Local and national concerns: criteria and indicators

Local and national concerns: criteria and indicators

“Best bet” Land-use Systems

Country reports

Alternatives To Slash-And-Burn In Indonesia

Unique id: IDAZAZYB

Source file: D:\Projects\ASB\ASB Country and Thematic reports\Indonesia PhaseII report\Part IV-V .xml

Authors: Thomas P. Tomich, Meine van Noordwijk, Suseno Budidarsono, Andy Gillison, Trikurnianti Kusumanto, Daniel Murdiyarso, Fred Stolle, Ahmad M. Fagi, Iswandi Anas, A.F.S. Budiman, Kenneth Chomitz, Rebecca Elmhirst, Chip Fay, Hubert de Foresta, Dennis Garrity, Danan P. Hadi, Suryo Hardiwinoto, Kurniatun Hairiah, Genevieve Michon, Nu Nu San, Cheryl Palm, Soetjipto Partoharjono, Djuber Pasaribu, Eric Penot, Robert Simanungkalit, Martua Sirait, S.M. Sitompul, F.X. Susilo, David Thomas

--------------------------------------------------------------------------

Alternative systems and technologies must be profitable and socially acceptable for smallholders; if not they have little prospect for adoption (hence impact). Part IV reports the empirical results of application in Indonesia of the methodological innovations of the ASB global working group on socioeconomic and policy issues (documented in Vosti et al. 1998). The GEF project did not provide funds for empirical research on these essential topics, which affect adoptability of land use alternatives by smallholders and also are the basis for assessing tradeoffs (if any) between national policy objectives and global environmental benefits. Thus, funding had to be sought from other sources – and was secured from the Asian Development Bank (ADB) and the Ford Foundation supplemented by additional funds from DANIDA, the Government of Japan, and others. The process of seeking additional funding delayed work on this key component of the research, which could not begin until funding, was secured in mid-1997.

Assessment Criteria. Empirical results for Indonesia for four sets of indicators – profitability, labor requirements, cash flow constraints, and household food security – will be presented in this part of the report. From among these, a sub-set of indicators will be selected for two of the sets of assessment criteria presented in Part I:

Criteria for smallholders’ socioeconomic concerns: production incentives, labor constraints, and household food security.

Criteria for policymakers’ objectives: growth and aspects of equity and stability

This part of the report will conclude with sections on tradeoffs and complementarities among smallholders’ concerns and policymakers’ objectives and on ‘scaling up’ the assessment from plots to landscapes and watersheds. Criteria for institutional barriers to adoption, which are concerns to both smallholders and policymakers, will be considered in Part V.

IV.1Profitability indicators

Since many of the land use alternatives in Sumatra involve perennials, the appropriate measure of profitability is the net present value (NPV, present discounted value) of revenues less costs oftradable inputs (fertilizer, fuel, etc) and of domestic factors of production (land, labor, management) over the full 25 year period considered in the analysis. Because it can account for input and factor costs as well as outputs and can handle time by discounting future values, this measure of total factor productivity is superior to partial measures of productivity (e.g., yield or output per unit labor).

The policy analysis matrix (PAM) technique provided the framework for estimating profitability indicators as well as the indicators of labor requirements and cash flow constraints discussed below. The ‘PAM’ is a matrix of information about agricultural and natural resource policies and factor market imperfections that is created by comparing multi-year land use system budgets calculated at private and social prices (Monke and Pearson 1989 is the basic reference). Private prices are the prices that households and firms actually face, so private profitability – the NPV at private prices -- is a measure of production incentives. Social profitability, calculated at economic (shadow) prices, removes the impact of policy distortions and market imperfections on incentives for adoption and investment. Thus social profitability —the NPV at social prices -- is an indicator of potential profitability(or comparative advantage). Divergences, the difference between private profitability and social profitability, are indicators of distortions, arising either from policy or from market imperfections and failures. The structure of the PAM is described in Table IV.1, which is taken from Monke and Pearson (1989, p. 19).

