Research Report

Results of Management Turnoverin Two Irrigation Districtsin Colombia

Douglas L. VermillionandCarlos Garcés-Restrepo

INTERNATIONAL IRRIGATION MANAGEMENT INSTITUTEP O Box 2075 Colombo, Sri Lanka

Tel (94-1)867404 • Fax (94-1) 866854 • E-mail IIMI@cgnet.comInternet Home Page http:/ /www.cgiar.org

ISSN: 1026-0862ISBN: 92-9090-329-5

4

2

Research Reports

IIMI’s mission is to create sustainable increases in the productivity of irrigated agriculturewithin the overall context of water basins and the analysis of water resource systems as awhole. In serving this mission, IIMI concentrates on the integration of policies, technologiesand management systems to achieve workable solutions to real problems—practical, rel-evant results in the field of irrigation and water resources.

The publications in this series cover a wide range of subjects—from computer model-ing to experience with water users associations—and vary in content from directly appli-cable research to more basic studies, on which applied work ultimately depends. Some re-search papers are narrowly focused, analytical, and detailed empirical studies; others arewide-ranging and synthetic overviews of generic problems.

Although most of the papers are published by IIMI staff and their collaborators, wewelcome contributions from others. Each paper is reviewed internally, by IIMI’s own staff,by IIMI’s Senior Research Associates and by other external reviewers. The papers are pub-lished and distributed both in hard copy and electronically. They may be copied freely andcited with due acknowledgment.

3

Research Report 4

Results of Management Turnoverin Two Irrigation Districts in Colombia

Douglas L. Vermillion and Carlos Garcés-Restrepo

International Irrigation Management InstituteP.O. Box 2075, Colombo, Sri Lanka

4

Vermillion, D. L., and C. Garcés-Restrepo. 1996. Results of management turnover in two irrigationdistricts in Colombia. Research Report 4. Colombo, Sri Lanka: International Irrigation Manage-ment Institute (IIMI).

/ irrigation management / irrigated farming / privatization / farmer participation / farmers’ associa-tions / farmer-agency interactions / farmer-managed irrigation systems / social aspects / economicaspects / operation / maintenance / agricultural production / sustainability / water distribution / Co-lombia / Coello / Saldaña /

ISBN: 92-9090-329-5ISSN: 1026-0862

© IIMI, 1996. All rights reserved.

Responsibility for the contents of this paper rests with the authors.

Editor: Kingsley Kurukulasuriya; Consultant Editor: Steven Breth; Artist: D.C. Karunaratne;Typesetter: Kithsiri Jayakody; Editorial/Production Manager: Nimal A. Fernando.

The authors: Douglas L. Vermillion is a Social Scientist and Management Specialist at theInternational Irrigation Management Institute (IIMI) and Carlos Garcés-Restrepo is anIrrigation Specialist at IIMI.

The authors wish to acknowledge, in memorium, the excellent work of Juan G. Fernandez, theprincipal field investigator, who died of cancer in 1995. They also thank David Seckler, ChrisPerry, R. Sakthivadivel, Harald Fredericksen, and Gil Levine for their valuable comments onearlier drafts of this paper.

IIMI is grateful to the Colombian Institute for Land Development (INAT) for its supportfor this study. IIMI gratefully acknowledges the financial and technical support for this re-search from the Bundesministerium für Wirtschaftliche Zusammenarbeit (BMZ) and theDeutsche Gesellschaft für Technische Zusammenarbeit (GTZ) of Germany and the Ford Foun-dation.

5iii

Contents

Summary 1

Introduction 3

Irrigated Agriculture in the Tolima Valley 4

Development of the Coello and Saldaña Districts 5

Turnover Process 7

Changes in Management 8

Financial management 8

Personnel 9

Operation and maintenance 9

Organization 11

Management Performance after Turnover 12

Impacts on government 12

Financial viability 12

Physical sustainability 17

Irrigation operations 20

Agricultural productivity 25

Perspectives of Stakeholders 28

Farmers 28

District staff 30

Agency staff 30

Conclusion 30

Perception 30

Main results 31

Literature Cited 32

1

Summary

In 1975, farmers in the Coello and Saldaña irrigationdistricts in the Tolima Valley, Colombia, petitionedthe government for the right to take over manage-ment of the districts. They based their argument onthe fact that, over the previous 20 years, they hadalready repaid their agreed 90 percent share of thecost of construction. They were also paying water feesto the government and were dissatisfied with the costand quality of management that the government wasproviding. They argued that they could manage thesystems more cost-effectively than the government.In 1976, the government agreed to the farmers’ de-mands, expecting that turnover would save moneyfor the government.

This paper assesses the extent to which turnoverof irrigation management to farmers in the Coello andSaldaña irrigation districts in Colombia has had animpact on the cost of irrigation to farmers and thegovernment, the sustainability of irrigation, and thequality of water distribution.

The sustainability of irrigation is assessed rela-tive to both the financial viability of the districts andthe physical condition of irrigation infrastructure 19years after turnover. The quality of water distribu-

tion is assessed relative to efficiency and equity ofdistribution and to productivity of water.

The study found that turnover did achieve thegovernment’s main objective of discontinuing gov-ernment financing for operations and maintenance.However, since the government retained partial con-trol over the irrigation districts after turnover, stafflevels declined slowly and the cost of irrigation tofarmers changed little. A detailed inventory of irri-gation infrastructure found that the vast majority ofstructures and canal lengths were in good functionalcondition.

