1. Tree crops have played a central role in Sri Lanka's development since the 19th century, when tea and rubber were introduced to the island. At that time coconuts were already a long-time staple, grown primarily by smallholders. Sri Lanka is currently the world's largest exporter of tea and one of the world's five largest exporters of coconut kernel products, such as desiccated coconut. While Sri Lanka's output of natural rubber is modest compared to its main competitors, export duties on rubber have been a major source of earnings for the government (12 percent of revenues in 1979).
2. Until the late 1970s, Sri Lanka adopted a generally inward-oriented development strategy that entailed increasing government intervention in the economy. The tree crops sector had long been an exception to the inward-oriented policy, accounting for more than three quarters of Sri Lanka's export earnings by 1979. However, heavy taxation of the sector's earnings, coupled with repeated threats of nationalization and declining commodity prices, led to inadequate investment in replanting. By the late 1970s, the actual and potential output of tree crops had been seriously eroded by years of neglect. Furthermore, new challenges had emerged after 1972 as a result of the nationalization of all landholdings over 20 hectares. Two government-owned plantation corporation - the Janatha Estates Development Board (JEDB) and the Sri Lanka State Plantation Corporation (SLSPC) - gained control of more than half the area under tea cultivation, 30 percent of the area under rubber and about 5 percent of the area under coconuts.
3. GOSL recognized these challenges and sought external assistance to reverse the decline in the sector. The Bank's overall strategy supported the government's policy of gradual trade liberalization, deregulation and privatization. However, since the privatization policy did not extend to JEDB and SLSPC, in 1978 the Bank embarked on a series of four projects that aimed to rehabilitate tree crops and diversify production on government-owned estates. Taken together, the four projects accounted for 43 percent of Bank-wide lending for tea during the last two decades. These four projects aimed at government estates were complemented by two projects targeted at smallholder rubber growers. This Impact Evaluation Report (IER) focusses on the first Smallholder Rubber Rehabilitation project (Credit 1017-CE) and on the Fourth Tree Crops project (Credit 1562-CE).
4. An IDA credit of US$16 million equivalent was approved for the first Smallholder Rubber Rehabilitation (SRR) project in May 1980. The objectives of the project were to improve production and enhance earnings for rubber smallholders by supporting the replanting of over-aged and degraded stands, in-filling of gaps in viable stands, rehabilitation or expansion of group processing centers, and institutional strengthening of the Rubber Control Department (RCD) and Advisory Services Department (ASD) of the Ministry of Plantation Industries (MPI). A small component was provided for the Rubber Research Institute of Sri Lanka (RRISL) and funding was provided for impact evaluation surveys and case studies by the Agrarian Research and Training Institute (ARTI). Most of the IDA credit was to be channelled to the government's Rubber Replanting Fund (RRF), which provides grants to smallholders for replanting rubber. While grant payments are recovered in part via a cess on exports of rubber, cess revenues have not been sufficient to cover the RRF's expenditures (paras. 2.39 ff).
5. The Fourth Tree Crops (FTC) project was supported by an IDA credit of US$55 million equivalent, approved in March 1985, and cofinanced with the Asian Development Bank (ADB); the Governments of the Netherlands, Norway and the United Kingdom; and the Bank of Ceylon, a state-owned commercial bank. The objective of the project was to reverse the decline in tea, rubber and coconut production on public sector estates operated by JEDB and SLSPC, two of the world's largest plantation corporations at that time. This objective was to be achieved by improving estate productivity through field development (replanting and infilling of degraded and over-aged areas under tea and rubber, replanting and under-planting of areas under coconuts, and diversification into fuelwood in areas considered unsuitable for tree crops), development of field nurseries to produce replanting materials, factory rehabilitation, strengthening the corporations' management and financial control capabilities, and improving housing, health and social amenities for estate workers through respective project components, collectively referred to as the Social Welfare Program (SWP).