As pointed out by our colleague, Arild Angelsen, the list of potential corrections to arrive at social prices is quite long. The adjustments to derive social prices in these analyses focus mainly on policy distortions arising from trade restrictions. As discussed below, we also used a lower real discount rate (15% instead of 20%) to capture a rough approximation of the impact of capital market imperfections on the private cost of capital. We have used the same wage rate in both sets of calculations, implicitly assuming that there are no imperfections in the market for unskilled labor. While this is not completely true, it also seems that these imperfections do not have a significant effect in the unskilled labor market (see discussion of labor markets in Section V.4 below). The main omission here is that prices are not adjusted to reflect costs and benefits of environmental externalities arising from these production activities, such as smoke, ecological changes, and loss of watershed functions. These adjustments, which probably would be significant and which are necessary for the complete analysis, are not possible at this time because of lack of data. Filling this gap is a priority for future research, as discussed below in Section IV.5.

These studies focus on primary production in agriculture and forestry. To get the complete economic picture, especially regarding comparative advantage and growth potential, it would be necessary to extend these analyses ‘downstream’ to include the private and social profitability of processing activities, especially for timber, rubber, cassava, and palm oil. Each of these studies of processing activities (described in Appendix E) would be a major undertaking in its own right and was not feasible during Phase II work in Indonesia.

Table IV.1 Policy Analysis Matrix

Costs

Revenues

Tradable

inputs

Domestic

factors

Profits

Private prices

Social prices

Effects of divergences and efficient policy

I³

J⁴

K⁵

D¹

H²

L⁶

¹Private profits, D, equal A minus B minus C.

² Social profits, H, equal E minus F minus G.

³ Output transfers, I, equal A minus E.

⁴Input transfers, J, equal B minus F.

⁵ Factor transfers, K, equal C minus G.

⁶ Net transfers, L, equal D minus H; they also equal I minus J minus K.

Ratio Indicators for Comparison of Unlike Outputs

Private cost ratio (PCR): C/ (A – B)

Domestic resource cost ratio (DRC): G/ (E – F)

Nominal protection coefficient (NPC)

On tradable outputs (NPCO): A/E

On tradable inputs (NPCI): B/F

Effective protection coefficient (EPC): (A – B)/ (E – F)

Profitability coefficient (PC): (A – B – C)/ (E – F – G) or D/H

Subsidy ratio to producers (SRP): L/E or (D – H)/E

Source: Taken from Monke and Pearson 1989, Table II.1, page 19.

To assure comparability across land use systems (and across ASB sites in Indonesia and Thailand), a regional short course on application of the PAM approach to natural resource management and policy analysis was be held in Chiang Mai, Thailand, 1-13 June 1997. Through participation in lectures and computer-based exercises, teams developed a common methodology for analysis of land use systems. The course, which was funded by ADB, involved eleven participants from Indonesia (see Annex D) plus eight from Thailand. The Indonesian teams trained in the course then undertook studies of six Sumatran land use systems selected for study in ASB Phase II. Five of these six studies were sub-contracted to Indonesian national partners listed in Table IV.2. The sixth, on transmigration systems, was completed by an ICRAF researcher (see Budidarsono 1998). Fortunately, except for the study of industrial timber, preliminary results of these ongoing socioeconomic assessments are available to be included in this report.

Table IV.2 ADB-Funded Grants for Socioeconomic Research in Indonesia

Research Topic

Researchers

Institution

Does shifting cultivation really cause deforestation? Economic analysis of shifting cultivation and five-year bush fallow in LampungProvince

Bustanul Arifin

Agus Hudoyo

Department of Agricultural Economics and Rural Sociology, University of Lampung

Economic analysis of land use system for large scale plantations of oil palm and industrial timber estates

Retno Maryani

Setiasih Irawanti

Forest Products and Forestry Socio-Economics Research and Development Centre, Ministry of Forestry

3. Economic analysis of large scale logging

Machfudh

Wesman Endom

Forest Products and Forestry Socio-Economics Research and Development Centre, Ministry of Forestry

Analysis of the economic efficiency and comparative advantage of the Sumatran small-holder rubber using ‘PAM’ method

PrajogoU. Hadi

Gelar Setya Budhi

Center for Agro Socio-Economic Research, Agency for Agricultural Research and Development, Department of Agriculture

Economic analysis of NTFP extraction in Rantau-pandan, Province of Jambi

Arif Aliadi

Wibowo A. Djatmiko

The Indonesian Tropical Institute (LATIN)

Operational definitions for the six land use types were given at the end of Chapter I.

1. Community-based forest management,

2. Large-scale commercial logging

3. Smallholder rubber, including both rubber agroforests and rubber monoculture.

4. Large-scale plantations of oil palm and industrial timber estates

5. Upland rice with bush fallow

6. Transmigration systems, focusing on cassava and Imperata cylindrica (alang-alang)

See Tables I.2, I.3, and I.4 for additional specifications of these systems. Annex E contains the PAMs for the various scenarios and more information on each of the studies.