The districts were able to continue to expand theirrigated area modestly and sustain high levels ofproduction after turnover, partly due to a policy tolimit rice production and deliver lower volumes ofwater per hectare. Perhaps the finding that would beof most concern to farmers was that while the cost ofirrigation did not increase after turnover, the grosseconomic value of production per hectare and perunit of water grew dramatically. After turnover, irri-gation constituted a relatively small and decliningproportion of the total cost and value of agriculturalproduction.

3

In 1975, farmers in two irrigation districtsin the Tolima Valley, Colombia, petitionedthe government for the right to take overmanagement of the districts. They basedtheir argument on the fact that, over theprevious 20 years, they had already repaidtheir agreed 90 percent share of the cost ofconstruction. In addition, they were dissat-isfied with the cost and quality of irriga-tion management for which they paid wa-ter fees to the government. They claimedthat they could manage the systems morecost-effectively. In 1976, the governmentagreed to the farmers’ demands, expectingthat turning over management of the irri-gation districts would save it money.

This paper assesses the extent to whichturnover of irrigation management to farm-ers in the Coello and Saldaña irrigation dis-tricts in Colombia has had an impact on:

• the cost of irrigation to farmers and thegovernment

• the sustainability of irrigation

• the quality of water distribution

The sustainability of irrigation is as-sessed relative to both the financial viabil-ity of the districts and the physical condi-tion of irrigation infrastructure 19 years af-

ter the turnover. The quality of water dis-tribution is assessed relative to the effi-ciency and equity of distribution and to theproductivity of water.

This study was conducted from 1993to 1995. It involved analysis of data fromthe records of the two irrigation districtsand from group and individual interviewswith farmers, district management staff,board members, and agency staff. Thestudy sampled 93 randomly selected farm-ers, 44 in Coello and 49 in Saldaña (Garces-Restrepo and Vermillion 1994).

The central finding was that after turn-over the gross value of output per hectareand per unit of water increased dramati-cally while the cost of irrigation to farmersheld steady. New policies imposed by thewater users’ associations after turnoverplayed a leading role in the improvementsin economic performance. In particular,policies that restricted rice production insandy areas and that reduced the averagevolume of water delivered per hectare sup-ported crop diversification and raised thevalue of irrigated output. This study lendssupport to the hypothesis that managementturnover programs worldwide will intro-duce incentives for improvement of man-agement efficiency and productivity of ir-rigated agriculture (Vermillion 1992).

Results of Management Turnover in Two IrrigationDistricts in Colombia

Douglas Vermillion and Carlos Garcés-Restrepo

Introduction

4

Colombia is a mountainous country (figure1) with relatively abundant water resources,including more than 1,000 perennial rivers.It has both tropical and temperate climatesand an average rainfall of 1,500 mm/year.

Coello and Saldaña irrigation districtsare located in the Tolima Valley in centralColombia at an elevation of 380 meters. Thevalley sits between the central and westmountain ranges of the country, and thelarge Magdalena River runs through it. Thevalley has mostly a flat topography with un-dulating terrain toward both mountainranges and has primarily alluvial soils, fans,terraces, and narrow valleys with minor riv-

ers. The main soil characteristics are sandyand loam in Coello and clay and loam inSaldaña. Soil erosion is evident on the hill-sides but is not yet a problem in the valleyfloor, except that it creates a high silt loadin rivers and irrigation canals. Yearly pre-cipitation in the valley is 1,000 to 1,500 mil-limeters. A marked bimodal distribution inApril/May and October/November makesthe need for irrigation primarily supple-mental. The median temperature is 27.9°C.The average relative humidity is 74 percent,and the yearly average tank evaporation is1,800 millimeters (figure 2).

In the 1930s, land re-form in the Tolima Valley re-placed the old hacienda sys-tem of peasant cultivationwith landownership forfarmers. Irrigation intro-duced in the 1950s trans-formed the valley’s agricul-ture. Cotton became an im-portant crop during the1950s and 1960s. It waseventually replaced by rice,which has been the main ir-rigated crop since the 1970s.Maize, sorghum, fruit, andvegetables are also irrigatedin the valley now.

Today, the Tolima Val-ley is a relatively prosper-ous farming area, located ata major transportationcrossroads. It has numeroustowns where agricultureand agro-business are themainstream of the localeconomy. A large number ofboth public and private or-ganizations that providetechnical assistance and ag-ricultural support servicesto farmer-managed irriga-tion systems are present inthe area.

Irrigated Agriculture in the Tolima Valley

FIGURE 1.Map of Colombia, with Coello and Saldaña irrigation districts.

5

The Coello-Saldaña irrigation district be-came operational in 1953 when 8 years ofconstruction were completed. The totalcapital cost for the district was US$5,500 perhectare (in 1993 dollars). Initially Coello andSaldaña were constructed and managed asa single district. They were separated in1976 after management was turned over tothe water users’ associations.

Coello irrigation district is a river di-version system. It has a lateral intake with

a design capacity of 28 m3/s. The intakeconsists of an approach channel formed byan earthen levee, which facilitates flow in-take during low river levels in the dry sea-son. The intake has two radial gates withprovision for controlling both sediment in-take and water depth (HIMAT 1991a). Theintake channel leads to the main convey-ance canal (Gualanday), which has a capac-ity of 25 m3/s and extends for 5.7 kilome-ters before reaching the command area. The

FIGURE 2.Climatic data for Tolima Valley, Colombia.

Development of the Coello and Saldaña Districts

6

main canal divides into four branch canals,each of which leads to unlined secondaryand tertiary canals. Field turnouts are slid-ing gates.

Coello district served an irrigated areaof 25,600 hectares in 1993 (table 1), makingit one of the largest schemes in the country.Because the district was not designed witha parallel drainage system, waterloggingand salinity are problems on as many as7,000 hectares. In 1993, about 38 percent ofthe farms were 5 hectares or less in size andonly 6.4 percent exceeded 50 hectares. Com-pared to 1968, we see a trend toward frag-mentation of holdings and less inequity infarm size among farmers (table 2).

