Kirindi Oya Irrigation and Settlement Project

7/31/2019 Kirindi Oya Irrigation and Settlement Project

1/53

ASIAN DEVELOPMENT BANK PPA: SRI 19025

PROJECT PERFORMANCE AUDIT REPORT

ON THE

KIRINDI OYA IRRIGATION AND SETTLEMENT PROJECT(Loans 324-SRI[SF], 612-SRI[SF], and 794-SRI[SF])

IN

SRI LANKA

December 2000


2/53

CURRENCY EQUIVALENTS

Currency Unit - Sri Lankan Rupee/s (SLRe/SLRs)

At Appraisal At Appraisal At Appraisal At Project At OperationsSupplementary

Loan

Phase II Completion Evaluation

(30 Sep 1977) (30 Sep 1982) (30 Sep 1986) (30 Jun 1995) (31 Mar 1999)SLRe1.00 = $0.116 $0.0478 $0.0357 $0.0200 $0.0150

$1.00 = SLRs8.61 SLRs20.91 SLRs28.05 SLRs49.95 SLRs66.80

ABBREVIATIONS

ADB Asian Development BankDOA Department of AgricultureEIRR economic internal rate of returnha hectareID Irrigation DepartmentIFAD International Fund for Agricultural DevelopmentIWMI International Water Management InstituteKfW Kreditanstalt fr Wiederaufbaukm kilometerOEM Operations Evaluation MissionOFC other field cropsO&M operation and maintenancePCR project completion reportt tonTA technical assistanceWUG water user group

GLOSSARY

Yala Southwest (dry) monsoon season (approximately April to September)Maha Northeast (wet) monsoon season (approximately October to March)

NOTES

(i) The fiscal year (FY) of the Government ends on 31 December.

(ii) In this report, $ refers to US dollars.

Operations Evaluation Office, PE-XXX


3/53

CONTENTS

Page

BASIC DATA iiEXECUTIVE SUMMARY ivMAP vi

I. BACKGROUND 1

A. Rationale 1B. Formulation 1C. Purpose and Outputs 1D. Cost, Financing, and Executing Arrangements 2E. Completion and Self-Evaluation 2F. Operations Evaluation 2

II. PLANNING AND IMPLEMENTATION PERFORMANCE 3

A. Formulation and Design 3B. Achievement of Outputs 3

C. Cost and Scheduling 4D. Procurement and Construction 5E. Organization and Management 5

III. ACHIEVEMENT OF PROJECT PURPOSE 7

A. Operational Performance 7B. Performance of the Operating Entity 8C. Economic Reevaluation 9D. Sustainability 10

IV. ACHIEVEMENT OF OTHER DEVELOPMENT IMPACTS 10

A. Socioeconomic Impact 10B. Environmental Impact 11C. Impact on Institutions and Policy 12

V. OVERALL ASSESSMENT 13

A. Relevance 13B. Efficacy 13C. Efficiency 14D. Sustainability 14E. Institutional Development and Other Impacts 14F. Overall Project Rating 15G. Assessment of ADB and Borrower Performance 15

VI. ISSUES, LESSONS, AND FOLLOW-UP ACTIONS 16

A. Key Issues for the Future 16B. Lessons Identified 17C. Follow-Up Actions 18

APPENDIXES 20


4/53

BASIC DATAKirindi Oya Irrigation and Settlement Project

(Loans 324-SRI[SF], 612-SRI[SF], and 794-SRI[SF])

Project Preparation/Institution Building1

TA No.2

Name Type Amount($)

ApprovalDate

168-SRI Lunugamvehera Reservoir (Irrigation andAgriculture Development) PPTA 49,000 7 Jul 1976

407-SRI Kirindi Oya Irrigation and Settlement PPTA 50,000 23 Jul 1981730-SRI Kirindi Oya Phase II PPTA 75,000 9 Dec 1985

Key Project Data ($ million)As per ADB Loan Documents

324-SRI 612-SRI 794-SRI Total ActualTotal Project Cost 66.80

328.00 33.10 127.90 93.40

Foreign Exchange Cost 24.00 12.00 10.20 46.20 43.10Local Currency Cost 42.80 16.00 22.90 81.70 50.30Cancellation 4.004 2.51 5 11.13 17.64 17.64

Funding Proposed ActualSource 324-SRI 612-SRI 794-SRI Total 324-SRI 612-SRI 794-SRI Total

Borrower 42.80 4.40 6.50 53.70 6.50 5.86 3.65 16.01ADB 24.00 10.00 26.60 60.60 20.00 9.47 15.48 44.95IFAD 6.00 6.00 12.00 4.06 16.06KfW 7.60 7.60 13.30 3.04 16.34Total 66.80 28.00 33.10 127.90 51.80

622.43 19.13 93.36

Key Dates324-SRI 612-SRI 794-SRI

Appraisal 12-31 Aug 1977 16 Mar-3 Apr 1982 2-20 Jun 1986

Loan Negotiations 26 Oct-3 Nov 1977 11-15 Oct 1982 22-25 Sep 1986Board Approval 9 Dec 1977 9 Dec 1982 30 Oct 1986Loan Agreement

- ADB 25 Jan 1978 17 Dec 1982 24 Nov 1986- IFAD 20 May 1978 8 Dec 1982 - KfW 19 Sep 1979 7 Apr 1983

Loan Effectiveness (ADB) 28 Apr 1978 21 Apr 1983 26 Feb 1987Loan Closing

- In Loan Agreement 31 Dec 1985 30 Jun 1986 30 Jun 1991- Actual 29 Jan 1987 13 Oct 1989 10 Jan 1995- Number of Extensions 1 4 3

Months (effectiveness to completion)- Original Project 92.0- Total Project 200.5 (+ 118%)

1

Only the report on TA 730 is retained in ADBs records with the date of May 1986.2

= not applicable, ADB = Asian Development Bank, IFAD = International Fund for Agriculture Development,KfW = Kreditanstalt fr Wiederaufbau, PPTA = project preparatory technical assistance, TA = technical assistance.

3 Includes $15.6 million equivalent in foreign exchange entitlement certificates (which have been abolishedsubsequently) and $1.38 million in customs duty. A major portion of the local cost has subsequently been financedby KfW and IFAD, at $13.3 million and $8.0 million, respectively.

4$4 million was cancelled from ADBs loan account when cofinancing from IFAD became available.

5 Consisting of $0.53 million under IFAD funding and $1.98 million under KfW funding.6

Estimated by project cost established at reappraisal for the supplementary loan, less identified financing gap.

ii


5/53

Key Performance IndicatorsAppraisal

324-SRI 612-SRI 794-SRI PCR7

PPAREconomic Internal Rate of Return (%) 17.6 11.0 13.6 6.3 2.6Cost

- New Irrigation Only ($/ha) 5,925 18,528 13,363 18,336 17,245- New and Existing Irrigation ($/ha) 3,852 8,919 8,648 9,931 9,340

Irrigation Area- New Irrigation (ha) 8,409 4,200 8,409 5,416 5,416- New and Existing Irrigation (ha) 12,934 8,725 12,934 10,000 10,000

Borrower Government of the Republic of Sri Lanka

Executing Agency Irrigation Department (lead agency)

Mission Data (No. of Persons/Person-Days)8

Type of Mission 324-SRI 324/612-SRI 612-SRI 794-SRI PCR/PPARPreappraisal 1/9Appraisal 1/19

Reappraisal 4/44 1/10Inception/Review 2/18Fact-finding 1/9Project Administration

Inception 1/16Review 8/72 10/143Reformulation 1/17Special Loan Administration 6/50Project Completion 1/15

Operations Evaluation 1/10

7 PCR = project completion report, PPAR = project performance audit report, ha = hectare.8

Summarized from PCR.

iii


6/53

EXECUTIVE SUMMARY

The Kirindi Oya Irrigation and Settlement Project was intended to develop underutilizedland in the dry zone of Sri Lanka. Its main objectives were to increase food and fiber cropproduction and create gainful livelihood for people barely covering their subsistence needs. TheProject included the construction of the countrys largest earthfill dam with a gated spillway and

canals to irrigate 8,400 hectares (ha) of newly developed command area. Additional waterwould allow cropping intensity to increase from about 135 to 200 percent on an existing area ofabout 4,600 ha irrigated from traditional reservoirs (tanks). With reliable irrigation, yields wereexpected to increase considerably.

The 1977 feasibility study assembled the considerable preparatory work undertaken bythe Government, supported by limited technical assistance from the Asian Development Bank(ADB). From the options considered by the Government over time, the one proposed under thefeasibility study was driven by political expediency. It was the best prepared option and theGovernment wished to demonstrate its concern for peoples welfare in an economicallydepressed area that had recently experienced civil disturbance. The feasibility study terms ofreference restricted analysis to the Governments specific request. Against that background, the

study was more detailed than many from that era. It estimated the cost of the Project at$112.2 million of which irrigation development would absorb $75.3 million and the settlementcomponent $14.3 million. The study drew attention to the high development cost ($7,150/ha in1977 dollars) and to other key factors for successful development. The estimated economicinternal rate of return (EIRR) for the option selected was 10.8 percent.