6. The SRR project was closed in June 1988, 2 1/2 years later than planned. About US$4.9 million of the credit was canceled, primarily due to depreciation of the Sri Lankan rupee (LKR), which increased the local value of IDA funds beyond what was required to implement the project. The final project cost was US$19.6 million, compared to an SAR estimate of US$28 million. The project got off to a good start but ran into trouble in 1983 as a result of drought conditions and a shortage of replanting materials. The appraisal objective of replanting 18,800 hectares was achieved at project closing in 1988, notwithstanding the need to uproot 1,500 ha of rubber in the project area due to an attack of Corynespora leaf spot (a fungal disease).
7. Civil unrest led to implementation delays in the FTC project, which was closed in December 1991, 1 1/2 years late. Actual costs were US$237.7 million, compared to US$211.8 million at appraisal. Much larger areas of tea and rubber were replanted than had been planned, whereas in-filling targets for tea were not met. This was due to the fixed cost nature of labor costs and the existence of surplus labor, for which estate superintendents attempted to create employment. The excessive replanting rate outpaced the flow of counterpart funds, leading to the reallocation of most Bank funds earmarked for nursery development and institutional development to the field development component. In addition, senior Bank management approved an increase in IDA's disbursement percentage for the field development component from 35 percent to 90 percent. Government intervention in labor matters precluded effective direct bargaining between the unions and plantation managers and resulted in substantial wage increases and employment guarantees that reduced efficiency and drove up costs. This is largely attributable to pressure from labor unions representing the largely Tamil estate workforce. The unions acquired considerable political clout as a result of the extension of voting rights to Tamils of Indian origin in the early 1980s and of the government's security concerns resulting from the ethnic conflict in Jaffna. While progress was made under the project on housing and health care for estate workers, conditions for estate residents are still poor, compared with those of the rural population.
8. Virtually the entire area replanted under the Smallholder Rubber Rehabilitation project (18,800 ha) is now mature, although the earliest replantings are just reaching full development. By 1999, it is estimated that production will reach 19,530 tons/year, equivalent to 22 percent of average annual output by private producers during the project period, 1981-88. Yields are expected to peak at 1,080 kg/ha, 15 to 17 years after replanting. These figures are the same as the PCR's, but are 20 percent lower than appraisal projections.
9. Both the SAR and the PCR calculations assume no replanting in the absence of the project, so that all production at full development is incremental. But given the age of rubber stands, there would clearly have been some replanting even without the project. The impact evaluation estimates that production would have reached only 9,080 tons by 1999 in the absence of the project, due primarily to funding constraints on the RRF and to more limited outreach by the extension services. The SRR project's net impact on production is therefore estimated at 10,450 tons/year.
10. At appraisal of the SRR project, economic rates of return were expected to reach 23 percent (paras. 2.17 ff). At project completion, ERRs were re-estimated at 18 percent, mainly because of reduced yields and lower prices. A further unforeseen decline in rubber prices coupled with a sharp upswing in real wages for tappers have reduced the re-estimated ERR for the project to 4.8 percent.
11. Re-estimated financial returns (FRRs) on replanted rubber were attractive at impact evaluation (14.4 percent) for smallholders who used family labor for tapping. However, FRRs were below ten percent for smallholders who relied on hired labor for tapping, i.e. about half of all smallholders with less than four hectares of land (paras. 2.35 ff). These returns are lower than estimates for FRRs for various farm models at appraisal (11 to 21 percent). One reason for the decline is real wage inflation prompted by the growth of Sri Lanka's textiles and services industries. Another reason is lower than expected yields. Eighty percent of the project area was replanted with a low-yielding clone (PB 86) that had been superseded in Malaysia in the 1960s. This was a serious error. The problem of insufficient clonal material of acceptable quality was compounded by poor cultivation practices. A further reason for the low returns is the low quality of smallholder-processed rubber. While the project attempted to address this via co-operative-style Group Processing Centers, the GPCs experienced numerous logistical problems and failed to attract members (para. 2.38).
12. Replanting grants and intercropping with bananas raised FRRs on rubber investments to acceptable levels. However, only one in eight farmers planted intercrops and most growers, especially those with less than one hectare, failed to maintain proper cultivation practices and were therefore ineligible for the full grant.