All of these studies use the macroeconomic parameters tabulated below because the data were collected in July 1997, when the exchange rate was about Rp 2400 / US dollar. By most assessments of economic fundamentals (e.g., purchasing power parity), the Indonesian Rupiah was not greatly overvalued at that time. The consensus was that the overvaluation of the Rupiah relative to the dollar may have been 10-15% in June 1997. Some expert analysts even expected the Rupiah to appreciate if it were floated in 1997 (Mc Leod 1997). To almost everyone’s surprise, the collapse of the Thai Baht in July 1997 spread to the Rupiah (among others). By January 1998, the Rupiah had fallen to over Rp 17,000 per US dollar. After a recovery below Rp 10,000, it had fallen again to over Rp 14,000 per dollar in June 1998. The reasons why Indonesia’s currency fell the furthest and has stayed down the longest rest with profound problems in its banks and other financial institutions compounded by the worst social instability and political uncertainty in 30 years.

The impact on land use incentives resulting from this monetary, social, and political crisis will be examined in Part VI. Although the causes of the regional financial crisis are not yet fully understood, they do not reflect fundamentals of the productive sectors of Indonesia’s economy. By any economic measure, the Indonesian Rupiah was extremely undervalued in mid-1998 as a result of the financial, social and political turmoil. (Under these conditions, people demand a huge premium to hold Indonesian currency.) To assess land use alternatives over the longer term, the macroeconomic parameters of July 1997 are a better guide than those that have prevailed during the crisis.

Macroeconomic parameters for PAMs	July 1997
Exchange rate	Rp 2400 / US$ 1
Wage rate in Sumatra	Rp 4000 / day
Real interest rates (net of inflation):
Private:	20 % per year
Social:	15 % per year

Real interest rates – that is interest rates net of inflation -- are the discount factors used to value future cash flows in current terms. As in most developing countries, capital markets in Indonesia are fraught with imperfections – some of which have been manifested in the financial crisis. Private interest rates (at least for smallholders, if not for large corporations that could secure subsidized credit) have been very high in real terms. In July 1997, formal sector lending rates were almost 30% pa and inflation was under 10% pa. Thus the private interest rate of 20% used in these analyses is a lower bound for the actual cost of capital for smallholders. The real social interest rate is less than the private rate and 10% is probably too low. So, somewhat arbitrarily, a rate of 15% has been used for the real social cost of capital, which is both the interest rate and the discount rate for calculating NPV at social prices. This difference between private and social interest rates is the main cause of divergences between calculations at private and social prices for many of the land use alternatives. The analyses are quite sensitive to the choice of discount rates, which unfortunately involves considerable uncertainty. Particularly for the private cost of capital, the subjective discount rate may be much higher (or lower) than the 20% real rate used here. Interest rates in the informal sector often exceed 100% per year. Stein Holden estimated that the average subjective discount rate (rate of time preference) among transmigrants in Riau exceeded 90% (Arild Angelsen pers comm).

On the other hand, as Angelsen has pointed out, ‘desire to claim or secure land rights may modify the effect of high discount rates.’

An activity with NPV less than zero is ‘unprofitable’ by definition. This does not necessarily mean that there are no positive cash flows. Instead, it means that it would be more profitable to do other things with the land, labor and capital than to devote them to this activity. If land is scarce, the NPV estimates measure returns to land because they are the ‘surplus’ remaining after accounting for costs of labor (including imputed value of family labor), capital (through discounting), and purchased inputs.[1] (To the extent that management is a scarce factor, it also would be included in the residual.) We also present a measure of returns to labor, the wage rate that sets the NPV equal to zero. This calculation converts the ‘surplus’ to a wage after accounting for purchased inputs and discounting for the cost of capital; no surplus is attributed to land. This measure of returns to labor is valid when land is abundant and labor is scarce. Returns that exceed the wage, Rp 4000 per day, mean the activity will be attractive to family members compared to off-farm work or would justify hiring labor.

Although local land abundance with household labor scarcity has prevailed historically and certainly continues in the ASB sites in Brazil and Cameroon, this fundamental relationship seems to be shifting in Sumatra. Nevertheless, it still is reasonable to believe that local land abundance and household labor scarcity continue in the forest margins, at least from the point of view of smallholder households in central Sumatra. This is supported by the result that returns to labor for rubber agroforests, the predominant smallholder land use, are almost identical to the wage rate (Table IV.3). This implies that no ‘rent’ accrues to land under the dominant system and is consistent with land abundance (since the ‘rent,’ its opportunity cost, is near zero).