ADBs original appraisal differed significantly from the feasibility study. Project cost wasreduced to 60 percent of feasibility study estimates while the proposed new irrigation areaincreased by 30 percent. The EIRR consequently rose to 17.6 percent. Both the reduced costand the expanded area proved optimistic and a supplementary loan was required in 1982 due tocost overruns. Phase II was approved in 1986 without change to the planned irrigation area andwas reformulated in 1991. At physical completion in 1994, the area of new irrigation totaled

5,400 ha, about 85 percent of feasibility estimates and 64 percent of the appraisal target. Withthe reduction in new area development, the feasibility study cost estimate was also close toactual.

The implementation period was more than double the 7.7 years originally envisaged,due to (i) delays in the commencement of major civil works, (ii) the effort required to obtain theadditional funding needed to complete the project, and (iii) insufficient contractor capacity due tothe demands of the large civil works program under the Accelerated Mahaweli DevelopmentProgram. Political disturbances also disrupted work progress occasionally. Institutionalconstraints were increasingly noted and compensated for by provision of additional consultants.

After a difficult period in the late 1980s and early 1990s, the irrigation scheme is

performing better. Yields have averaged almost 4 tons/ha despite a number of dry years suchas 1999. Cropping intensity has varied between 114 percent and 200 percent over the past10 years, with an average of 171 percent since 1992 compared with the design target of189 percent. Farmers have focused on rice production with less development of other fieldcrops than envisaged, though banana areas are expanding. Annual crops are grown on areasonable scale in drier years, though often without irrigation.

A total of 4,924 families were settled under the Project. These included 1,450 familiesdisplaced by the dam and irrigation development with the balance selected from other areas of

iv


7/53

southern Sri Lanka. Appropriate settlement infrastructure, including safe drinking water, wasprovided. For many farmers resident in the new irrigation area, family income remains low with asurvey conducted for the Governments project completion report indicating an average incomeof $135 per capita in 1994, just below the $140/capita poverty line (in 1991 prices). However, inabsolute terms their income has improved somewhat, particularly in the new irrigation areas.Malnutrition continues to be a severe problem in the project area, particularly for mothers and

children.

The environmental record is mixed with some significant disbenefits, notably a reductionin salinity of the ecologically valuable coastal lagoons. However, the human environment withinthe command area is now aesthetically attractive compared to adjoining slash and burnagricultural areas, while malaria incidence is reported to have fallen significantly.

Project economic performance has been poor due to (i) implementation delays; (ii) highproject cost; (iii) below target development of new land; and (iv) reduced cropping intensity,particularly on the new areas due to limited water availability. Both the livestock and forestrycomponents, added in Phase II, performed poorly compared to target. The EIRR is estimated at2.6 percent.

While irrigation infrastructure is mainly being maintained by the Irrigation Department,water user groups will need to take more responsibility for scheme operation and maintenance.However, the main threat to sustainability comes from irrigation, agriculture, and humansettlement development in the catchment above the dam.

With the notable exception of the settlement component, the Project overall hasperformed below expectation, although it remains relevant to Sri Lankas developmentobjectives. However, its relevance is reduced somewhat by the high cost and relatively smallnumber of beneficiaries. Cost-effectiveness and implementation efficiency were low. The Projectsuffered from institutional problems and had little impact on institutional development. Althoughthe Implementing Agencys accounting systems improved, project monitoring remains

inadequate. Overall, the Project is rated less than successful.

Aspects requiring particular attention in the future include (i) the threat to projectsustainability due to upper catchment degradation and increasing use of water for irrigationupstream of the dam, which are reducing reservoir recharge and irrigation water availability;(ii) irrigation management, with improvements in this regard likely to have positive impacts onincome, nutrition, and the environment (requiring significant sociocultural change on the part ofthe irrigation administration as well as the farmers); and (iii) settlement, with a need to allocateland titles to resident farmers and address nutritional problems in the project area and widerregion.

v


8/53


9/53

I. BACKGROUND

A. Rationale

1. The Kirindi Oya basin lies within Sri Lankas dry zone, which includes 70 percent of thecountrys land area. At the time of project preparation in the 1970s, the zone included

30 percent of Sri Lankas cultivated land and 25 percent of its population. Opportunities wereseen for more balanced development and improved land use in the zone, to increaseagricultural production, reduce imports, and limit population pressure in higher rainfall areas.

B. Formulation

2. Beginning with a reconnaissance report on the areas natural resources in 1956, theGovernment undertook a series of studies on the potential for development in the Kirindi Oyabasin. Together with topographical surveys in 1973 and 1974, these studies provided the basisfor preparation of an Asian Development Bank (ADB) loan1 supported by technical assistance(TA).2 The subsequent feasibility study prepared by the Government was completed in June1977. The study included substantial information and assessed the technical and economic

feasibility of a design proposed by the Irrigation Department (ID). ADBs limited TAsupplemented the considerable preparatory work done by the Government. The original loanwas approved by ADB in December 1977, with a supplementary loan approved in December1982 and an extension, referred to as Phase II, in October 1986. Further small amounts of TAwere used for the preparation of the supplementary loan3 and the formulation of Phase II.4

C. Purpose and Outputs

3. The Project was intended to increase rice and other field crop (OFC) production,generate employment, save foreign exchange, and use underutilized land resources. The scopeof the Project closely followed the design requested by the Government and defined in thefeasibility study. The main components were (i) the construction of a 5-kilometer (km) earthfill

dam with sluiced spillway and gates, (ii) rehabilitation of the 3,675-hectare (ha) Ellegalairrigation system which was irrigated from a series of small traditional reservoirs (tanks) andimproved water supply to the 850 ha Badagiriya system, (iii) development of new irrigation on8,410 ha, and (iv) settlement of 8,320 families including 333 households from the reservoir areawho would be given land for settlement and irrigation together with supporting infrastructure.Agricultural extension services were to be strengthened. The scope of the Project remainedessentially unchanged over three appraisals, apart from the deferment of some of the irrigationimplementation to a second phase and adding relatively small components for rural credit,social forestry, livestock, and marketing. Reformulation in May 1991 reduced the developmentof new irrigation by about 3,000 ha due to limited irrigation water availability and irrigationmanagement problems. Details of the original scope and achievement by completion are givenin Appendix 1. Cropping intensity on the existing irrigation areas was expected to increase from

153 to 200 percent and to reach 177 percent on the new areas, with 4,400 ha planted to cotton.The inclusion of cotton and other subsidiary field crops was intended to support the national

1 Loan 324-SRI(SF): Kirindi Oya Irrigation and Land Settlement Project, for $24 million, approved on 9 December

1977.2

TA 168-SRI: Lunugamvehera Reservoir (Irrigation and Agriculture Development), for $49,000, approved on 7 July1976.

3 TA 407-SRI: Kirindi Oya Irrigation and Settlement, for $50,000, approved on 23 July 1981.4

TA 730-SRI: Kirindi Oya Phase II, for $75,000, approved on 9 December 1985.


10/53

2

policy of promoting crop diversification and limit irrigation water demand on the large areas ofpermeable soils to be developed under the Project. However, due to problems experienced byother projects in establishing cotton, plans for the crop were dropped under the supplementaryloan.

D. Cost, Financing, and Executing Arrangements

4. ADB provided $20.0 million and $10.0 million, respectively, for the original andsupplementary loans, to finance the direct and indirect foreign exchange costs. Due to ADBsrestriction on financing local currency costs, the Government sought cofinancing with assistancefrom ADB. The International Fund for Agricultural Development (IFAD) provided loans of$12.0 million and SDR5.5 million ($6.0 million equivalent), administered by ADB.5 The Germanagency Kreditanstalt fr Wiederaufbau (KfW) financed $13.3 million and $7.6 million,respectively, under parallel cofinancing. Phase II was funded by ADB alone, which by that timewas able to fund local currency costs and provided a loan of $26.6 million. The residual costswere borne by the Government and are estimated at $6.5 million for the original Project, $5.9million for the supplementary phase, and $3.7 million for Phase II. All funding was providedunder preferential terms, in the case of ADB from the Asian Development Fund resources.

Appendix 2 gives details of project financing. Project cost estimates varied greatly over thecourse of design and implementation as discussed in Appendix 3. At the time of Phase IIappraisal, total project cost was estimated at $113 million, of which the dam and associatedstructures accounted for 34 percent and irrigation development 15 percent.

5. Over the life of the Project were four executing agencies: (i) ID was responsible for damconstruction and irrigation development; (ii) Land Commissioners Department for settlement,livestock and dairy development, and social forestry; (iii) Department of Agriculture (DOA) foragricultural support services; and (iv) Central Bank of Sri Lanka for the credit program throughselected participating banks.