13. Under the Fourth Tree Crops project, annual production of tea on government estates is expected to reach 185,200 tons by the year 2014, compared to estimated without-project production of 141,300 tons as a result of no replanting, for an increment of 43,900 tons/year. The SAR and PCR estimated incremental production at 47,900 tons/year and 39,700 tons/year, respectively. At impact evaluation, yields on replanted areas are expected to average 2,125 kg/ha made tea at full development (12 years after replanting), which is fairly standard for vegetatively propagated (VP) clonal teas, and is consistent with SAR and PCR estimates.
14. With respect to rubber, annual production on government estates is expected to peak at 65,800 tons in the year 2002, compared to without-project production of 49,200 tons/year. The incremental production of 16,600 tons/year at full development is explained by a larger replanting program and by higher yields than in the absence of the project (para. 2.12). The project's net impact on production is unlikely to reach appraisal projections of 20,100 tons/year at full development.
15. With regard to coconuts, annual production on government estates is expected to reach 57.1 million nuts/year at full development of the FTC project (in 2006). Without a replanting program, production on the estates would have declined to 49.6 million nuts/year in 2006. Thus the FTC project will result in incremental production of up to 7.5 million nuts/year at full development. This is less than half the appraisal estimate of incremental production, 18 million nuts/year, due to major shortfalls on both yields and area replanted.
16. Economic rates of return at appraisal of the FTC project were projected to reach 28 to 31 percent for tea, 24 to 28 percent for rubber and 18 percent for coconuts. The ERR for the project was estimated at 26 percent. The PCR re-estimated the project ERR at 10 to 15 percent. Rates of return were found to be 10 to 11 percent for tea and 11 to 15 percent for rubber. The PCR did no calculations for investments in coconuts, since these accounted for less than 5 percent of project costs and information was inadequate. The decreases in ERRs for tea and rubber were due to higher production costs and lower output prices than foreseen at appraisal. The impact evaluation revealed a further modest decline in returns. Revised ERRs range from 8.2 percent for tea to 7.4 percent for rubber and only 2 percent for coconuts. The ERR for the project (excluding unquantified welfare benefits and related costs) is estimated at 7.9 percent (paras. 2.17 ff).
17. The financial performance of the two government corporations has been dismal. JEDB lost money every year after Bank approval of the FTC project, while SLSPC reported a profit only twice. Both companies became increasingly reliant on external borrowing and on overdrafts from government-owned commercial banks for day-to-day expenditures. For example, JEDB's debt-equity ratio soared from 0.8 in 1982 to an unsustainable 9.8 in 1989. "Creative accounting techniques" were employed to make the financial position of SLSPC appear less dire (para. 2.25). In 1990, the two corporations had to be bailed out at a cost of LKR 3.6 billion. Since then, the tree crops sector has been a major burden on the fisc, as GOSL's expenses on the estates have risen while export duties on tea and rubber and the ad valorem sales tax on tea have been phased out (para. 2.32).
18. The main reasons for the losses at JEDB and SLSPC were depressed world prices, stagnant productivity and rising costs of production. The latter two were attributable to internal factors, including restrictions on the movement and use of labor, political patronage and mismanagement of the estates. Project investments added to the corporations' short term liquidity problems, while yielding unattractive returns in the longer term: FRRs at impact evaluation range from 8.3 percent for rubber to 0.4 percent for coconuts.
19. The FTC project included social welfare objectives that were largely achieved (paras. 3.11 ff). After appraisal, beneficiary surveys were conducted and social welfare services were identified and prioritized on the basis of the resulting "Needs Assessments". Beneficiary participation was high in the construction of latrines and re-roofing of houses. These items were not included at appraisal but were identified as high priorities by estate residents. Appraisal targets with respect to dispensaries and health centers were exceeded. These successes are clearly associated with some of the improvements in health and demographic indicators for estate residents that occurred during the project period.
