Project Number:	BGD/91/CO6
Project Title:	Integrated Aquaculture (Duckweed)
UN Cooperating Agency:	UNOPS
Govt. Executing Agency:	Ministry of Fisheries and Livestock
National Implementer:	PRISM (Bangladeshi NGO)
Sector:	Fisheries (440), Crops (420)
Sub-Sector:	Aquaculture
UNCDF Budget:	US$     1,803,000
Gov't. Budget:	39,662
Total Budget:	1,842,662
Actual Total Expenditures at Evaluation:	1,407,933
UNCDF Expenditures, Y/E 1995:	1,407,933
Date Project Approved:	Sept 1993
Date Project Began:	Sept 1993
Date Project Evaluated:	Nov-Dec 1995
Type of Evaluation:	Mid-term

II. Background

Studies have revealed that minimum caloric intake levels are not reached by two-thirds of rural Bangladeshis. Fish account for more than 80% of the modest amount of animal protein intake, and its consumption is falling at an alarming rate. Even maintaining current supplies requires rapid sector expansion. Primary constraints to a national increase of fish production, in an effort to keep up with population growth, are the lack of extension, credit, and local fish feed production due to competition for Bangladesh's scarce nutrient biomass.

Duckweed is an aquatic plant that is native to Bangladesh and which grows in brackish ponds and ditches. It has potential use as fish and animal feed and can also be used as a natural waste filter. Since 1989, a local NGO, PRISM, has operated a pilot project to demonstrate the potential of duckweed on lands donated by the Kumudini Welfare Trust, sharing the property with a hospital. The duckweed there is used to purify hospital wastes and is also cultivated to provide fish feed to local pilot fish ponds. Fish production levels in these ponds have been high, and, as a result, UNCDF assistance was requested to expand the integrated duckweed/fish technology to two sites in Mirzapur and Shibaloy. Another local NGO, Projects in Agriculture, Rural Industry, Science, and Medicine (PRISM) was identified for implementation.

III. The Project

In addressing the constraints to increased fish production, the project provides extension services, credit, and affordable fish feed through a replicable integrated system of small-scale intensive fish farming and duckweed production. A primary goal is to develop and test suitable organizational forms to support local production and dissemination of this technology among the rural poor. The project was implemented by PRISM, with approval of the Ministry of Fisheries and Livestock. The project was approved and startup began in September 1993 with a planned five-year duration. The total cost is $US 1,842,662; UNCDF funding amounts to $1,803,000, with the remainder supplied by the Government of Bangladesh (GOB).

The promotion of fish and duckweed production, to be managed on an integrated basis by small groups averaging ten members, was central to the project's design. These groups were provided with loans (averaging US $580 per member) for capital investments (earthwork, company registration, and shallow tubewells for extra dry season water supplies) and for working capital (inputs, labour, cost of leased land, etc.). Each consolidated fish pond/duckweed venture was established as a joint stock company, with shares and equity points allocated based on contributions of land and/or labour (Poorer members without land are assigned equity points based on an assessed value of labour contributions). The project set a 75% membership goal for landless and/or marginal farmers, and a 50% participation target for women.

IV. Purpose of the Evaluation

An early mid-term evaluation was scheduled, after two years of a five-year planned project. The evaluation team was asked to review overall project design, performance, and efficacy, and to make recommendations for changes in project strategy, scale, scope and implementation arrangements.

V. Findings of the Evaluation

A. Assessment of Results Achieved

• Annual fish yield targets were achieved; duckweed production targets were not. This reflected an overemphasis on fish production over the production of duckweed fish feed. In practice, not enough pond surface area was devoted to duckweed production to meet the needs for fish feed. As a result, farmers still rely heavily on purchased fish feeds, diminishing the project's impact on fish production costs.

• A sample of operating groups showed that in the two project areas, 40% in one and 50% in the other of the pond operators were landless or marginal farmers. These figures fell below project goals.

• Only 11% of operators were women, which also was well below project goals.

• Duckweed technology had a very positive effect on sanitary conditions: installation of pour-flush latrines with pond outflows provided nutrients and high growth levels to the duckweed. Duckweed water purification reduced odors and water-borne insect populations.

• Sales of duckweed in the project areas indicate a natural market for duckweed fish feed, which can develop into an added income source and a more efficient use of scarce land and biomass production. A more independent duckweed market may also allow for more flexibility at the levels of groups and the market, de-linking poorer duckweed ditch farmers from the wealthier fish pond farmers who currently control demand.