E. Completion and Self-Evaluation

6. The Project, as reformulated in 1991, was completed in early 1995. The Governmentengaged the services of the International Water Management Institute (IWMI), based inColombo, to prepare a project completion report (PCR). IWMI was familiar with the Project andprepared a comprehensive report based on available project data and a sample survey of479 households. ADB undertook a PCR mission in June 1995, generally accepting the findingsof the IWMI report, and adding information from ADB records. The Operations EvaluationMission (OEM) confirms much of the information in both the Governments and ADBs PCRs ascorrect. However, the economic internal rate of return (EIRR) of 6.3 percent reported in the PCRis considered optimistic due to unrealistic rice yields and expectations of real rice priceincreases (discussed in Appendix 4). The PCR rated the Project partly successful.

F. Operations Evaluation

7. This project performance audit report presents the findings of an OEM that visited SriLanka in March 1999 supported by a review of ADB files, records, and reports. These includedappraisal reports, loan agreements, minutes of loan and TA committee meetings as well as thestaff summaries of Board discussions, the PCR, missions back-to-office reports, and other

5

Including $4 million for foreign exchange costs cancelled from ADBs originally approved loan of $24 million.


11/53

3

reports by consultants and IWMI. Information was sought from ADB staff involved in the Project.Considerable time and effort were spent in obtaining key documents that could not be locatedwithin ADB, such as the original feasibility study and the report on the appraisal undertaken byKfW about one year after ADBs appraisal. Only the last of the three TA reports could belocated. Copies of the draft project performance audit report were provided to the Borrower, theExecuting Agencies, and ADB staff concerned for review. Their comments assisted in finalizing

the report.

II. PLANNING AND IMPLEMENTATION PERFORMANCE

A. Formulation and Design

8. At the time of formulation, a number of development options for the Kirindi Oya area hadbeen debated for many years by irrigation engineers and local politicians. Alternatives rangedfrom small-scale projects with small reservoirs in various locations, to major schemes involvingwater transfer between river basins. Although much of this analysis was at the reconnaissanceor conceptual stage, consideration of alternatives during project preparation might have defineda more cost-effective solution than the design selected by the Government and eventually

implemented. For example, according to the Governments PCR the much-discussed decisionabout the dam location was apparently motivated by the need to maximize the new area to bebrought under irrigation rather than by consideration of what would be the most cost-effectiveapproach. The Lunugamvehera location required building a much larger dam than sitesupstream would have required, thus greatly raising the cost. In the light of the low water inflowssubsequently experienced, this proved to be a costly decision.

9. The terms of reference for the original project preparatory technical assistance studyrequired the consultants to develop the project concept proposed by the Government. This wasintended to demonstrate the Governments concern for an impoverished area that hadpreviously faced serious civil unrest. ID engineers were also attracted by the challenge ofbuilding the largest earthfill dam in Sri Lanka. The feasibility study drew attention to aspects

that warranted further assessment such as the high permeability of soils in most of thecommand area, the lack of experience in the production of crops other than rice, and potentialproblems with continuous flow water management. The option recommended (with newirrigation on 6,500 ha) yielded an EIRR of 10.8 percent. The feasibility study was well preparedwithin the limits set and generated a high level of ownership by the Government.

10. The appraisal mission adopted the design proposed in the feasibility study. However, thearea to be developed for new irrigation was increased by 30 percent, which the feasibility studyhad deferred pending further study. More significant was a large reduction in project costs. The1977 feasibility study had estimated investment cost at $90 million excluding transfer payments.The appraisal assessed the project cost at $50 million to fund an increased irrigation area. Inpractice, project unit costs exceeded feasibility study estimates. The feasibility study and

appraisal estimates and actual expenditure are given in Appendix 3. The EIRR of the Project atappraisal improved to 17.6 percent due to the reduced costs and increased benefits.

B. Achievement of Outputs

11. The Project was largely implemented as planned. The earthfill dam and a total of 51 kmof main canal (10 km less than initially planned) were constructed. The main divergence fromappraisal relates to the reduction in new irrigation development from 8,400 to 5,400 ha. Thereduction resulted from overoptimism at appraisal and declining water inflows to the reservoir,


12/53

4

due to a reported decline in rainfall, increasing water use in the catchment area, and catchmentdegradation. Due to the reduced irrigation area, the number of families settled fell from 8,320 to4,900 and construction of other settlement infrastructure was reduced in proportion (Appendix1). The quality of infrastructure inspected by the OEM was generally satisfactory. However,because the design did not anticipate the water shortages experienced in the dry season, waterregulation structures are not ideal and operation is labor intensive. Tree planting in woodlots

and homelots under Phase II was about one third of target. The livestock component alsoexperienced implementation difficulties and did not achieve its targets in relation to livestockdistribution or the development of milk collection and marketing facilities.

C. Cost and Scheduling

12. The total cost of the Project is estimated at $93.4 million or about 83 percent of therevised budget of $112 million (Appendix 2).6 In constant value terms, the project cost increasedby about 53 percent over estimates of the original appraisal, reflecting (i) the reduction in projectcost at appraisal over feasibility study estimates; (ii) the lack of contractor competition;(iii) increased costs due to inadequate maintenance of construction equipment; (iv) someincreases in cost due to inadequate information at appraisal; and (v) the addition of social

forestry, agricultural credit support, agricultural support services, and livestock developmentcomponents valued at about $4.5 million under Phase II. The Government is estimated to havecontributed $16.0 million (17 percent of total project costs) while ADB contributed $44.95 million(48 percent), IFAD $16.1 million (17 percent) and KfW $16.3 million (18 percent). Project costsare compared to appraisal estimates in Appendix 3. Joint cofinancing with IFAD functionedsmoothly with loan administration delegated to ADB. For the parallel cofinancing with KfW, itwas not always easy to keep informed of the status of procurement and disbursement, anddifficulties arose in scheduling joint reviews.

13. Implementation took nearly 17 years from loan effectiveness compared with eight yearsexpected at appraisal. The major delay related to construction of the dam, which was completedin 1986, four years behind schedule. The new irrigation area was to be progressively developed

with the main canals completed one year after completion of the dam and the distributionsystem one year after that. The process of selection of settlers and their settlement was to startwith the completion of the dam. This sequence of events was logical and could have beenfollowed if more preparatory engineering work and investigation had been undertaken atfeasibility study stage and if contractor performance could have been relied on, particularly inthe area of construction management. The original schedule assumed a degree of coordinationamong various agencies that proved unattainable.

14. Many of the 3,950 settlers of Phase I were given their land allocations several yearsprior to the supply of irrigation water. This process is termed advanced alienation, agovernment policy designed to discourage squatters from occupying tracts of land earmarkedfor development. These settlers had to be sustained by food assistance under the World Food

Program for longer than anticipated due to implementation delays. The first new tracts of the

6

Costs have been reestimated based on the ADB Controllers records for ADB and IFAD loans and from informationobtained from KfW and ADB records on the government contribution. In the earlier years with the noteddeficiencies in the accounting function of ID, advances may not have been fully liquidated or leases for IDequipment not fully deducted from progress payments. This may result in a minor overstatement of project costs.Records on KfWs parallel disbursements are incomplete. However, information has been received on loanutilization and cancellation. There is less information on government expenditure which includes food assistancefrom the World Food Program during the early years of settlement and system operation.


13/53

5

command area were irrigated in 1987 and the last from 1994. Implementation of the livestock,woodlot, and credit components continued until loan closing in January 1995, with targets notmet.

D. Procurement and Construction

15. A number of implementation problems were experienced. ID, the lead ExecutingAgency, had no experience in implementing ADB-financed projects and faced the complicationof parallel cofinancing under the KfW loan. Both loans required departures from governmentprocurement procedures. Limited English language capability at the field level also limited IDsability to follow ADB procedures. However, the factor that most severely affected thecompetitive bidding process was the implementation of the Mahaweli Ganga development underthe newly created and powerful Mahaweli Authority. Mahaweli was given priority in theallocation of financial resources and attracted the best contractors and qualified engineers fromother agencies because of higher pay scales and benefits. The limited national contractingcapacity for large-scale projects was stretched over a large number of civil and engineeringworks with reduced competition and increased prices. The shortage of qualified staff led to slowimplementation.

16. These factors caused particular difficulties for construction of the damthe Projectslargest civil works contract. Three contractors expressed interest in bidding, with two submittingbids, both significantly higher than IDs engineering estimate.7 The lower bidder finally declinedto extend bid validity, while ADB was trying to impress on ID the need to include furthertechnical specifications in the draft contract. This left the River Valley Development Board as theonly bidder. The Board had been the Executing Agency for the Walawe Irrigation Project, thefirst ADB-assisted irrigation project in Sri Lanka, but had lost this status and become agovernment-owned contractor. Many of its qualified engineering staff had moved to theMahaweli Authority and, at the same time, it was required to absorb workers from othergovernment organizations. ID signed the contract with the Board before ADB had reviewed orapproved the final draft contract. In the interest of making up for delays already incurred, ADB

acquiesced in the absence of competition and compliance with its guidelines. For a number ofsmaller contracts, ADBs objections were respected, for example when award of contract wasproposed to other than the lowest qualified bidder. With few exceptions, the quality of work wassatisfactory. Most of the defects noted on some minor civil works and on land leveling havebeen corrected. All civil works were tendered under local competitive bidding.