For these reasons, and to facilitate cross-site comparisons, returns to labor valued at private prices was selected as the indicator of profitability for smallholders’ production incentives. Private prices are used in this indicator to reflect actual incentives smallholders faced under policies in effect in mid-1997.

At the same time, local and national policymakers increasingly are making public policy decisions under conditions of land scarcity and labor abundance. Land certainly is a constraint that should be considered by policymakers in choices regarding development of large-scale estates versus smallholders and there are other reasons to believe these development strategies are mutually exclusive (Tomich et al 1995).

Returns to land valued at social prices will be used as the indicator for potential profitability from policymakers’ perspective. Social prices are used to indicate potential value added from this alternative if policy distortions and market imperfections were removed. This impact on value added is directly linked to policymakers’ growth objectives.

Table IV.3 Profitability Matrix, July 1997

Land Use System

RETURNS TO LAND

RETURNS TO LABOR

Wage to set NPV to Zero

NPV Private Prices

NPV Social Prices

Divergences

Private Prices

Social Prices

Rupiah 000 / ha

Rp / person-day

Community - based forest management

8.0 to 16

9.4 to 18

(1.5) to (2.5)

11,000 to 12,000

11,000

Commercial Logging

(804) to (131)

(32) to 2,102

(2,233) to (773)

(17,349) to 2,008

7,,917 to 31,400

Rubber agroforest (seedlings)

1.6

4,000

4,100

Rubber agroforest (clones)

(95) to 2,202

234 to 3,623

(330 to (1,420)

3,900 to 6,900

4,200 to 7,700

Rubber monoculture

(167)

(993)

(826)

3,683

2,600

Oil palm monoculture

275

1,480

(1,204)

5,797

9,981

Upland rice/bush fallow rotation

(220) to (76)

(180) to 53

(37) to (130)

2,700 to 3,300

3,000 to 4,500

Monoculture cassava/Imperata cylindrica

(71) to 360

(315) to 389

135 to 243

3,895 to 4,515

4,085 to 4,455

Estimates of returns to land and returns to labor, each evaluated at private and at social prices, are presented in Table IV.3. The upland rice / bush fallow rotation stands out as being unprofitable, either in terms of potential profitability (returns to land at social prices) or smallholder production incentives (returns to labor at private prices). For the upland rice / bush fallow system, the higher (less negative) returns are for the fallow of ten years or more, which is no longer feasible. The lower (or more negative) numbers in the range correspond to short fallow shifting cultivation. These results are consistent with the disappearance of shifting cultivation in most of Sumatra’s peneplains and piedmont. Sustainable forms of continuous foodcrop production may be technically feasible in Sumatra’s peneplains, but often are not financially attractive because they require too much labor and too many purchased inputs. For this report, we have focused on cassava, which may be among the most profitable of the continuous foodcrop alternatives for the peneplains. The most profitable cassava system studied was an extensive fallow system without any fertilizer applications. Profitability at private prices was estimated at over Rp 545,000 per ha (see Appendix E). However, this example is not included in Table IV.3 because, as noted in Part III, these systems mine nutrients, exhausting the soil and reducing the range of future land use options. Two cassava systems that use fertilizer are included in Table IV.3, one with fertilizer applications from the first year and one with fertilizer beginning in the seventh year after forest clearing. Application of fertilizer from the first year after clearing (30 kg N; 60 kg P; and 60 kg K per year) is not profitable privately (negative Rp 71,000 per ha) or socially (negative Rp 315,000 per ha). These treatments and the agronomic results are taken from experiments conducted at the Biological Maintenance of Soil Fertility (BMSF) research project at the ASB benchmark area in Lampung. However, an intermediate approach (also reported in Table IV.3) with fertilizer applications beginning in year seven (50 kg N; 50 kg P) does produce relatively attractive returns at both private prices (Rp 360,000 per ha) and social prices (Rp 224,000 per ha). However, the longer-run sustainability of this system requires further study. Note that, because of chemical fertilizer price subsidies that were still in effect in mid-1997, cassava is one of the few cases where estimated ‘divergences’ are positive, indicating that policy increases private profitability.