Saldaña district, located south of theCoello district, is also a river diversionscheme. It diverts water from the SaldañaRiver through a direct intake without anapproach levee. The intake also has radialgates and water head controls. It has a de-sign capacity of 30 m3/s into the main con-veyance canal. This canal conveys water to

three partially lined branch canals. As inCoello, each branch canal leads to unlinedsecondary and tertiary canals. Field turn-outs are sliding gates (HIMAT 1991b).

In 1993, Saldaña district irrigated13,975 hectares (table 1). Because it lacks acomplementary drainage system, waterlog-ging affects up to 1,600 hectares. Its 1,500water users have 1,850 farmholdings. In1993, 63.9 percent of the farms were 5 hect-ares or less in size and only 1.0 percent ex-ceeded 50 hectares. As with Coello, thetrend between 1968 and 1993 was towardfragmentation of holdings and less inequityin farm size among farmers (table 2).

Along the main canals, both schemeshave composite underflow and overflowcross regulators, which consist of gates andside weirs. This design protects againstovercropping and facilitates desilting. Italso allows gate adjustments to be made atlonger intervals than conventional designsdo. Hence, the design facilitated turnoverby simplifying O&M requirements.

TABLE 1.Basic information, Coello and Saldaña districts.

Item Coello Saldaña

State Tolima TolimaPeriod built 1949�53 1949�53Transferred Sept. 1976 Sept. 1976Design area (ha) 44,100 16,428Gross irrigated area, 1993 (ha)a 25,628 13,975Net irrigated areab (000 ha) 15.3 12.5Water users� association Usocoello UsosaldanaMain soil type Sandy, loam Clay, loamMain crops Rice, soybean, cotton RiceWater source River Coello River SaldañaSystem type Run-of-the river Run-of-the riverIntake structure Radial gates Radial gatesIrrigation structures 1,666 756Lowest water measurement point Secondary canal Secondary canalWater delivery efficiency (%) 69 69Length of main canal (km) 69 61Length of canal network (km) 250 162Area served per km canal (ha) 102 86Turnout type Sliding gates Sliding gates

aTotal of the areas irrigated for all crops grown in a year.bArea irrigated for at least one crop a year.

7

The rehabilitation of both irrigation ca-nals and natural drains was done in Coelloand Saldaña well before turnover. In Coello,US$8.69 million (in 1988 dollars) was spenton rehabilitation between 1968 and 1973.In Saldaña, US$2.28 million (in 1988 dol-lars) was spent on irrigation and drainageworks between 1969 and 1972.1 At the timeof management turnover in 1976, the sys-tems were in good physical condition, andrehabilitation was not an issue in negotia-tions between the government and thefarmers. Rehabilitation was not done in con-nection with turnover.

However, the issue of who should beresponsible for financing rehabilitation wasalways a matter of dispute. The users ar-gued that because the government had notturned over ownership of the infrastruc-ture, it belonged to the nation and was thegovernment’s responsibility to maintain.Despite pressure from the government,farmers have refused to repay the cost ofrehabilitation in either system, except foran agreement with farmers in Coello to re-pay the cost of building a feeder canal tosupplement their water supply (which isstill under construction, today).

1Drainage systems inboth schemes are natu-ral drains. No drainagesystem was ever con-structed. The rehabilita-tion and maintenance ofdrains refer to thedesiltation of smallstreams, the redirectionof natural outlets, etc.

TABLE 2Number of farms by size, Coello and Saldaña districts, 1968 and 1993.

1968 1993Farm size Coello Saldaña Coello Saldaña (ha) no. % no. % no. % no. %

0-5 264 26.6 589 56.4 703 38.5 1,255 63.9

5.1�10 200 20.1 146 14.0 386 21.1 279 14.2

10.1�20 207 20.8 141 13.5 300 16.4 231 11.7

20.1�50 180 18.1 115 11.0 322 17.6 181 9.2

> 50 143 14.4 54 5.1 115 6.4 19 1.0

Total 994 100 1,045 100 1,826 100 1,965 100

Turnover ProcessIn the early 1960s, the Government of Co-lombia entrusted the O&M of its irrigationdistricts to INCORA, the government landreform agency. The performance of theagency in irrigation management was notsatisfactory. Water users of the Coello-Saldaña district were not only unhappywith the poor O&M service provided butwere also concerned about the high man-agement costs. In the early stages of devel-opment in the 1950s, more than 90 percentof the farmers paid the water fee, but thepercentage of payers declined over timedue to farmer dissatisfaction with the qual-ity of system management. Declining fee

collections further limited the ability of thegovernment agency to provide an effectiveirrigation service. Inefficient system O&Mfurther motivated farmers to seek to takeover management of the district.

In 1975, the farmers, who had alreadyformed associations, decided to formallyrequest that the government transfer man-agement responsibility for the system to theassociations of water users. The associationsargued that the scheme should legally be-come their property because farmers hadalready repaid the government their shareof the costs of construction (Vermillion andGarces-Restrepo 1994).