E. Organization and Management

17. ID staff performed satisfactorily in engineering design. The task of supervisingcontractors posed difficulties in an environment of institutional rivalry and divided loyalties, asituation in which ADB could provide little assistance. Reporting was weak with senior stafffluent in English needing to allocate their scarce time for other tasks. Project accounting was

complicated, with procurement under three different lending agencies and the Governmentbudget. Being experienced only in ordinary accounting for budgetary releases in local currency,ID staff encountered major problems. The first report by the Auditor General was prepared in1984 for the years 1979-81 and contained many qualifications.

7 According to the KfW appraisal, only three Government-owned contractors had the capacity to undertake a

contract of that magnitude.


14/53

6

18. Project coordination was complicated by the number of executing agencies, particularlyafter 1981 when responsibility for the provision of infrastructure and associated facilities for landsettlement under the Project was passed to the Land Commissioners Department. The projectcoordination committee was encumbered by having two project managers, one for irrigationappointed by ID and one for settlementthe additional district government agentwith thelatter acting as chairman. ID remained the lead Executing Agency at the national level, but was

limited to irrigation at the project level resulting in implementation problems.

19. ADB fielded 36 missions in support of the Project, an average of two per year.Preappraisal, appraisal, fact-finding, and reappraisal totaling eight missions helped withadjusting the project funding and scope to increasingly difficult circumstances. The initialpreappraisal and appraisal missions involved only 28 staff days (involving seven specializationsduring appraisal), which seems to reflect overconfidence in dealing with a new project in anuntested institutional environment. Three reappraisal missions for the supplementary loan wereconducted over a span of one year, again with a large number of staff but of short duration. Asimilar pattern is noted for the review missions prior to Phase II. Many of these missions werehelpful in resolving specific difficulties in engineering design and in providing support in thestruggle to adjust cost from the reduced levels accepted for appraisal. The six special loan

administration missions assisted in resolving backlogs in withdrawal applications. With theapproval of the Phase II loan, reviews changed from crisis management during Phase I to amode of regular monitoring with fewer staff spending more time in the field on each mission.This led to reformulation in 1991 on the basis of a good understanding of project developmentand finally to its closing in 1995.

20. Most of the consulting services required for project implementation were financed underADB and other donor loans. They related to the construction work undertaken in Phase I up tothe end of 1986 and included (i) an international consulting firm for design and constructionsupervision funded by ADB; (ii) a team of project implementation consultants for aspects of thedam and irrigation development funded by KfW; and (iii) a specialist geotechnical service forspillway investigation, design, and construction supervision. Inputs totaled 425 person-months,

almost four times the initial 112-month estimate. The project implementation support was usefulin supplementing the capability of ID and the principal contractor. The expertise in geotechnicalservices and related spillway design features was essential in ensuring the integrity and safetyof the dam. In general, the provision of international expertise has drawn greater attention tosafety features. As in many projects, it proved difficult to achieve a lasting institutional impactfrom the consultant services provided. The efforts made in water management planning andoperations, irrigation agronomy, and systems operation and maintenance (O&M) have onlypartly been institutionalized within ID.

21. Domestic consulting services costing $575,000 were provided under Phase II. The majorportion was used for an accounting firm to upgrade IDs project accounting following earlierattempts that had not achieved the desired results. Project benefit monitoring was undertaken

by the Agrarian Research and Training Institute, which undertook a baseline survey and severalspecial studies. Lack of interest and motivation by the implementing agencies made this effort oflimited value in contributing information for project management. IWMI undertook a projectimpact assessment, on which ADBs PCR was based. The study included analysis of an April1994 survey and was comprehensive. Minor assistance was extended to the LandCommissioners Department on project programming and agricultural marketing. Advisory TAs


15/53

7

were also undertaken by IWMI to assist in improving irrigation management on the Kirindi Oyaand Walawe schemes.8

22. A standard covenant of the loan agreement for Loan 324-SRI(SF) Section 4.01, statesthat the Borrower shall cause the Project to be carried out with due diligence and efficiency andin conformity with sound administrative, financial, engineering, agricultural and land

development practices. Overall project implementation performance and outcomes wouldsuggest that this covenant was in practice not fully complied with. No detailed seasonalmonitoring information is available, as required under the Loan Agreement, and DOA was notable to provide the required level of irrigated crop extension. Progress reports were infrequentand lacking in substance while financial reporting was much delayed and inadequate,particularly under Phase I. Slow progress was made to recover housing loans withSLRs1 million recovered by September 1999, against a total outstanding of SLRs11.5 million.Grants had been issued on 8,150 allotments out of the total of 9,660. The collection of irrigationservice fees was replaced by agreement under the ADB-supported Agriculture Program Loan9

to establish and assist in the development of farmer organizations.

III. ACHIEVEMENT OF PROJECT PURPOSE

A. Operational Performance

23. Irrigation was improved for the existing irrigation area and new irrigation was developedon 5,400 ha (64 percent of target). The reduction in area was necessitated by uncontrolledirrigation leading to high per hectare water use, and lower than expected inflows to thereservoir. Cropping intensity on the new irrigation area has averaged about 134 percent. Overallcropping intensity has averaged about 146 percent over the past 14 years though with asignificant increase in the last few years (to 170 percent for the period 1993-1999). Estimates ofirrigation efficiency (water used for plant growth as a proportion of water released from thereservoir) have varied. However, analysis by IWMI in 1997 suggests irrigation efficiency to be ata low 28 percent compared with 44 percent planned.10 Because farmers in the old (Ellegala)

area claim prior water rights, supply restrictions are applied mainly to the new area farms.

24. Available data indicate rice yields in the wet (maha) season of about 4.0 tons (t)/ha andin the dry (yala) season 3.8 t/ha, which is close to the average expected at appraisal of 4.0 t/ha.OFCs occupied an average of about 9 percent of irrigated area including, in declining order,green gram, cow pea, groundnut, vegetables, chili and, more recently, banana. Bananasmallholdings are now being established on part of the command area (about 400 ha in1999/2000) with financing from Seylan Bank and government support through tissue culture andextension/research. As a substitute for rice, most OFCs are grown in situations of decliningcertainty of irrigation water supply. They are often grown on residual moisture from the previousrice crop, with limited or no irrigation. Their area has fluctuated between 50 and 1,860 ha of thetotal area, influenced by price volatility as well as by irrigation water availability. They have not

established a firm place in the cropping pattern. Water availability has frequently been low in thedry season, resulting in a cropping intensity of only 72 percent from 1992 to 2000. Despite often

8

TA 846-SRI and TA 1480-SRI: Study of Irrigation Management and Crop Diversification, Phases I and II, for$350,000 and $750,000, approved on 16 January 1987 and 19 February 1991, respectively.

9 Loan 994-SRI(SF): Agriculture Program Loan, for $80 million, approved on 28 November 1989.10 Kirindi Oya System Management Improvement: A Strategy to Achieve 200% Irrigation Intensity. Project proposal to

ID. IWMI, Colombo, December 1997.


16/53

8

severe irrigation water deficits, water is treated as if plentifully available and the flow isinadequately controlled.

25. No account was taken of livestock under Phase I. According to the Government PCRsince the emphasis of the Project was on settler-based crop production, about 1,800 ha ofscrub used for grazing were cleared and a large number of irrigation tanks that provided

drinking water and wallowing sites for cattle and buffalo were leveled. These activities had anadverse impact on the livestock husbandry practiced by herdsmen in the old area. No thoughtwas given to these pervasive problems at the planning stage of the ProjectMost of theproblems faced by herders today are the result of this negligence.

26. Phase II recognized the importance of livestock farming and included a number oflivestock components. Most of these were unsuccessful, notably the cattle and buffaloupgrading program and the milk collection, processing, and marketing component. The formerfaced low demand for stock, while the latter could not compete with private milk companies. TheProject planned to construct about 20 milk collection centers, a mini processing factory, and twosales centers for milk and dairy products. However, by project completion, the centers were onlycollecting around 6 percent of milk produced, and this was sold to the two private milk

companies. Little incremental milk production resulted from the Project, and milk production inthe area has declined, with total production from the Project area in 1994 estimated at2.6 million liters as against the incremental milk production planned under the Project of 4.1million liters. However, cattle and buffalo continue to play an important role in the regionalagricultural economy. Within the irrigation area, they cause considerable damage to canals andcrops, leading to conflict between herders and cultivators and frequent claims for compensation.Although it was planned to develop 1,100 ha for livestock grazing and agroforestry under PhaseII, this has made not made much contribution to livestock production and little evidence of theintended production system could be located by the 1994 survey team. However, a good rate oftree planting was reported for the woodlots and along 36 km of roads.