20. Privatization became the most important institutional issue for the FTC project, although it was not an objective at appraisal. Following the mid-term review of the project in September 1988, the Bank encouraged GOSL to establish a Core Group to solve the growing financial crisis at JEDB and SLSPC. This led eventually to the establishment of a Task Force in 1990 with a broader mandate, namely to consider the "...appropriate structure, organization and management of the two corporations.." (para. 3.36). The Task Force met with considerable resistance, due to the political sensitivity of the privatization issue.
21. The Bank was the main source of external guidance on the privatization process, but its objectives and approach differed significantly from GOSL's (para. 4.10). In particular, the Bank supported the transfer of ownership to the private sector, including to foreign firms, so as to attract skills and capital. The government favored and finally adopted a management contract approach to privatization that relied on management contracts with no equity stakes for management companies and no foreign participation. GOSL wanted to retain control of the assets and ensure that employment would not be adversely affected. What emerged in 1992 was a highly restrictive management arrangement with 22 private Sri Lankan Estate Management Companies (EMCs, para. 4.27). The winners were identified via open bids in April 1992 after more than 150 local and foreign firms had expressed an interest. The 22 EMCs signed management contracts with 22 new, wholly government-owned Regional Plantation Corporations, which were given control over most of JEDB's and SLSPC's 500 estates.
22. In January 1992, the Bank withdrew its intended support for a follow-on to the FTC project, because GOSL was not prepared to make a time-bound commitment to privatizing the ownership of the estates or to awarding long term leases to the EMCs as a second-best solution. In practice, private management has reduced losses on the estates and increased efficiency. However, the private sector has been unwilling to inject funds into state-owned companies in which it has no equity share and with which it only has five-year management contracts. Furthermore, GOSL interference in wage decisions has persisted. Thus costs have not been reduced sufficiently, nor has private capital been attracted to the sector. These problems have led to rising debts that are undermining the sustainability of the current arrangement.
Findings at Impact Evaluation
23. Both projects are rated as unsatisfactory at impact evaluation. To be sure, they were implemented under challenging circumstances, arising from an ethnic conflict that erupted in the north of the island in 1981 and from violent political confrontation in the south in the late 1980s. The FTC project was particularly affected by these conflicts, both indirectly via the northern conflict's effects on wage determination and directly via the murder of 25 estate superintendents following the southern outbreak. Furthermore, during 1980-92 world prices declined for all three tree crops.
24. Successes in other sectors, notably the high rates of growth achieved in the labor-intensive textiles and services sectors, resulted in labor shortages on rubber smallholdings and drove up the cost of labor. The labor issues on government estates were somewhat different: a blend of ethnic and political factors contributed to high wage increases, constraints on the movement of labor between estates, year-round employment guarantees, low work norms, declining productivity and severe underemployment of estate labor, particularly in the mountainous tea-growing areas in the centre of the island. In addition, political and cultural factors have militated against reforms to a highly inefficient work rule for tea pluckers, who are almost exclusively women. They are required to work eight hours to receive the daily wage even if they can complete their assigned task (and more) in less than eight hours. Male workers, on the other hand, are free to leave as soon as they finish their assigned task and are still guaranteed a full day's wage.
25. The price and labor factors, combined with low yields due to weak management, outdated cultivation and processing practices, and poor quality inputs, undermined the expected returns under the two projects. Notwithstanding improvements in cultivation practices on government estates under private management, these factors continue to make the sustainability of project benefits uncertain.
Lessons at Impact Evaluation
26. The tree crops sector is likely to prosper if and only if appropriate policy measures are taken. Past approaches have failed to be sufficiently comprehensive or structured. For example, the FTC project was part of a medium term investment program that did not address pressing labor concerns. The ultimate objectives should be to: i) generate sustainable growth, and ii) improve the welfare of those who earn a living from the tree crops sector. These ultimate objectives can only be achieved by galvanizing the sector via improved technology and increased factor productivity, and by competing more effectively in international markets, in ways that are specified below. Success with regard to these intermediate objectives in turn requires a significant injection of capital, investment in human resources, exploitation of the industry's global expertise, and elimination of inefficient practices.