8

Negotiations for management transferwere completed within a year, between1975 and 1976. The associations hired a law-yer to represent them in negotiating theterms of the transfer. Issues to be resolvedincluded the disposition of district staff,ownership status of scheme assets, and thefuture degree of involvement of HIMAT2

in the districts. It was finally agreed thatmost of the existing staff would be retainedby the districts and others would be trans-ferred out. The ownership of irrigationstructures would remain with the govern-ment, although some equipment and facili-ties were transferred to the farmers’ asso-ciations. The government concluded thatunder existing laws it could not relinquishthe ownership of scheme assets. HIMATwould retain a role of oversight for districtmanagement to ensure that the systemswere properly maintained. In practice, thismeant that HIMAT continued to give its ad-vice and consent for annual budgets, O&Mwork plans, setting water fee levels, andpersonnel changes. The farmers’ associa-tions obtained direct control over the O&Mof the entire system, including the intakes.

As part of a policy to improve the per-formance of irrigation districts, the govern-ment created HIMAT in 1976. Its initial taskwas to turn over management of the Coello-

Saldaña district to the two farmers’ asso-ciations. The district was divided into twoseparate districts, Coello and Saldaña. Thiswas the first irrigation management turn-over in Colombia (Plusquellec 1989). It seta precedent for later transfers.

The transfer process employed a legalrule in the country’s constitution referredto as the “delegation of administration”(Delegación de Adminis-tración), by whicha public good (in this case, an irrigation dis-trict) could be turned over to a private cor-porate entity (a water users’ association) foradministration on behalf of the state. Theusers were then empowered to recruit staffand organize and take charge of O&M forthe two systems with the proviso that itwould be financially self-reliant and gov-ernment subsidies for O&M would be dis-continued. The delegation of administra-tion created a continuing labor relationsconflict between the districts and the gov-ernment, which resulted in numerous legaldebates and proceedings until the 1990s.Labor laws prohibited the firing of existingstaff hired previously by the government.In 1993, Colombia enacted a new land de-velopment law, which was intended togrant full control over irrigation districtmanagement to farmers’ associations(Ministerio de Agricultura 1993).3

2HIMAT, or the Institutefor Hydrology, Meteo-rology and Land Devel-opment, was created inorder to implement theturnover program andprovide governmentoversight and supportto irrigation districts af-terwards. Negotiationsabout turnover in-cluded the issue ofHIMAT’s mandate afterturnover. In 1994, thisacronym was changedto INAT, when meteo-rology was removedfrom the institute’smandate.3However, this law iscurrently being chal-lenged in the courts re-garding the issue of re-leasing staff who wereoriginally hired by thegovernment.

Changes in Management

Financial management

Farmers expected that through turnoverthey would not only improve managementbut would also contain or reduce the costof irrigation. However, it soon became ap-parent that the delegation of administrationwould not give farmers’ associations com-plete control over their budgets and O&Mplans. Although the farmers had wantedHIMAT to play an advisory role, the gov-ernment continued to influence budgeting

and staff decisions. After turnover, the dis-tricts were unable to reduce staff and costsas much as they wanted, due to resistancefrom HIMAT.

Two kinds of water charges are assessed:a flat area charge (based on farm area irri-gated) and a volumetric charge (based onbasic water requirements by crop type). Therevenues from the area charge are used toguarantee the coverage of fixed costs suchas personnel. The volumetric fees are usedmore for variable costs such as operations.

9

Before and after turnover, farmers paidthe area-based water fee prior to the sea-son for which water was ordered. The volu-metric fee was paid after the season and hadto be paid entirely before any irrigation or-ders could be approved for the next crop-ping period. Farmers are charged volumet-ric fees according to the type of crop plantedand its respective “base” or target alloca-tion. Because water is only measured rou-tinely down to the heads of secondary ca-nals, volumetric charging is based on theo-retical, as opposed to measured, water de-liveries. Farmers may complain if they be-lieve that their actual deliveries are less thanadequate or less than the assessed amount,in which case district staff may make spe-cial measurements at tertiary offtakes withsmall flumes. This can result in either anadjustment in the volume delivered or inthe fee assessment. This system did notchange with turnover.

Because the associations did not receiveownership of system assets and becausethey had not paid for previous rehabilita-tion costs, the farmers expected the govern-ment to pay for any future rehabilitation orstructural replacement. Hence, after turn-over, farmers did not raise a capital-replace-ment fund (although they did raise anequipment-replacement fund).

PersonnelOne of the more noticeable outcomes ofturnover was the significant reduction ofpersonnel. Before transfer, in 1975, the twodistricts combined had 300 employees. By1993, the total staff for both districts was189, a 37 percent reduction since transfer(table 3). Accounting for changes in area ir-rigated, in 1975 there were 62 hectares ofservice area per district staff member. By1993 the ratio had risen to 147 hectares perstaff member (table 3). Most reductionswere made in maintenance and technicalsupport staff and occurred gradually,mostly through attrition and nonreplace-ment. Labor laws made it difficult for dis-trict managers to release dispensable gov-ernment employees. Nevertheless, districtboard members and agency officials re-ported that paperwork was diminished andadministration became more efficient aftertransfer, especially in irrigation scheduling,fee processing, and communications be-tween users and district management.

Operation and maintenanceIn the Coello and Saldaña districts, water isallocated to farmers on the basis of area andcrop type. In theory, all users who plant the

TABLE 3.Staff levels before transfer (1975) and after transfer (1993), Coello and Saldaña districts.

1975 1993Program Both districts Coello Saldaña Both

Administration 36 18 18 36Maintenance 161 60 50 110Operation 51 19 24 43Technicala 52 0 0 0Total 300 97 92 189Net irrigated areab (000 ha) 18.7 15.3 12.5 27.8Area/staff member (ha) 62 158 136 147

aSeveral �technical staff� members were retained after turnover but were shifted to other programs, including staff forhydrologic measurement, design, and accounting.bArea irrigated for at least one crop a year.