27. Settlement infrastructure is operated and maintained adequately. The piped water

supply system, roads, schools, clinics, police stations, and post offices are much appreciated bythe settlers. The piped water supply is unique among land settlement schemes in Sri Lanka andviewed as greatly improving living conditions. Where service is fairly reliable, householdsgenerally pay their water feeswhich cover O&M costsand are keen to invest in individualmetered household connections. Total cost at nearly $4.0 million (25 percent of land settlementcost) was high but appears to be a good investment in support of future semi-urbandevelopment.

B. Performance of the Operating Entity

28. Many of the farmers who were relocated from the reservoir area (333 households) andfrom the new command area were shifting cultivators who had resided in the area for a number

of years. Annual farm income from rain-fed subsidiary crops and, infrequently, a small area ofrice grown during the rainy season with water from small local reservoirs, was estimated in thefeasibility study at $100 (updated to 2000 dollars) with additional earnings of $35 from casuallabor on rice farms. For the existing rice farmers drawing water from the large tanks of theEllegala system or the separate Badagiriya tank, annual household income was estimated at$430 with no off-farm earnings.

29. The farm budgets developed for the OEMs economic reevaluation of the Project aregiven in Appendix 4, Table A4.12. In the with-project situation, an irrigated 1 ha farm in the old


17/53

9

Ellegala irrigation area could annually earn around $700 and a new area farm about $610. Thisdifference is less than at project completion, due partly to the introduction of bananas to thecropping system combined with more reliable rainfall in the last few years. Without the Project,overall cropping intensity on an irrigated Ellegala farm is estimated at 128 percent, generatingannual household earnings of $270. This level is less than the $430 estimated at appraisal dueto the worsening terms of trade for rice. Average earnings per day worked are estimated to have

risen from around $3.5 to $5.0 under the Project.

30. The drinking water supply system is expected to recover its O&M costs. According tosystem managers, this is accomplished by the threat of disconnection in the case of delayedpayment of water dues. Others interviewed by the OEM were less positive about the level ofcost recovery. Overall, the water supply system is performing reasonably well, with consumerswillingness to pay seen in the context of reliability of service.

31. Cost recovery for irrigation, initially combining investment and O&M cost as covenantedin various loan agreements, could never have been implemented successfully. Settlers havegenerally not been prepared to meet even full O&M cost, since they consider this to be agovernment responsibility, and many lack the financial resources to make a cash contribution.

The frequent lack of water in the dry season has limited settlers financial performance and thustheir capacity to pay. Since completion of the Project in 1994, the Government has thereforeattempted to provide sufficient funds to meet O&M requirements of the main structures andcanals. The OEM observed some signs of delayed maintenance with slumping canal banks anddeferred control of vegetation, but the system remains functional. Farmer groups areresponsible for maintaining their field distribution system, largely by the provision of voluntarylabor. The result is affected by the strength of water user group (WUG) leadership and theproportion of nonresident farmers who usually do not contribute. Again, the systems seen by theOEM remain functional, even though maintenance may be less than optimal.

C. Economic Reevaluation

32. EIRR as reestimated by the OEM is 2.6 percent, including the investment in irrigationand half the agricultural research and extension, and land settlement costs (see Appendix 4).Inclusion of the entire land settlement cost would reduce the EIRR to 2.3 percent. The low rateof return reflects a higher project cost than initial estimates, the shortfall in area developed, andlow cropping intensities. It also reflects the delay in generating benefits. If upstream waterretention and diversion are not controlled, there is a risk that the targeted cropping intensitieswill not be realized, further reducing the EIRR.

33. The EIRR in the feasibility study was estimated at 10.2 percent, increasing to17.6 percent shortly thereafter at appraisal. It went through a number of declines, namely 11percent at Phase I extension, 8 percent (for the whole Project) at the time of approval for PhaseII in 1986,11 and 6.3 percent in the PCR. The economic analysis presented in the original

appraisal was weak and no supporting tables were presented. ADBs PCR reported the resultsof the economic analysis conducted by IWMI as part of the Governments PCR in 1994. Whilethe analysis in the latter is detailed, it contains serious flaws including overestimation of riceyields and price. Modifying the PCRs analysis for these factors reduces the EIRR to a levelclose to that calculated by the OEM. Details are provided in Appendix 4 Section E.

11

On a sunk cost approach, 19 percent was computed for Phase II alone.


18/53

10

D. Sustainability

34. The main threat to long-term sustainability, at the time of the OEM, was lower thanpredicted river flows. Their initial effect was to reduce the planned area of new irrigationdevelopment and to limit cropping intensity, particularly in the new areas. During the 14 yearssince dam completion, flows have been erratic and generally low. In part, this was due to a

change in rainfall patterns compared with earlier years, but the greatest impact has come fromdevelopment that is upstream of the reservoir (see Box). Lower demand for water is, however,likely following restoration by ID of the Mau Ara reservoir and diversion of its waters to theMalala-Ara basin, which supplies to the 850 ha Badagariya irrigation system fromLunugamvehera.

Recent Developments in the Kirindi Oya Basin

Rapid development is taking place in much of the Kirindi Oya basin, due to a huge influx of population.Many people settle by the road that follows the stream. Irrigation development is evident on each side ofthe river with large-scale pumping. Over five years, the number of pumps has increased from about200 to around 2,000, due mainly to the ready availability of cheaper pumps and PVC (polyvinylchloride)pipe. Other major developments include the construction of many large (6-9 meter) diameter wells andthe repair, rehabilitation, and new construction of small tanks (dams) within the catchment. It is estimatedthat around 150 tanks are in operation, each irrigating an average of 30 ha. Farmers mainly irrigatevegetables, especially high value capsicum, onions, and sugarcane. Numerous diversion weirs are beingconstructed and towns draw their water from the river. Land use is changing from scrub to settledagriculture where bunded fields also reduce runoff. Conversion of forested hill areas to pine plantationand vegetable farming is changing runoff characteristics. Overall, the watershed is undergoing quite rapiddegradation with a marked effect on streamflow. Of 100 m

3of water emanating from the headwaters, only

about 10 m3

on average reach the reservoir. In future, with the present rate of development, even thiswater may be lost. This has implications for project sustainability.

Source: Based on Success Story of Yala 1999 Paddy Irrigation in Kirindi Oya. October 1999. Peter Droogers et. al.IWMI.

35. The other key area affecting sustainability is the level of O&M. At present, ID is primarilyresponsible for this, though it is attempting to pass responsibility for scheme maintenance toWUGs in line with national policy and budget availability. Substantial further work is required todevelop user-funding of scheme O&M. However, for the time being, scheme maintenance,though severely underfunded, is sufficient to prevent major operational problems. All structuresthat the OEM visited were adequately maintained, apart from some canal damage. Schememanagement has adopted a firm attitude to unauthorized modification of irrigation structures(e.g., to increase water supply to a particular distributary). When evidence of modification isfound, management has reported that water is cut off until the damage is repaired. In this way,project management has so far managed to minimize damage to the irrigation system (a causeof great problems to some other irrigation schemes).

IV. ACHIEVEMENT OF OTHER DEVELOPMENT IMPACTS

A. Socioeconomic Impact

36. The Project has had a significant impact on poverty in the project area. Farm families,including both original Ellegala farmers and new settlers (but particularly the latter), havebenefited from project investment in irrigation, village, and communication infrastructure. Whiledelays occurred between settlement and first irrigation water release, food distribution under theWorld Food Program reduced hardship. A majority of settler families are likely to be better off


19/53

11

now than in the early days of settlement or before that, particularly the slash and burn farmersfrom the project area and the unemployed. However, poverty and malnutrition continue to bemajor problems in the project area, leading to stunted growth, often resulting from nutritionaldeficits or intestinal infections that start after weaning. Many preschool children demonstrate ahigh incidence of low weight for age and/or low weight for height.

37. At the time of the impact evaluation survey in March 1994, average income among theEllegala farm families was SLRs5,566 per month. New area settlers were less well off withfamily incomes averaging just over SLRs3,000 per month. With an average new area family sizeof 5.6, this equated to an annual income of SLRs6,500 per capita or around $135 at theprevailing exchange rate of SLRs48/$.12 At this level, at least the bottom 50 percent of new areasettlers were likely to be below the poverty line. Nonetheless, a striking feature of the area is thenumber of new houses that have been built and a dynamic economy, with active markets andtransport services for example. The level of affluence that has been achieved, patchy though itis, can largely be attributed to the Project.

38. The provision of public services has significantly changed the living conditions of settlersresiding in the project area. Where previously children were sent to stay with relatives for

schooling, they can now attend local schools. Literacy, especially for girls, has improved to ahigh level. A health clinic close to the dam site operates effectively. The open spaces createdhave reportedly reduced the prevalence of malaria. Safe drinking water contributes toimproved health, though other ailments due to unsanitary conditions and intestinal infectionsremain serious problems. Project road infrastructure facilitates interaction with surroundingareas for purchases and sales. All settlers met during the OEM acknowledged these benefits.Even so, there are some who continue to resent their relocation due to the Project. Newcommunities with settlers from different locations take time to become cohesive. The largenumber of nonresident landholders does not help in this process.