Returns to labor are highest for community-based forest management (extraction of NTFPs), but these high returns are dependent on some mechanism to exclude outsiders. Thus, this system plays an important role for existing communities that can regulate access to forest lands. If, on the other hand, communities could not regulate access to their forests, one would expect the returns to labor from extraction of forest products to decline toward the wage rate. However, even under ‘open access’ one would still expect returns to labor to exceed the wage rate by some margin equal to a risk premium. The risks involved include possibility of failure to find products to extract and also the risk (and associated costs) of detection by officials, since many of these activities are prohibited.

The relatively low returns to land – only slightly above rubber agroforests – suggest that NTFP extraction is not a feasible alternative for large numbers of people, because there is not enough land for everyone to practice this extensive livelihood strategy. These results must be interpreted with some care, however, for three reasons. First, these extractive activities are highly site-specific. It may be that the study site is not representative. Only additional studies can resolve this. Second, as often is the case, at least part of this community forest is on StateLand and it is not clear how this problem of tenure insecurity might bias these results. On one hand, long run profitability may be overstated because of unsustainable harvesting (viz., songbirds and rattan). On the other hand, if the community or individual members had secure property rights, this might induce them to invest and to manage resources to increase productivity over time. Finally, as already noted, it was not possible to put a value on timber extraction, but it is likely that this is significant. We hope to be able to conduct a study of the economics of smallholder timber extraction in the future.

The results for commercial logging appear paradoxical, but this is because of policies that produce the biggest divergences for any of these land uses. First, the sustainable logging regulations – if they really are followed – reduce profitability, mainly by slowing timber extraction. Second, high export taxes (effectively an export ban) for logs and sawn timber depressed the domestic prices of logs from 50-70% below comparable world prices. (Timber export taxes were to be reduced to 30% by the end of 1998.) However, timber companies could get around both of these problems. First, as mentioned above, many companies circumvent regulations on timber extraction. Second, these typically are vertically-integrated firms producing products like plywood for the export market. Therefore, the best indicator of profitability of these activities for logging companies is the figure of just over Rp 2 million per ha, valued at social prices that reflect world prices of forestry products. When comparable estimates are available for industrial timber plantations, it seems likely that these will be more profitable than logging.

By all accounts, illegal logging is common, which seems inconsistent with these results of negative returns to logging at private prices. However, the major cost item for logging concessions -- establishing and maintaining logging roads -- is not incurred by illegal loggers. If one can get access to timber without having to invest in infrastructure (and at the same time circumventing various fees), logging can be very profitable.

One could argue that the estimated NPV of logging activities of over Rp 2.1 million per ha (about US$ 875) in mid-1997 should be added to the social profitability for all the other activities and to private profitability, at least for large-scale estates that often can market timber felled as a by-product of land clearing. Recall that natural forest cover is the starting point underlying these calculations (and all the other estimates in this report). Thus all the forest-derived land uses (rubber, oil palm, cassava, and even upland rice) started out with felling of forest timber. And, as already noted, there is substantial (but as yet unquantified) timber felling in conjunction with NTFP extraction. Thus, in many cases it would be appropriate to add the value of the harvested wood to the profitability of each activity overall. This modification is debatable for private profitability of smallholder systems, however, because most of the felled timber is burned instead of marketed. Yet, this simply may be a result of trade restrictions that make it artificially difficult for smallholders to sell timber legally (Section VII.2). The estimate of timber values was not added to other land uses in the tables presented in the report, however, because the one-off value of timber extracted as a by-product of land clearing often exceeds the value of the derived land use. Thus, although it is technically correct to do so, adding the value of timber – which admittedly is subject to considerable uncertainty – would simply obscure differences in profitability among the derived land uses. This problem in presentation is linked to a problem in conservation: if regulations can be circumvented – as often is the case -- forest conversion is privately profitable simply for the value of timber regardless of the subsequent land use. Of course, for the social profitability calculations, timber values would have to be balanced against losses of ecological and other environmental functions of natural forests.