10

same crop type receive a basic allotment,and they pay area and volumetric chargesbased on assumed deliveries relative to per-hectare targets by crop type. Before turn-over, irrigation was scheduled on the basisof preseason crop plans, modified duringthe season through water requests submit-ted by registered farmers. The district man-agement prepared irrigation schedulesbased on orders received from farmers. Ir-rigation requests were approved to the ex-tent that predicted water availability metthe aggregate demand. The user was re-sponsible for going to the district office be-fore the season to sign an agreement withthe seasonal irrigation plan. The user wasinformed of the day and time of his or herirrigations, and the ditch tender made aninspection of the farm to be sure that thecanals, turnout, and measuring devices (ifany) were in working condition. The ditchtender was responsible for delivering wa-ter to farm turnouts according to the agreedschedule and recording the total water de-livered for the season. This system contin-ued after turnover and was implementedmainly by the same staff as before, but theywere employees of the farmers’ associations.

The districts estimate water require-ments for each crop type, and these esti-mates become the base allocation. The dis-tricts measure discharge at the intake andalong the main canal at offtakes into sec-ondary canals. Water is distributed accord-ing to the base allocation, but sometimes isreduced when water is scarce.

Prior to turnover, the governmentagency compiled annual plans for mainte-nance and repairs. Such plans were pre-pared by the head of the maintenance unit,based on field inspections and sometimeson complaints from farmers. The most com-mon maintenance works were desilting andcleaning of canals, road maintenance, andstructural repairs. Targets were establishedin advance, but deviations were commondue to funding constraints for repair or op-eration of heavy equipment.

The district management has adminis-trative and operational manuals detailingroles and responsibilities of staff and users.The districts have maintained data on dailyrainfall, temperature, and relative humid-ity since the inception of the project. Dataon river flows and main and branch canaldischarges are also recorded daily. Recordsof seasonal crop and irrigation plans, feecollection levels, registration of farmers,inventory of equipment and supplies, ac-counting, and yearly budgets have beenkept regularly, before and after turnover.

After turnover, the new district admin-istrations introduced practices to improveirrigation efficiency and allow continuedexpansion of irrigated area. Attention waspaid to reducing staff where possible andrevising cropping patterns to be consistentwith the relative scarcity of water in the twosystems. Water is more scarce in Coello. In1993, the average target discharge or dutyin Coello was 8.6 mm per day versus 15.5mm per day in Saldaña. In Coello, the an-nual water demand was 1,097 millimeters,948 millimeters of which was supplied byirrigation (table 4). In Saldaña, the annualwater demand was 1,318 millimeters,which was exceeded by an annual irriga-tion water supply of 1,517 millimeters. Dur-ing the same year, the relative water sup-ply (i.e., ratio of total supply, including ef-fective rainfall, to demand, calculated at thesecondary-canal level) was 1.4 in Coelloand 1.8 in Saldaña (table 5). Water is rela-tively more scarce in Coello during otheryears as well.

In Coello, where water supplies wereinsufficient for planting rice over the entiresystem, the association introduced a ricerotation and zoning plan to enable all farm-ers to plant rice at least once per year. InSaldaña, where water was more abundant,the association introduced a continuous,staggered planting arrangement for rice,which allowed 2,000 hectares of rice to beplanted every month, year-round. This im-proved water distribution and, according

11

to farmers, it also improved profit marginssomewhat by spreading rice marketingthroughout the year.

OrganizationAfter turnover, the general assemblies of thefarmers’ associations for Coello andSaldaña districts elected boards of directorsto supervise their districts. It was a require-ment that each board had seven memberswith fixed quotas for two categories offarmers—four members with small farmsthe total area of which is less than 20 hect-ares4 and three members with large farms,each with an area of more than 20 hectares.

Directors were elected in a general assem-bly every 2 years. The board recruits andselects the general manager and partici-pates with the general manager in the se-lection or release of other district staff.

After transfer, each board recruited ageneral manager who was an engineer. Thedistricts then became responsible for day-to-day O&M of the systems. The reductionin personnel allowed the management tostreamline the organizational structure bycombining sections and integrating func-tions. In both districts, the general managersupervises an administrative unit and threetechnical units—operations, maintenance,and technical services (see figure 3).

TABLE 4.Basic system parameters, Coello and Saldaña districts, 1993.

Data set Coello Saldaña

A. Irrigation water supply (mm) 948 1,517B. Effective rainfall (mm) 554 793C. Pumped water supply 0 0D. Irrigation duty (mm/day) 8.6 15.5E. Maximum irrigation demand (mm/day) 9 8.5F. Annual demand (mm) 1,097 1,318G. Seasonal maximum irrigation intensity (%) 54.4 93.7H. Annual irrigation intensity (%) 101 161I. Production (rice) (t/ha) 7 7J. Gross margin per year ($/ha) 1,146 1,593K. Gross irrigated area (ha) 25,628 13,975L. Regulation area (ha) 10 10M. Farmer management area (ha) 50 50N. Farm size (ha) 14 7.5O. Capital cost ($/ha) 5,500 5,500

TABLE 5.Basic performance indicators, Coello and Saldaña districts, 1993.

Indicator Formulaa Coello Saldaña

Return to land ($/ha) J 1,146 1,593Return to irrigation ($/100 m3) J/(A+C)/10 12.00 10.50Return to water ($/100 m3) J/(A+B+C)/10 7.60 6.90Return to economy (%) J/O 21 29Fee/cost ratio (%) P/Q 102 109Water use efficiency (WUE) (%) F/(A+B+C) 73 57Relative water supply (ratio) 1/WUE 1.4 1.8Delivery efficiency (%) F/(D+D1) 104 55O&M area per staff member (ha/person) 324 189

aP = irrigation fee, Q = cost of irrigation, D1 = duty for wells. See table 4 for other definitions.