39. Gender issues were not discussed at project design. However, the Project has led to anumber of positive impacts on women. In particular, the piped water within easy walking

distance or connected to the house saves time and effort in obtaining drinking water, whichtraditionally has been womens work. To a lesser extent, the establishment of woodlots hasreduced the time spent on gathering fuelwood. Improved health care and reduction in malariarisk take some of the worries away from women who do most of the nursing of sick familymembers.

B. Environmental Impact

40. No comprehensive environmental impact assessment was required when the originalloan was prepared in 1977. Subsequent reports and the project impact evaluation study (theGovernments PCR) drew attention to several environmental issues and impacts. The Projectconverted over 4,000 ha of land with soils and development potential similar to the soils found in

the command area into a reservoir. Aside from secondary jungle and scrub (with the timberpotential reportedly already exploited at feasibility stage), the reservoir area contained ninehamlets or villages, two minor tanks irrigating 35 ha of rice for one season, and 565 ha ofshifting cultivation. The population in the reservoir area was estimated at feasibility stage at1,400, living in 333 households. Two ruins of Buddhist temples, Rakingala and

12 The poverty line defined in the World Banks 1994 Sri Lanka Poverty Assessment was $140 per year in 1991

prices.


20/53

12

Digawapichettiya, were known to be located in the reservoir site. These had not beeninvestigated by archeologists but were not considered significant by the ArcheologicalDepartment. The inundation of the reservoir area somewhat reduced habitat for wildlife in theadjacent national park. On the positive side, the reservoir supports a substantial fishery withannual landings estimated at 1,300 t. The water from the reservoir also supplies the pipeddrinking water system. Anecdotal evidence suggests that the reservoir and the land clearing for

command area development have led to or coincided with a reduction in the prevalence ofmalaria.

41. The operation of the reservoir has had a number of environmental consequences, someforeseen at feasibility stage, others not. Even with controlled irrigation water distribution, a riskwas seen that drainage water would affect the brackish water lagoons and bird sanctuariesdownstream. Appropriate diversion of drainage water was suggested. Free flow management,that is typical of irrigation practice in Sri Lanka, has aggravated the problem. The Embilikala andMalala Lagoons have effectively changed into fresh water lagoons due to excessive flow ofdrainage water. This has destroyed the environment for brackish water shrimp from which manyresidents in the area made a living and on which birds in the adjacent bird sanctuary depended.The shrimp fishermen had to change to less lucrative fishing for freshwater fish. Diversion of the

flow of drainage water to restore the lagoons to their original state would be an expensiveundertaking and a corresponding cost of the Project.

42. Changes in hydrology led to a temporary increase in salinity in irrigation water.Underground saline sediments became dissolved (the main reason for the non-feasibility ofwells for villages at that time). Salinity in parts of the newly developed irrigation area peaked inthe 1989 wet season. Salinity was recognized as a sufficiently serious problem for the project-supported adaptive research station to place special emphasis on salinity-tolerant varieties andcrops. However, since the wet season in 1993, soil salinity has been found only in small areaslocated adjacent to natural drainage lines. More serious complaints with regard to salinity havebeen made in the preexisting Ellegala irrigation system. Measurements of salinity wereundertaken at biweekly intervals and, up to 1993, showed considerable variation in salinity. The

cause was thought to be seepage from the left and right bank irrigation canals, which carriedsalt into the tanks forming the Ellegala system. Another view is that silt with soluble salts,deposited over many years at the bottom of the tanks, was disturbed with the increasedthroughput of water. Whatever the reason, the problem has gradually diminished over time andit is now a problem only in small areas in seasons with low availability of irrigation water.

43. The Project has reduced the grazing areas for herds of cattle and buffalo, while irrigationcanals disrupt their movements from one area to another. As a consequence, livestock strayinto rice fields and cause damage to canal embankments where grass is usually plentiful.Conflicts continue between the influential and fairly prosperous herdsmen and the settlers. Withrespect to wildlife, irrigation development has obstructed the established migratory pattern ofelephants, whose habitat is increasingly encroached upon, leading to damage to crops and

occasional injury or death of humans or elephants; both populations are large for theirsupporting environments. On the other hand, the Project has expanded the elephant habitatthrough the creation of the Lunugamvehera Park that is north of the reservoir and connectedwith the Yala Park.

C. Impact on Institutions and Policy

44. The feasibility study did not include institutional analysis, nor did subsequent appraisals,beyond the presentation of organization charts and assessment of staffing levels. Institutional


21/53

13

development was largely treated as a by-product of project implementation rather than as aproject objective or component. Performance weaknesses, as they emerged, were addressedby the assignment of more consultants, which assisted in completing the Project. However, theconsultants had little sustained impact on institutional development, since they addressedspecific technical problems rather than institutional issues. A number of local staff did benefitprofessionally from their association with consultants, though language barriers, different skill

and experience levels, and time pressures often tended to limit benefits. All the institutionsparticipating in the Project continue to suffer in various degrees from structural, budgetary, andmotivational constraints.

45. WUGs have been established for the irrigation system. These have helped to articulatemembers concerns in the meetings preparing seasonal plans for cropping and irrigation waterrelease. These plans are often overtaken by subsequent events such as less than anticipatedrainfall and farmers cropping preferences at variance with plans. For the management of scarcewater, cultural change is needed, involving the gathering and analytical interpretation ofinformation, and publication and enforcement of firm guidelines in cooperation with WUGs.Many settlers brought in from outside the project area appear to have obtained their allotmentby political influence. They have not established residence in the project area and only visit the

area for plowing, sowing, and harvesting. This hampers the development of community spiritand the functioning of WUGs. ID staff report that gradually more nonresidents are movingpermanently to the area, mainly due to improved reservoir performance.

V. OVERALL ASSESSMENT

A. Relevance

46. At the time of its design, the Project was afforded high priority by the Sri LankanGovernment. The southern part of the dry zone was a focus of poverty and civil unrest, whileirrigation development was seen as the optimal course for providing a living for southern arearural residents and minimizing urban drift. By postevaluation, the focus of both Sri Lankas and

ADBs strategies had changed to private sector development, policy and institutional reform,human resource development, poverty reduction, and natural resource and environmentalmanagement. It is not surprising that strategic objectives have changed over the 22 years sinceproject design and it is unlikely that the Project as designed would be approved today. Theexpenditure of close to $100 million to benefit a relatively small number of settlers can beconsidered of limited relevance in the context of overall development needs and povertyalleviation in the southern dry zone. With the benefit of hindsight, a different approach toirrigation and dryland development would have been more appropriate, including rehabilitationof existing tanks, development of small-scale pump or tank-based irrigation, and assistance toslash and burn and livestock farmers. This would have benefited a larger number of families atlower cost. Despite these reservations, the Project is classified as relevant.

B. Efficacy

47. The Project overall has been reasonably effective in terms of developing irrigation withinthe command area. However, efficacy is significantly reduced by the limited area of newirrigation, which resulted in a greatly increased cost per hectare of new or upgraded irrigation.The Project has broadly achieved its objectives in terms of increasing rice production, withyields (4 t/ha) and cropping intensity (171 percent) close to design targets. However, OFCproduction has been less than expected, with areas averaging less than 20 percent ofanticipated levels. The reduced area of new irrigation limited the number of settlers that could


22/53

14

be brought into the area by around 40 percent compared with appraisal, while many settlershave yet to take up residence, leading to a less than effective settlement outcome. The efficacyof livestock, forestry, and agricultural extension was low, contributing to an overall efficacy ratingfor the Project of less satisfactory.

C. Efficiency

48. As noted in para. 8, considering alternative options at project preparation rather thanbasing the selected design on the desire to maximize the new area to be brought underirrigation could have led to a more cost-effective solution. The new irrigation area is only64 percent of target (para. 23) due to lower than expected inflows into the reservoir combinedwith uncontrolled irrigation leading to high per hectare water usage. The shortage of qualifiedconstruction staff due to overstretched national building capacity led to substantialimplementation delays (dam construction was four years behind schedule) and inflated costsdue to lack of competition. The low EIRR of 2.6 percent (paras. 32-33) compares with a rateexpected at appraisal of 17.6 percent. Overall, the Project is considered to have represented aninefficient use of resources.