Oil palm is widely viewed as the most profitable alternative for Sumatra’s peneplains and Indonesia’s oil palm producers have the lowest unit costs in the world. Thus, it is no surprise that large-scale oil palm monoculture is among the most profitable alternatives, either in terms of returns to land valued at social prices or in terms of returns to labor valued at private prices. The later measure is of limited relevance, however, because the official wages for plantation workers are well below these estimates of returns to labor. But, much as they had earlier in Malaysia (Barlow 1986); plots of 2-5 ha of oil palm planted by independent smallholders began to appear in Sumatra beginning in the 1980s. These merit study for their possibility to combine high potential profitability from a national perspective with attractive returns to smallholders’ labor. For the time being, however, Government development strategies discriminate against the emergence of independent smallholder oil palm producers. For example, some provinces will not license palm oil mills unless the enterprise also has its own oil palm plantation or associated smallholders in nucleus estate/smallholder (NES) schemes. This is intended to prevent NES participants from selling their produce outside the project (as happened in the case of rubber) in order to avoid repayment of loans. But not licensing independent mills in an effort to prevent free trade in fresh oil palm fruit also retards development of the market for independent smallholder oil palm producers.

The three contrasting rubber systems produce a wide range of results. First, as already noted, it is encouraging that returns to labor at private prices are virtually identical to the market wage for rubber agroforests planted with seedlings. Although these smallholders are the lowest cost producers of natural rubber in the world (Barlow et al., 1994), returns to land at social prices are not much above upland rice with a long bush fallow rotation and are well below oil palm monoculture.

Perhaps the most striking result in Table IV.3 are the returns to land at social prices for rubber agroforests planted with PB 260 clones, which rival large-scale oil palm monoculture. This system also produces attractive returns to labor at private prices. These data must be treated with caution – which is why they are in italics – since they are based on projections from farmer-managed trials and have not been verified through broader experience by smallholders. The top of the range of profitability estimates might actually be attained by 10-25 % of smallholders (E Penotpers comm.) However, the lower figure in the range represents an expert’s best guess about a ‘worst case’ scenario for yields in this system for the bottom quartile. The big question is where the middle of the profitability distribution would be for this system – and that can only be answered through farmers’ experience. But these results support the idea that potential profitability of rubber agroforests planted with clonal material (and other smallholder agroforests planted with appropriate, higher-yielding germplasm) may be comparable to large-scale oil palm plantation monoculture, at least as long as wages are low.

The profitability estimates for smallholder rubber monoculture planted with GT 1 clonal seedlings provide a cautionary tale to balance the encouraging projections for rubber agroforests planted with PB 260 clones. These monoculture plots were part of a government-sponsored rubber replanting project that was undertaken with high expectations. But the disappointing yields that were obtained because of institutional shortcomings involving supply of planting material, technical information, and credit – these will be taken up in Part V -- could not offset the high costs of that project’s approach. Instead of the high-cost approach in this case of rubber monoculture, the strategy to introduce clones into smallholders’ agroforests seeks a moderate increase in yields at minimal incremental costs. Yet the costly lessons of earlier failures in smallholder rubber development should be borne in mind (Tomich 1991), including difficulty in supplying clonal planting material. The sites studied, for example, were designed to be planted with clones but were actually planted with clonal seedlings because of this problem.

IV.2 Labor requirements indicators

Table IV.4 presents three different indicators of labor requirements. First is total person-days required to establish a system, where ‘establishment’ refers to the period before positive cash flows begin. The two systems with highest potential profitability in the previous section – smallholder rubber agroforests planted with clones and large-scale oil palm—both have very high labor requirements for this phase. However, recall that each system also had high returns to labor. Thus, problems in the labor market or credit market that will be discussed in Part V could impose a serious barrier to adoption, but returns to labor itself is not a problem here.

More generally, returns to labor valued at private prices, which was selected above as an indicator of smallholders’ production incentives, also is a good indicator for smallholders’ concerns with labor constraints if combined with assessments of institutional barriers in markets for labor and capital.

The two other indicators of labor requirements in Table IV.4 are closely related, labor requirements for the operational phase (defined as the period after positive cash flow begins) and total labor. Both measures are averaged over time and the units are person-days per hectare per year.

From the perspective of policymakers concerned with employment generation, total time-averaged labor requirements is a good indicator that is related to equity and stability criteria. Note, however, thatwhile labor-intensive alternatives should be attractive for policymakers who are concerned with job creation, these alternatives will only be attractive to households if they provide attractive returns to labor, the indicator discussed above.

For the rubber and oil palm systems that were evaluated, total time-averaged labor requirements are similar, ranging between 100 and 150 person-days per ha pa. Harvesting labor is the biggest component in these systems. Because of lack of pronounced seasonality in much of Sumatra, harvesting of rubber and oil palm can go on roughly 10 months a year. The two extractive activities – community based forest management and commercial logging – fall at the opposite extreme. Neither of these extractive activities nor the upland rice / bush fallow rotations can provide many employment opportunities compared to treecrop-based alternatives.