4The required quotas forrepresentation of largeand small farm ownerswas officially droppedin 1994 but the sameproportional constitu-tion of board memberscontinued up to thewriting of this report, in1996.

12

FIGURE 3.Organizational structure, Coello district.

Management Performance after Turnover

In taking over management of the irriga-tion districts, farmers hoped to enhance thecost efficiency and quality of O&M, with-out sacrificing the agricultural productiv-ity and financial and physical sustainabilityof the districts. The performance of the dis-tricts is assessed according to these crite-ria. The government’s main interest was toreduce its own recurrent costs of irrigationwithout sacrificing the productivity of irri-gated agriculture.

Impacts on governmentThe government’s interest in the transferwas initially to accede to political pressuresand later to reduce government subsidiesto the irrigation sector through a nationalpolicy of management transfer. In Coelloand Saldaña, the government was success-ful in discontinuing its subsidies for O&M,

which were costing it about US$9/ha at thetime of turnover. However, it continues tofully finance rehabilitation. If farmers de-fer maintenance, expecting that the govern-ment will finance future rehabilitations, thegovernment may not conserve as muchmoney in the irrigation sector as it wouldlike.

Financial viabilityAfter transfer, the farmers’ irrigation policywas essentially to balance the budget, con-tain the cost of management, and achieve amore responsive irrigation service. This wasonly partially successful. Figures 4 and 5show the changing patterns of revenue andexpenditure in Coello and Saldaña beforeand after turnover. In Coello, during the ini-tial stages of scheme development, expen-ditures exceeded revenues, partially be-

13

FIGURE 4.Revenue and expenditure per hectare, Coello district, 1955�92.a

aIn constant 1988 U.S. dollars (US$1 = Col$333).bData for years 1960 to 1982 are not available.

FIGURE 5.Revenue and expenditure per hectare, Saldaña district, 1953-92.a

aIn constant 1988 U.S. dollars (US$1 = Col$333).bData for years 1960 to 1976 are not available.

cause of external subsidies and develop-ment assistance. The early drop in revenueand expenditure was due to the transitionfrom scheme development to scheme man-agement.

Between 1983 and 1992 in Coello, rev-enues always exceeded expenditures, ex-cept for 1984 (figures 4 and 6). Expendituresrose in real terms by 51 percent, while rev-enues grew by only 44 percent during the

14

period (table 6). Sideline revenue sources—such as rental of farm equipment and dis-trict property, technical services, finesagainst members, sale of materials, andcharges for transporting equipment andmaterials—expanded from about 10 per-cent of revenues to 20 percent between 1983and 1992 (figure 7). Before turnover, rev-enue was taken to at least regional levels ofthe government. Part of the reason farmerswanted to take over management was theirperception that they were financing over-head costs of the government outside thesystem. After turnover, revenues that are inexcess of annual budget costs go into anequipment-replacement fund, are allocatedto the next year’s budget (to help limit therise of fees), and are used by the water us-ers’ association for events and assembliesconnected to public relations. Sideline rev-enues also help limit the level of water fees.

Maintenance costs (including relevantstaff costs) accounted for 55 to 60 percentof total expenditures in the Coello district.This was followed by costs of administra-tion and operation. The proportion of each

to total costs has remained roughly the samesince the early 1980s (figure 8).

Saldaña district was in a weaker finan-cial position than Coello after turnover, withexpenditures exceeding revenues for 6 ofthe 10 years between 1983 and 1992 (fig-ures 5 and 6). However, the district gradu-ally strengthened its position. The level ofrevenue per hectare in Saldaña fluctuatedwidely, but between 1983 and 1992 realgrowth in revenues was 29 percent, com-pared with the 20 percent growth in expen-ditures (table 6).

Both districts improved their financialpositions after turnover, although from op-posite directions. Coello reduced its surplusand enhanced efficiency; Saldaña dimin-ished its pattern of deficits.

Coello and Saldaña both have a fixedarea-based water fee and a volumetric wa-ter fee. These fees vary slightly, dependingon type of crop and whether the farmer is asmallholder or a large operator (i.e., > 20ha). The emphasis of the farmers on con-taining the cost of management resulted ina decline in the area fee after turnover. How-

FIGURE 6.Budget balances in Coello and Saldaña districts, 1983�92.a

aIn constant 1988 U.S. dollars (US$1 = Col$333).

15

ever, the volumetric water fee rose aftertransfer in real terms. In Coello, the area feefor rice (in 1988 US dollars) dropped fromabout US$9/ha in 1976 (at transfer) toUS$5.55/ha in 1993 (see figure 9), while thevolumetric fee for rice rose from aboutUS$0.13/100 m3 (Col$42) in 1976 toUS$0.16/100 m3 in 1992 (see figure 10).5

In Saldaña both area and volumetricfees for rice are higher than in Coello. InSaldaña the area-based fee dropped only

slightly after transfer, from US$9/ha toUS$7.96/ha in 1993 (figure 9). In 1993, thevolumetric fee rose from US$0.13/100 m3

at transfer to US$0.19/100 m3 (figure 10).The difference in the cost of water betweenthe two districts may be due to the fact thatSaldaña has a serious problem of siltationin the intake canal and employs boatsequipped with costly floating draglines todesilt the canal year-round. The most sig-nificant finding from figures 9 and 10 is that

TABLE 6.Total revenue and expenditure, Coello and Saldaña districts, 1983-92 (In 1988 U.S. dollars).a

Coello Saldaña

Year Revenue Expenditure Revenue Expenditure

1983 756,760 633,930 715,920 664,5601984 705,710 711,410 700,900 655,5601985 860,060 660,660 842,940 842,9401986 855,260 825,530 578,380 652,5501987 936,640 791,890 657,360 695,8001988 936,040 795,200 647,150 757,3601989 1,054,650 822,220 733,930 737,5401990 1,063,060 904,200 711,410 713,2101991 1,014,950 948,050 624,320 730,9301992 1,086,790 955,260 923,420 725,230Change1983�92 (%) +44 +51 +29 +20

aUS$1 = Col$333. Only post-turnover data, which were available for both systems, are included, to indicate financialviability after removal of subsidies.