27. The analysis in this impact evaluation suggests that these objectives cannot be attained under government management of the industry, nor will the restrictive private management arrangement currently in place suffice. A coherent set of measures should be considered and adopted by GOSL to create an enabling environment in which the private sector can flourish (see the lessons below). The government's full commitment to this approach is as yet uncertain, but will be essential in order to restore domestic and external investors' confidence and provide investors with appropriate incentives. The Bank and other donors can and should continue to play an important supporting role in the fulfillment of these objectives. These policy proposals are consistent with the intentions communicated in a Policy Framework Paper for 1994-96, that was prepared by GOSL in collaboration with staff of the International Monetary Fund and the World Bank. Finally, the proposals are consistent with the lessons of experience drawn from the evaluation of the SRR and FTC projects. These are summarized as follows:
(a) Support of Viable Tree Crops Projects. In the event of sharp movements in prices and costs during implementation, the Bank should re-assess economic returns to ensure that: i) the focus is maintained on development objectives rather than just on implementation indicators; ii) projects that become inviable are restructured or suspended in a timely manner without saddling the borrower with additional debt.
(b) Support of Viable Rubber Replanting Schemes. The Bank should only support a cess-based, accelerated replanting program for tree crops if- i) the clonal material to be replanted is of an internationally acceptable standard and has been thoroughly tested; ii) the replanting grant is adjusted regularly to reflect changes in investment costs and guarantee a minimum acceptable level of support; and, iii) cess rates are adjusted to reflect changes in the level of grants.
(c) Privatization as a Process. A process of privatization should: i) be transparent; ii) enjoy unwavering commitment from the government; iii) be open to foreign investment and expertise; iv) encourage participation by all stakeholders; v) secure popular backing, rather than just the support of vested interests; and, vi) publicize the positive case for privatization.
(d) Attracting Private Capital. The privatization strategy adopted to date fails to provide the private sector with adequate incentives to invest its capital and is therefore unsustainable. The experience of the past two years suggests that privatization will only succeed if private managers are given the option of purchasing a substantial share of the equity of the corporations. This will: i) enable them to borrow and invest against their shares; ii) provide incentives to invest in long-term projects and manage the estates accordingly; and, iii) provide them with some down-side risk so as to discourage excessive risk-taking.
(e) Criteria for Management Contracts. The design of the management contracts adopted thus far has limited their effectiveness. The contracts are short-term and grant private managers little operating autonomy. If management contracts are to be used for tree crops plantations, they should be given for an extended period of time, e.g. 50 years, and should provide substantial operating autonomy to private managers in order to encourage long-term planning and investment. Even then, they must be regarded as second-best arrangements for attracting private capital.
(f) The Policy and Regulatory Environment. i) the elimination of export taxes was a sound policy decision that should be maintained, together with an exchange rate policy that takes into account its effects on the tree crops sector; ii) other regulatory reforms warrant consideration, e.g. the removal of restrictions on direct sales outside the auction system, in order to encourage vertical integration, the development of brand names, price stabilization through forward contracts, and reductions in the lag between production and disposal of tea.
(g) Labor Wages and Productivity. i) Wages and benefits should be subject to collective bargaining between managers and unions; ii) specific contractual commitments are required to safeguard private sector investments against arbitrary administrative decisions on labor matters; iii) jobs can only be preserved if productivity is increased, e.g. by allowing workers to move from labor surplus to labor deficit areas and by renegotiating work norms that have been set at unduly low levels.
(h) "Putting People First". i) Needs assessments and beneficiary participation were key elements in the success of the social welfare program; ii) the underemployment of estate youth is an acute problem requiring urgent attention, e.g. via training and promotion of off-farm opportunities; iii) women's work norms as tea pluckers result in compulsory overtime that is inefficient and lowers women's effective wage rates. Therefore the best practices applied in tea growing countries in Eastern Africa should be piloted and then, if successful, applied more widely on Sri Lankan estates.