Table IV.4 Labor requirements matrix, July 1997

(Total labor inputs for establishment and averages over time for operations and total labor)

Land Use System	Establishment phase (Person-days/ha)	Operation phase (Person-days/ha/yr)	Total Labor (Person-days/ha/yr)
Community - based forest management	na	0.2 - 0.4	0.2 - 0.4
Commercial Logging	15 to 100	17 to 41	31
Rubber agroforest (seedling)	271	157	111
Rubber agroforest (clones)	444	74	150
Rubber monoculture	344	166	133
Oil palm monoculture	532	83	108
Upland rice / bush fallow rotation	na	15 to 25	15 to 25
Monoculture cassava / Imperata cylindrica	na	98 to 104	98 to 104

IV.3 Cash flow constraints indicators

Because perennials are so important among the Sumatran alternatives, our analysis of cash flow constraints focused on multi-year (rather than seasonal) cash flow constraints in order to assess whether the investments required by these systems are barriers to adoption by smallholders. Table IV.5 takes two perspectives on multi-year cash flow constraints: years to positive cash flow and the NPV of establishment costs, which we define as costs prior to positive cash flow. The imputed value of family labor is included in these establishment costs because these labor inputs presumably represent foregone earnings in other activities even if they do not require cash outlay.

By either measure, community-based forest management is the only profitable system without any multi-year cash flow constraints. For the other systems, years to positive cash flow range from 2 years for logging and cassava to 6-10 years for smallholder rubber and 10 years for large-scale oil palm. Time is not a constraint by itself, as evidenced by almost 3 million ha of rubber agroforests that have been planted by smallholders without any formal credit. The NPV of establishment costs at private prices, which is derived directly from the PAM cash flows, probably is the best indicator of cash flow constraints for smallholders. In interpreting these estimates, keep in mind that the existing rubber agroforests are evidence that the Rp 1.3 million required to establish them has not been an insurmountable barrier for smallholders. These estimates suggest that replacing seedlings with higher-yielding clones in rubber agroforests more than doubles investment costs to roughly Rp 2.6 –2.9 million per ha. Since there is no long-term institutional credit for smallholders in Sumatra, whether these investment requirements are barriers to adoption depends in large part on the divisibility of the activity (i.e., is it possible to plant a bit at a time?).

At Rp 8 million per ha, investment costs for large-scale oil palm plantations are the highest of all. Investments of this magnitude would be difficult for many smallholders. But capital costs for large-scale plantations may be inflated for at least two reasons. First, large-scale oil palm plantations formerly received heavily subsidized credit from the Government, which would tend to make them artificially capital intensive. Second, there may have been a tendency among respondents to overstate investment costs in order to mask the profitability of these investments. Even more than rubber, adapting high-yielding oil palm systems as alternatives for smallholders will require research to develop options that are less capital intensive.

Table IV.5 Cash flow constraint matrix, July 1997

Land Use System	Years to positive	NPV of	Years to positive	NPV of
	Cash flow	Establishment cost	Cash flow	Establishment cost
	Private prices	Private prices	Social Prices	Social Prices
	(Years)	(Rupiah / ha)	(Years)	(Rupiah / ha)
Community - based forest management	na	na	na	na
Commercial Logging	2	820,669 to 869,199	2	716,917 to 764,238
Rubber agroforest (seedlings)	10	1,305,536	10	1,477,735
Rubber agroforest (clones)	6 to 7	2,593,458 to 2,862,422	6 to 7	2,950,338 to 3,303,338
Rubber monoculture	10	2,085,257	10	2,192,584
Oil palm monoculture	10	8,041,847	9	8,182,015
Upland rice/bush fallow rotation	never	na	never	na
Cassava / Imperata cylindrica	2	na	2	na

IV.4Household food security indicators

Food nutrient content measures, as in Table IV.6, can be seriously misleading because food security derives from the ability to obtain food, including purchases, and not just capacity to grow it. An unsustainable, low-productivity shifting cultivation system that is suffering decreasing yields because of nutrient depletion and increasing variability in yields because of pest problems may be a riskier basis for securing household food supply than a rubber plot that reliably produces a steady stream of output that can readily be marketed in exchange for rice that trades at a stabilized price.