FIGURE 7.Revenue sources, Coello district, 1983�92.

5These fee levels are forrice for smallholderfarmers.

16

FIGURE 8.Types of expenditure, and proportion (%) of each to total cost, Coello district, 1983�92.

trends in both fees reversed directions at thetime of transfer. Volumetric fees rose for tworeasons: (1) the need for revenue to belinked to rising operating costs and (2)board policies of discouraging rice produc-tion and encouraging crop diversification,reducing allocation of water per hectare,and encouraging expansion of irrigatedarea.6

The rising area-based fee reversed along-term decline, while the volumetric feereversed from a decline to a long-term riseafter transfer. Farmer boards in both dis-tricts preferred to charge farmers more inrelation to volume of water used than to thearea planted.

The area and volumetric fees collectedin the Coello district (for all crops) totaled

FIGURE 9.Area water fee for rice, Coello and Saldaña districts, 1967�93.a

aIn constant 1988 U.S. dollars (US$1 = Col$333).6It should be noted thatfee structure differs bycrop and by whetherfarmers are classified as“small” (< 20 ha) or“large.” The rate ishigher for larger farms.The fee for rice for smallfarmers is used hereinas this has been an im-portant crop in bothsystems. In 1993, 76percent of all farmerswere “smallholders” inCoello and 90 percentwere “smallholders” inSaldaña.

17

US$75,990 in 1983 and US$92,041 in 1992(in 1988 dollars). Taking into accountchanges in annual net area irrigated, thismeans the actual cost of irrigation to farm-ers rose slightly from US$50.57/ha in 1983to US$53.86/ha in 1992 (table 7).

Physical sustainability

In both districts, 55 to 60 percent of totaldistrict income goes toward maintenanceof the irrigation network. This percentagedid not change significantly after turnover,

TABLE 7.Annual cost of irrigation to farmers, Coello district, 1983�92.a

Charges paid ($) Net Cost toirrigated farmers

Year Fixed Volumetric Total area (ha) ($/ha)

1983 246,246 439,039 685,285 13,550 50.571984 259,459 414,830 675,375 13,890 48.621985 292,793 486,786 779,579 16,925 46.061986 281,081 458,258 739,339 16,070 46.011987 342,643 491,591 834,234 17,565 47.491988 342,642 494,294 817,417 17,900 45.671989 330,330 584,384 914,714 18,550 49.311990 306,306 577,177 883,483 18,410 47.991991 319,219 496,996 816,216 16,120 50.631992 309,909 520,120 830,030 15,410 53.86

aIn 1988 U.S. dollars (US$1 = Col$333).

FIGURE 10.Volumetric water fee for rice, Coello and Saldaña districts, 1967�93.a

aIn constant 1988 U.S. dollars (US$1 = Col$333).

18

because O&M budgets continued to bebased on the previous year’s budget andcontinued to be reviewed and approved bythe government. However, district manag-ers in HIMAT expressed concern that thefarmers’ strong emphasis on cost reductionwas compromising the physicalsustainability of the systems.

To address this issue, a comprehensiveinventory and examination of all structuresand canal lengths in both systems were con-ducted in 1994. The survey classified canalsections as fully functional, partially func-tional, or dysfunctional. The criteria wereprimarily the extent to which original hy-draulic design conditions were supported.Canal sections that were less than fullyfunctional but that still operate at 70 per-cent of design capacity or higher were clas-sified as partially functional; sections thatoperate at less than 70 percent of designcapacity were classified as dysfunctional.

The Coello survey shows that 68 per-cent of the total canal length was fully func-tional (table 8). This constituted 250 kilo-meters of main, secondary, and tertiary ca-nals. Partially functional canal sections weredistributed relatively evenly between main,

secondary, and tertiary canals. Dysfunctionalcanal lengths were only found at the ter-tiary level, and at this level only 4 percentof the sample lengths were found to be dys-functional. Of the 1,666 total structures ex-amined in Coello, 71 percent were consid-ered fully functional; 15 percent were dys-functional (table 9). Of the dysfunctionalstructures, two-thirds were small flumesused for measuring water at field turnouts(figure 11). These were installed in the re-habilitation period during the late 1970s andearly 1980s. They had not been requestedby the farmers’ association and are rarely usedby the new management. Fifteen percent ofdysfunctional structures were culverts.

In Saldaña, 48 percent of the total net-work was fully functional (table 8). Seventy-nine percent of the main canal was fullyfunctional, whereas only 33 percent of sec-ondaries and 28 percent of tertiaries werejudged fully functional. Forty-four percentof the total canal length was partially func-tional, mainly in secondaries and tertiaries.Dysfunctional sections were located only intertiary and secondary canals with 19 per-cent of observed tertiary lengths and 5 per-cent of secondary lengths being found dys-

TABLE 8.Results of canal maintenance survey, Coello and Saldaña districts, 1994.