• Despite high production levels, village enterprises had very low "return to capital investment" ratios compared to other fish pond projects in Bangladesh. Net return on investment in the sample villages ranged from low to negative. This was due to the very high investment costs for the ponds.

• The joint stock company model for fishery enterprise development created unequal relationships between the poorest and least poor farmers, based on land contributions and resultant company shares. In addition, the allotment of "sweat equity" shares was much lower than expected. This put the invested asset of the poorest farmers — their land — at an unacceptable risk level, exacerbated by the high infusions of credit.

B. Assessment of Project Design

• The basic strategy to link duckweed and fish production is viable, and is being replicated elsewhere in the country. The sanitation technology design was particularly successful, and an initial aversion to excrement-based technology has been overcome by the experience of duckweed ponds receiving latrine outflow, where odors and water-borne insect populations have been reduced.

• Despite this limited success, however, the evaluation determined that the project had serious design flaws. Sweat equity share allocation methodologies were poorly planned, and the availability of company shares to the poorest members was too limited, diluting their ownership stake. The primary development objective of raising the income levels of the poor was directly at risk, because of the potential for land loss under the joint stock company strategy.

• Project design and implementation strategies placed too little emphasis on duckweed production in relation to fish farming.

• The use of large credit infusions to finance pond construction in Bangladesh was unnecessary, where Food for Work (FFW) and other subsidies were available for capital investments to ponds. Further, the high investment contributed to excessive project costs.

• One of the project sites was an area known to be flood-prone, a serious investment risk.

• The joint stock company enterprise development strategy was counterproductive to the project's development goals, and placed unnecessary burdens on the members.

VI. Recommendations

An extended evaluation debriefing was arranged at UNCDF headquarters, and included two evaluators and the implementing agency, PRISM. The following key recommendations were made as a result of the evaluation:

• Technology. There should be no new pond development in the flood-prone areas. Pond criteria should include the ability to maintain a sufficient year-round water level to avoid costly pumps. Earthwork investments should be minimized, and World Food Programme assistance for pond rehabilitation should be sought. Supplementary fish feed purchases should be minimized, and movement toward exclusive use of duckweed for fish feed should be stressed where feasible; expansion of nutrient supply to duckweed through flush latrines should be strongly encouraged. The project should also make every effort to record and analyse production data to establish optimal parameters for duckweed production technology.

• Institutional Framework. No new joint stock companies should be formed. The project should shift to an informal group structure based on similar social and economic conditions among the membership, largely separating groups of men and women; record-keeping responsibilities of the groups should be reduced to reflect existing capacities. Legal and financial consultants should be engaged to propose effective strategies for the existing companies, including debt burdens and the general assessment of company benefits by their members. The new groups, or individual group members, may undertake production of fish, duckweed, or both in their own ponds or in leased ponds, so long as the ponds meet the selection criteria. Design goals for beneficiaries to participate in the project could be maintained, but only if clearer guidelines to screen participants are identified.

• Credit. If the recommendations to maximize participation by the poorest sectors of society and to reformulate the groups on less formal basis are to be met satisfactorily, the credit program will have to be redesigned, both in terms of overall and individual loan size. The adequacy of current remaining loan funds should be assessed based on expected loan repayments and the impact of possible loan write-offs, and the determination of new requirements for less capital-intensive technology.

VII. Policy Implications and Lessons Learned

• All future project goals that include asset development and commercialization should fully consider the realities of both formal and informal economic sectors. When the informal sector holds development potential, goals to increase its income generating capacity may be more secure than goals to formalize sector activities. In the case of BGD/91/C06, application of the experimental joint stock company model ran counter to much of the experience and learning in Bangladesh regarding organizational innovation.

• Development strategies that pool assets stand the risk of exacerbating economic differences because they create inequity at different levels based on the initial level of investment. Furthermore, when asset development is attempted among the poorest of the poor, an unacceptable risk can be incurred. Any low investment ventures with secure and relatively high returns should be encouraged as tools to increase productivity among the poor, as opposed to high investment, high production models.

• Village enterprise development ventures should, in general, pool common levels of risk and investment among their members. This, however would require a degree of economic and/or social homogeneity among the membership. Furthermore, any level of formalization should reflect existing capabilities among the members, so as not to impose onerous costs and responsibilities.

• In developing new technologies (in this case, the linkage of fish and duckweed production), sufficient attention should be given to extension work and support. Extension support overshadowed by cumbersome administrative duties often creates conditions where problems, lessons, and project corrections can be easily overlooked.

VIII. Evaluation Team

The mid-term evaluation mission team consisted of Monawar Sultana, a socio-economist and team leader; and Roy Jenson, aquaculture specialist/marine biologist.