To accommodate land use alternatives that do not involve foodcrops, our food security indicator is based on Sen’s (1982) concept of risk of food entitlement failure, which encompasses trade-based and production-based entitlements to food as well as security of property rights over productive assets (inheritance and transfer entitlements). Moreover, one of the key dimensions of this analysis is the ‘path’ of food entitlement – is it derived from consumption of one’s own food production, exchange of one’s own production for food, or working for wages to buy food? These ‘paths’ determine the measure of risk of entitlement failure. If the path is production of one’s own food, one simple indicator of production risk is the coefficient of variation of yields.[2] If the path is exchange for food, terms of trade risk must be considered in addition to production risk. A simple indicator of terms of trade risk is the coefficient of variation of the ratio of revenue (price of output times yield) to the price of the staple food, which for Sumatra is rice. This can also be viewed as the coefficient of variation of purchasing power in terms of rice. (Note that if one’s product is rice, the prices cancel out and all that is left is the coefficient of variation of yields, our indicator of production risk.) Finally, if the path is wage labor, risk of entitlement failure is a function of the employer’s financial situation, which is only partly related to production or terms of trade risk. These simple measures do not adapt easily to multiple output systems, such as extraction of non-timber forest products. Although many of these commodities may be important to households’ food and nutritional security, data for food security indicators are not available for NTFPs. Calculations in Table IV.6 indicate that production risk for rubber agroforests may be less than the upland rice/bush fallow rotation. Terms of trade risk for rubber is twice its production risk, as measured by its coefficient of variation. Although these measures suggest upland rice/bush fallow is less ‘risky’ than rubber, the superior production incentives for rubber agroforests are the reason why they have displaced upland rice over the past century.

IV.5 The ‘missing middle’: scaling up from plots to landscapes

Work is needed to expand the assessments of sustainability from plot-level agronomic issues to include environmental externalities at the landscape level and watershed functions. In addition to the two existing study areas in Lampung, the ASB-Indonesia Consortium is planning to have a serious look at the issues of watershed degradation and rehabilitation in the foothill/ mountain zone of Lampung . This is a zone of major conflicts between migrants who are attracted by the fertility of the soils (allowing for coffee production), but who come into conflict with forestry officials who try to maintain this zone as 'protection forest'. This site, together with Mae Chaem in Northern Thailand and Manupali in Mindanao, the Philippines, are the 3 areas that will be the focal points for our regional program’s research on policies and technologies to address environmental externalities at the landscape level.

The policy-driven agenda will require new biophysical insights into landscape-level processes

Of soil and water conservation, as current plot-level insights can not be easily scaled up (Figure IV.1). The Sumber Jaya area, halfway between Krui and the North Lampung ASB benchmark area seems eminently suitable to take up this challenge (see Map 3). In order to complete the landscape transect, it is necessary to expand from the present focus on the peneplains and piedmont agroecological zones in order to include the montane zone and coastal swamps.

[1] In some figures, we will use an alternative measure called the internal rate of return (IRR), which is the discount rate that brings the NPV to zero. The IRR is technically inferior to NPV for assessment of mutually-exclusive alternatives (Gittinger 1982), but using it makes the same point with greater clarity.

[2] The coefficient of variation is the standard deviation in a series divided by the mean of the series. It is a relative measure that expresses variation as a proportion of the average level.

Land Use System	Nutritional Value of Food Produced by the System			Food Entitlement via: Own Production, Exchange, or Wages		Risk of Food Entitlement Failure
	Nutritional Value of Food Produced by the System			Establishment	Operation	Production Risk		Terms of Trade Risk
	Calories: avg kcal /ha/yr	Protein: Avg. kg /ha/yr	Micro-nutrients	Establishment	Operation	Food	Non-food
Community-based forest management	?	?	Important	n.a.	Own prod’n & exchange	?	?	?
Commercial logging	Nil	Nil	Nil	Wages	Wages	n.a.	n.a.	n.a.
Rubber agroforests	118	2.2	?	Own prod’n	Exchange	n.a.	0.13	0.26
Oil palm	19,800	Nil	Nil	Wages	Wages	n.a.	n.a.	n.a.
Upland rice / bush fallow rotation	441 - 490	8.3 - 9.2	Nil ?	n.a.	Own prod’n	0.18	n.a.	n.a.
Monoculture cassava degrading to Imperata cylindrica	9,900	13.6	Nil	n.a.	Own prod’n & exchange	0.06	n.a.	0.22