D. Sustainability

49. The irrigation system is now operating reasonably well, within the confines of a restrictedwater supply. The main threat to sustainability comes from developments in the catchment areaand restricted water inflows to the reservoir. While water shortages over the past 14 years havebeen severe on many occasions, developments in the catchment suggest that increasingproblems will be experienced. This will require great efforts on the part of scheme managementand water users to develop water saving systems, such as increasing the area of upland crops,introducing rice transplanting, and improving irrigation efficiency. Substantial improvement isalso required in catchment area planning and management. The second key area affectingsustainability is the level of, and responsibility for, O&M with the need to develop WUG fundingof scheme O&M. Experience in the ADB-financed Walawe Irrigation Improvement Project13

suggests that development of fully effective user-funded O&M is a long-term undertakingrequiring substantial outside assistance. In combination, these two major factors lead to asustainability rating of less likely.

E. Institutional Development and Other Impacts

50. The Project did not have explicit institutional strengthening objectives and consequentlyhad little institutional impact. Implementation and funding structures were complex, makingmanagement difficult. Weaknesses continue to affect the project institutions in collecting,recording, and interpreting data for accounting, technical assessment, or monitoring. In relationto accounting, ID finally improved its performance and submitted audited statements that wereaccepted by ADB. All agencies suffer from severe deficiencies in monitoring their activities and

evaluating impact. The occasional surveys and studies done by the Agrarian Research andTraining Institute were no substitute for effective ongoing monitoring and evaluation by theinstitutions themselves. Information on agricultural production from the area remains sparse.WUGs have been formed and many are operating satisfactorily. However, the nonresidence ofmany farmers restricts their operation, and most have yet to take on responsibility for O&M. TheProject is assessed as having little institutional development impact.

13

Loan 695-SRI(SF): Walawe irrigation Improvement Project, for $11 million, approved on 27 September 1984.


23/53

15

F. Overall Project Rating

51. Major parts of the Project have been built as originally designed, including the dam andmost of the main canals. The new irrigation area fell one third short of target while wateravailability was reduced by an apparent decline in rainfall and significant irrigation developmentin the catchment, threatening project sustainability. While the past few years have seen

improved rainfall and irrigation scheme management, the environment remains risky and limitsinnovation and farmers ability to pay for scheme O&M. Settlement infrastructure is welldeveloped and appreciated by the settlers, in particular the piped supply of drinking water. Interms of family income, many settlers remain poor and malnutrition remains widespread in bothnew and old areas. Positive and negative environmental impacts are more or less balanced withsome reasonable expectation that the remaining problems may be corrected. Significant costoverruns, time overruns, and benefits below expectation have resulted in poor economicperformance. In spite of some positive achievements, the Project is considered less thansuccessful based on the ratings in Table 1.

Table 1: Assessment of Overall Project Performance

Criteriona

Weight (%) Assessment Rating(0-3)

WeightedRating

1. Relevance 20 Relevant 2 0.402. Efficacy 25 Less Satisfactory 1 0.253. Efficiency 20 Inefficient 0 0.004. Sustainability 20 Less Likely 1 0.205.Institutional and Other Development Impacts 15 Little 1 0.15Overall rating Less than Successful 0.6> 2.5

Successful 1.6 - 2.5Less than Successful 0.6 - 1.6Unsuccessful < 0.6

G. Assessment of ADB and Borrower Performance

52. Overall, the performance of ADB was satisfactory, though with hindsight, it is clear thaterrors were made at the time of the original design. By combining an increase in expectedirrigable area with a decrease in costs compared with feasibility study estimates, the appraisalartificially increased the level of expected performance to an unsupportable level, necessitatingtwo additional loans to complete development. At the same time, the Governments preferreddevelopment option was accepted too readily; consideration of alternatives at the feasibilitystage may have led to a more cost-effective solution. Following the poor initial design, ADB and

the Government made major efforts to salvage the Project, by ensuring that sufficient financewas available to complete development. The negative impact of the initial project design onlivestock herding was recognized in the Phase II appraisal, but the design of the livestockcomponent did little to reduce problems and conflicts.

53. The performance of the Implementing Agencies was satisfactory. However, majordeficiencies were evident in the accounting function in the early years with late submission ofaccounts and qualification of accounts by the Auditor General. Lack of familiarity with ADBprocedures and a complex funding program caused implementation problems for ID. While in


24/53

16

most cases, ADB procedures were followed, the award of the major contract for damconstruction was made without approval (para. 16). Project monitoring was unsatisfactory, andfew useful data are available on irrigation system performance and agricultural production. Inaddition to making scheme management more difficult, the lack of data preclude objectiveevaluation. Local ID management has improved its communication with farmers and WUGsover the past few seasons. This factor, combined with a moderately successful rice crop grown

with limited water in the 1999 dry season, suggests that some of the problems of the early yearsmay have been overcome.

VI. ISSUES, LESSONS, AND FOLLOW-UP ACTIONS

A. Key Issues for the Future

1. Reliability of Irrigation Water Supply and Equitable Distribution

54. Despite improvements in water management over the past few years, irrigationefficiency remains low. More effort is required to distribute water efficiently, to reduce deeppercolation losses, and to recapture runoff and drainage water for irrigation use. This will require

a major campaign to change attitudes among irrigation staff and farmers. Recapture of drainagelosses is already being undertaken and may be expanded.

55. Much water could be saved on the more permeable soils by growing crops other thanrice. However, farmers are unlikely to follow the recommendations of the extension service foryield-increasing production technology as long as the risk of drought is not brought undercontrol and equitable water supplies are not assured. Neither ID and agricultural extension staffnor farmers have adequate experience to support the changes in irrigation managementrequired by non-rice crops. Such a development will need a major change in attitudes, based onthe necessary political will, incentives, and knowledge. A good start has been made with theestablishment of bananas on about 400 ha, though the crop is presently experiencing marketingproblems. Further diversification into OFCs is suggested for the porous soils in much of the new

irrigation area. With the more efficient use of water on porous soil areas, more water could bemade available to the lowland soils best suited to growing rice (including the old irrigationsystem and parts of the new system) allowing cultivation of two crops of rice per year asoriginally envisaged. Efficient management, based on regular data collection, monitoring, andtransparent documentation is a key step in the process. This includes institutional support toenable local staff to resist pressures that currently restrict the efficient operation of the system.

56. If no water flows into the reservoir, the question of water management will becomeirrelevant, since there will be no water to manage. Irrigation development and degradation in thecatchment area are thus major concerns to Kirindi Oya managers and farmers, requiring urgentaction if the existence of the scheme is not to be placed in severe jeopardy. A number ofsuggestions are made for follow-up action in Section C.

2. Nutrition

57. Early childhood malnutrition continues to be a problem in Sri Lanka in general. It is ofconcern that this should also be a problem among all groups in the project area, according tothe Governments PCR. This was confirmed by the OEM through discussions at the project-supported clinic. Stunted child growth is the most obvious symptom of such malnutrition, oftencombined with intestinal infections or parasites. Less obvious is the negative impact onintellectual development. The observed lack of entrepreneurial spirit among farmers is a likely


25/53

17

cause for concern in this regard. Improved water management and higher incomes could go along way to reducing the problem, at least for the next generation. The project-supported clinicshould be placed in a position to support a child nutrition program, in particular in the criticalphase of weaning and to expand preventive maternal and child health care.

3. Environment

58. Improved water management and reduced drainage water flows can reverse much ofthe environmental damage to coastal lagoons. Diversion of fresh water around the lagoons isalso possible but would be expensive. A study of options for water reuse within the irrigationscheme and/or diversion of drainage water away from the lagoons is merited. With regard to thedamage done by free-grazing herds of cattle, enforcement of existing laws, an incentive toreduce herd size (in favor of better yielding cattle), and a program of reintegration of cattle intothe farming system may assist in resolving the conflict between settlers and herd owners.However, it is not easy to make changes to farming systems. Improvement in the medium termis likely be found through the formation of sufficiently strong WUGs that can negotiate andenforce agreements with the powerful herd owners in relation to seasonal grazing restrictions.The size of the elephant population needs to be controlled in line with the carrying capacity of

their remaining habitat, either by birth control or by culling.

4. Irrigation Service Fees

59. Farmers cannot be expected to pay irrigation service fees as long as the service doesnot provide for a comparatively risk-free production environment. Under the circumstances inthe Project (and in many other locations in Sri Lanka), the covenanted irrigation service feeshad no chance of implementation. Improved irrigation management may change that prospectand ensure equitable treatment of resident and nonresident settlers, with nonresidents requiredto contribute in cash to compensate for the maintenance work that they do not undertake. Feecollection by the WUGs may in future allow O&M costs to be fully met, at least to the branchcanals and eventually the main canals, if not the dam.

5. Land Ownership and Residence

60. At present, a majority of settlers have been provided with land grants that entitle them toreside on the land but not to subdivide it or sell it without permission. Around 1,500 out of 9,659grants remain to be issued, mainly because of the nonresidence of the farmers. The question ofnonresident settlers needs to be resolved as soon as possible and the remaining grants issued,or where necessary land reallocated. The granting of freehold title and development of a marketfor land is desirable, though this may need to await wider developments in Sri Lankan law. Theissue of nonresident farmers is also important in the effective functioning of WUGs for themaintenance of irrigation systems and for developing a community spirit. It is equally relevant tothe efficient operation of agricultural extension.

