaman, I. & Azman, H. 1992. Growth response of hill dipterocarp forest following two methods of logging in Peninsula Malaysia. In: Wan Razali Mohd., Shamsudin, I., Appanah, S. and Abdul Rashid, M.F. (eds.), Proceedings of the symposium on harvesting and silviculture for sustainable forestry in the tropics. Kuala Lumpur, Malaysia. 5-9 October 1992. pp. 24-31.

Appanah, S. & Putz, F.E. 1984. Climber abundance in virgin dipterocarp forest and the effect of pre-felling climber cutting on logging damage. Malaysian Forester, 47: 335-342.

Bossel, H. and Krieger, M.H. 1994. Simulation of multi-species tropical forest dynamics using a vertically and horizontally structured model. Forest Ecology and Management, 69: 123-144.

Campbell, E.J.F. 1990. Ecological relationships between lianas and trees in a lowland tropical rain forest in Sabah, Malaysia. Unpublished M.Sc. thesis, University of Stirling, UK.

Chai, E.O.K., Tan, S.S. & Lee, H.S. 1994. Relative performance of dipterocarp trees in natural forest, managed forest, logged forest and plantations throughout Sarawak, East Malaysia. In: Wan Razali Mohd., Chan, H.T. & Appanah, S. (eds.), Proceedings of the seminar on growth and yield in tropical mixed/moist forests. 20-24 June 1998, Kuala Lumpur, Malaysia. pp. 161-175.

Chai, D.N.P. & Udarbe, M.P. 1977. The effectiveness of current silvicultural practice in Sabah. Malaysian Forester, 40: 27-35.

Chiew, K.Y. & Garcia, A. 1998. Growth and yield studies in the Yayasan Sabah Concession Area. In: Wan Razali Mohd., Chan, H.T. & Appanah, S. (eds.), Proceedings of the seminar on growth and yield in tropical mixed/moist forests. 20-24 June 1998, Kuala Lumpur, Malaysia. pp. 192-204.

Fox, J.E.D. 1968. Logging damage and the influence of climber cutting prior to logging in the lowland Dipterocarp forest of Sabah. Malayan Forester, 31: 326-47.

Healey, J.R., Price, C. & Tay, J. 2000. The cost of carbon retention by reduced impact logging. Forest Ecology and Management, 139: 237-255.

ITTO. 1996. Reduced impact, increased costs? Tropical Forest Update, 6: 10-12.

Leslie, A.J. 1987. A second look at the economics of natural management systems in tropical mixed forest. Unasylva, 39 (155): 46-57.

Marns, H.M. & Jonkers, W. 1981. Logging damage in tropical high forest. UNDP/FAO Working paper no. 5, FO:MAL/76/008. Sarawak Forestry Department, Malaysia.

Marsh, C.W., Pinard, M.A., Putz, F.E., Tay, J. & Sullivan, T.E. 1996. Reduced impact logging: a pilot project in Sabah, Malaysia. In: Schulte, A. & Schone, D. (eds.), Dipterocarp forest ecosystems: towards sustainable management. pp. 293-307.

Newberry, D., Still, M.J. & Campbell, E.J. 1992. Primary lowland dipterocarp forest at Danum Valley, Sabah, Malaysia. I. Structure and family composition. Phil. Trans. Roy. Soc. London, 355 (1275): 323-457.

Ong, R.C. & Kleine, M. 1995. DIPSIM: a dipterocarp forest growth simulation model for Sabah. Research paper no.2. Forest Research Centre, Sabah, Malaysia.

Pinard, M.E. & Putz, F.E. 1996. Retaining forest biomass by reducing logging damage. Biotropica, 28: 278-295.

Price, C. 1989. The theory and application of forest economics. Blackwell, Oxford.

Tay, J. 2000. Economics of reduced impact logging techniques in Sabah, Malaysia. Ph.D. Thesis. University College of North Wales, Bangor, Wales, UK.

Thang, H.C. & Yong, T.K. 1994. Status on growth and yield studies in Peninsula Malaysia. In: Wan Razali Mohd., Chan, H.T. and Appanah, S. (eds.), Proceedings of the seminar on growth and yield in tropical mixed/moist forests. 20-24 June 1998, Kuala Lumpur, Malaysia. pp. 137-148.

Previous Page Top of Page Next Page