B. Lessons Identified

61. Many of the developments observed during the project period would not happen thesame way again. With the experience of much reduced rates of return for new irrigationprojects, ADB is now focusing on the rehabilitation of existing schemes, adoption of participatoryapproaches from the outset, and early transfer of operation to water users. For new projects, anengineering loan would precede investment to ensure cost-effective design and reduceimplementation delays due to design issues. Nonetheless, a number of lessons can be drawn:


26/53

18

(i) ADB has a special responsibility in approving projects. It should ensure that theoptimal investment option has been selected. It must be transparent in explainingwhy major changes are made at appraisal relative to feasibility study analysisand when major assumptions are revised in subsequent economic evaluations.Substantial changes should not be made to costs or benefits unless supported byrigorous analysis.

(ii) ADB must be able to challenge underlying assumptions and require that theproposed design reflects circumstances found in the target area. It should not behurried into making politically driven decisions.

(iii) Getting the design correct from the beginning, including institutional aspects, is anecessary condition for project success. This is particularly true because neitherADB nor a Government has a specific capacity to withdraw from a project if itsdesign proves to be flawed.

(iv) Where irrigation schemes include areas of existing irrigation, it is important toclarify existing formal and informal water rights, and to ensure that water

allocation between new and existing users is equitable. The needs of existinglivestock owners must also be taken into account.

(v) Particularly for schemes that are planned to utilize a significant proportion of totalcatchment water discharge, detailed analysis of catchment hydrology isessential, both at the time of design and thereafter at regular intervals. Improvedstream gauging and rainfall recording may be required to support such analysis.Irrigation development must take a conservative view of future river flows, andrecognize that water demand for all purposes is likely to increase over time.Schemes that use a high proportion of predicted seasonal or annual flow areinherently risky and should usually be avoided.

(vi) Settlement schemes can be highly charged politically. Major efforts arewarranted to ensure that settlers move to their allotments to promote both socialdevelopment and user management of irrigation and agricultural development.

(vii) Good record keeping is essential for management as well as facilitating theunderstanding of past developments. In particular, all details of economicevaluation should be electronically stored for greater accountability and futurereference.

C. Follow-Up Actions

62. A number of the issues raised invite follow-up action by the Government. Many requirechanges in institutional culture and incentive structures. The draft proposal for improvedirrigation management, prepared by IWMI in late 1997 (ibid), is considered a good starting point,but should be expanded to capture sociocultural and institutional dimensions. The followingactions in Table 2 should be considered to promote the further achievement of benefits and theirsustainability.


27/53

19

Table 2: Follow-Up Actions

Measure/Action Responsible Unit Period1. Watershed Management

i. Prepare improved water balances for the catchment, irrigation system, and lowervalley to define systems for efficient water allocation, minimization of drainagewater wastage, and reduction of damage to the coastal lagoon environment. This

will require improved measurement, recording, and analysis of climatic andhydrological data.

ID 2001

ii. Control watershed and irrigation development upstream of the reservoir to securewater availability and project sustainability. A coordinated planning approach toriver basin development is necessary given the range of agencies responsible fordevelopment in the area.

ID, AgrarianService

Department,Agricultural

DevelopmentAuthority

Starting 2001

2. Irrigation Scheme Managementi. Review options for perennial crops (in addition to bananas) that may be

established by farmers on porous soils with a view to reducing irrigation waterdemand. Pursue this strategy and apply water savings to increase croppingintensity on both new and old irrigation areas.

DOA 2001-2005

ii. Strengthen the authority of WUGs, empowering them to fine or withhold waterreleases from farmers who neither contribute their labor nor pay the equivalent incash to undertake maintenance activities.

ID 2001-2005

iii. Promote more efficient irrigation scheme O&M. Continue to review operationalprocedures for water release in discussion with WUGs.

ID 2001-2005

iv. Control excess drainage flows by a more judicious release of water and improvedon-farm management, and capture and reuse irrigation drainage water. Accountwill be taken of current water reuse by squatters and landholders outside theProject.

ID 2001-2002

v. Review, revise, and enforce rules, regulations, and procedures with regard toirrigation management, maintenance, role of WUGs, and dam disasterprocedures.

ID 2001

vi. Establish a monitoring system providing information on the different productionenvironments of the Project (porous soils, lowland soils, old irrigation system, newirrigation area, areas reusing drainage water). Information should includecropping patterns and yields, and be supported by surveys assessing production

technology, cost of production, use of labor, yields, and income derived.Information should allow irrigation management and extension service to makedecisions on a sound basis rather than on assumptions or anecdotal evidence.

ID 2001-2002

vii. Reduce crop and canal damage by free-ranging cattle by preparing andsupporting a program to improve livestock management, reduce herd sizes, andpromote pasture and fodder development. Authorize and encourage farmers toharvest grass along irrigation canals for cut-and-carry livestock feeding.

ID 2001

viii. Reorient adaptive agricultural research to aspects of direct relevance to intensiveand diversified farming in the project area.

DOA 2001-2005

3. Settlementi. Issue land titles to resident settlers and withhold titles from nonresident farmers,

and provide for freeholding of farms when permitted by Sri Lankan land law.Land

CommissionersDepartment

Starting 2001

ii. Review the status of nonresident farmers and expropriate their land if they do not

relocate to the settlement within a period of about two years. Reallocate land topoor people, from surrounding areas, with qualifications and credentials to bereviewed and approved by the relevant WUGs.

Land

CommissionersDepartment

2001-2002

iii. Review nutrition data and medical records and develop strategy to combatmalnutrition in the project area.

Department ofHealth

2001


28/53

20

APPENDIXES

Number Title Page Cited on

(page, para.)

1 Project Scope and Accomplishments 21 1, 3.

2 Project Financing Arrangements 23 2, 4.

3 Comparison of Project Costs 24 2, 4.

4 Economic Reevaluation 27 2, 6.


29/53

Appendix 1, page 1

PROJECT SCOPE AND ACCOMPLISHMENTS

Original Up to End of Up to End of

Scope Phase I as Phase II as Actual

Component 1977 Approved 1982 Approved 1986Accomplishme

nt

A. Irrigation

1. Earthfill dam with Gated 5 km 5 km 5 km 5 km

Spillway and Main Canal Gates 25 m high 25 m high 25 m high 25 m high

2. Reservoira(ha) 3,577 3,577 3,577 3,577

3. Main Canal and Distribution System(km)

61.6 47 69.3 51.4

4. Rehabilitation of Existing Irrigation (ha) 3,675b

4,525c

4,525c

4,525c

5. Land Clearing, Leveling, and Field 8,409 4,346d

8,454 5,416Distribution System for New Irrigation(ha)

B. Agricultural Development

1. Rehabilitation of Existing 4,525 4,525 4,525 4,525

Irrigation (ha)

2. New Irrigation Area (ha) 8,409 4,346 8,454 5,416

3. Cropping Intensity 189 200 170 146d

4. Yields, Paddy (t/ha) 4.00 3.9

Porous Soils (t/ha) 3.50 3.50 3.8

Lowland Soils, Maha (t/ha) 4.50 4.50 3.8

Lowland Soils, Yala (t/ha) 4.80 4.80 4.0

5. Agricultural Research Station (no.) 1 1

6. Farmer Training Center (no.) 1 1

= not applicable, ha = hectare, km = kilometer, m = meter, t = ton.a

Other sources state 3,402 ha as the reservoir area.b

According to appraisal report, page 15, para. 5, the Badagariya irrigation system (850 ha) fed from anothercatchment, did not require rehabilitation, and was excluded originally.

cBadagariya system included.

dIncluding about 9 percent of area grown with other field crops as a last resort in substitution of failedpaddy or in view of known water shortage.

21


30/53

Appendix 1, page 2

Project Scope and Accomplishments (continued)

Original Up to End of Up to End of

Scope Phase I as Phase II as Actual

Component 1977 Approved 1982 Approved 1986Accomplishmen

t

C. Land Settlement

- No. of Settlers 8,320 4,200 8,320 5,100

- No. of Hamlets 28 18 35 23

- No. of Village Centers 4 2 4 4

- Township 1 1 1 1

- Roads (km) 200 235 609 453e

- Dug Wells 420 210 0 37f

- Piped Water Supply System

for about 5,000 Households 1 1

- Primary Schools 28 18 35 20

- Junior and Senior

Secondary Schools 5 3 5 4- Cooperative Stores 28 18 35 25

- Produce Stores/WFP Stores 23 23

- Community Centers 28 18 35 22g

- Hospitals/Health Centers 33 20 33 4

- Post Offices 5 3 5 4

- Sales Centers 5 3 5 4

- Markets 1 1

- Agrarian Service Centers 1 3

- Police Station 1 1

= not

Date post:	05-Apr-2018
Category:	Documents
Upload:	independent-evaluation-at-asian-development-bank
View:	225 times
Download:	0 times