Main canal Secondaries Tertiaries Total networkConditiona km % km % km % km %

CoelloFunctional 46.8 68 54.6 76 68.5 63 169.9 68Partially functional 22.3 32 17.3 24 35.7 33 75.3 30Dysfunctional 0 0 0 0 5.0 4 5.0 2Total 69.1 100 71.9 100 109.2 100 250.2 100

SaldañaFunctional 47.8 79 14.6 33 15.2 28 77.6 48Partially functional 13.0 21 27.6 62 30.8 53 71.4 44Dysfunctional 0 0 2.4 5 10.6 19 13.0 8Total 60.8 100 44.6 100 56.6 100 162.0 100

aFully functional: Original hydraulic design conditions are intact, including canal capacity, bed slope, slide slopes,and freeboard. Any canal erosion, breaches, cave-ins, siltation, or weeds are not significant enough to noticeablyinterfere with operational objectives. Partially functional: Original design conditions are compromised by some dete-rioration in bed slopes, side slopes, freeboard, etc., although operational capacity is still at least 70% of originaldesign. Dysfunctional: Operational capacity is below 70% of design capacity, and major rehabilitation is needed.

19

functional. Nineteen percent of the total ter-tiary length was judged dysfunctional. InSaldaña, 69 percent of the 756 structures ob-served were judged to be fully functional;12 percent were dysfunctional (table 9). Ofthe dysfunctional structures, 64 percentwere small measurement flumes at turn-outs, 11 percent were control structures, and10 percent were larger flumes upstream ofturnouts (figure 11).

It is not surprising that the more wa-ter-abundant system has a lower rating inmaintenance. But the large majority ofstructures in both districts is still fully func-tional. In Coello, 98 percent of the total ca-nal length was fully or partially functional;in Saldaña, 92 percent of the canal lengthwas fully or partially functional. This is aremarkable record, given that construction

was completed in 1953, that only limitedrehabilitation had been done in both dis-tricts in the late 1960s and early 1970s, andthat management was transferred to thefarmers’ associations in 1976.

In 1984, HIMAT, in agreement with theusers, conducted feasibility studies on mod-est rehabilitation and system expansion inboth Coello and Saldaña. Some portions ofthe canal and road networks had deterio-rated and were in need of repairs. Drain-age improvement was needed in Saldañaand a supplemental feeder canal wasplanned for Coello. Farmers in Coelloagreed to pay 90 percent of the cost of thefeeder canal, but farmers in Saldaña refusedto pay any of the cost of the rehabilitation.Construction is underway in Coello but notin Saldaña.7

TABLE 9.Results of structure maintenance survey,a Coello and Saldaña districts, 1994.

Coello (no.) Saldaña (no.)Func- Partially Dysfunc- Func- Partially Dysfunc-

Structure tional functional tional Total tional functional tional Total

Headgatesb 30 16 4 50 0 0 0 0Control-dropsc 40 7 18 65 0 0 0 0Control 53 24 2 79 100 17 10 127Drops 121 14 0 135 0 0 0 0Distribution box 15 1 1 17 0 0 0 0Culverts 102 20 38 160 16 3 2 21Siphon 25 6 0 31 40 5 0 45Aqueducts 10 0 0 10 0 0 0 0Radial gates 0 0 0 0 36 3 2 41Box culverts 0 0 0 0 0 0 2 2Flumes 0 0 0 0 0 0 9 9Gates 496 32 9 537 0 0 0 0Bridges 91 11 0 102 120 20 7 147Regulation damd 7 0 0 7 0 0 0 0Canaletase 186 88 166 440 210 98 56 364Measuring 2 17 14 33 0 0 0 0Total 1,178 236 252 1,666 522 146 88 756Proportion (%) 71 14 15 100 69 19 12 100

aFully functional: Keeps design conditions; no elements missing; no modifications apparent or needed. Partiallyfunctional: Some deterioration is evident; minor components missing; requires minor maintenance (painting, grease);still functions with 15% of design requirement. Dysfunctional: Heavy deterioration; broken, damaged, or missingcomponents; is not functional at all.bInclude radial gate structures; all others are sliding gates.cInclude four types: vortex (14), box (23), vertical (23), and chute (5) .dSmall dams that capture drainage, which is reutilized in the system.eFlume-type measuring structure.

7The government re-cently dropped its insis-tence that farmers mustpay for rehabilitationafter turnover. How-ever, it has a new policyto gradually phase outexisting subsidies.

20

Irrigation operationsThere is no indication that the operationalperformance of the Coello or Saldaña sys-tems changed significantly as a result ofturnover. Water continues to be deliveredwithout being measured below the maincanal offtakes. In Coello, the average annualdischarge at the intake varied from 14 m3/sin 1977 to 16 m3/s in 1993, with an averagefluctuation between minimum and maxi-mum discharge levels of 4 m3/s (figure 12).The average annual water supply, ratherthan declining over time, has shown a slightrise. Historical data on discharge at the in-take were not available for Saldaña.

The comparison of the annual volumeof water diverted at the headworks with theaggregate amount of water delivered to alltertiary canals, provides a measure of whatis termed, herein, total conveyance effi-ciency (figure 13). The annual average to-tal conveyance efficiency for 1982 to 1993was 60 percent in Saldaña and 69 percentin Coello. Part of the reason for relativelylow efficiencies may be the reportedly highsediment loads in main canals. This is themost serious management problem inSaldaña and is a major problem in Coelloas well.

As a simple effort to assess equity ofwater distribution along tertiary canals, afield check was made on 15 July 1993 com-paring actual and target discharges intofarm outlets along a tertiary canal locatedat the Florencia secondary canal in the

Saldaña district. The ratio between actualand target discharges is termed the deliv-ery performance ratio, or DPR. From thefirst outlet at the head end to the eighteenthoutlet at the tail, the DPR exhibited a cleardownward trend, ranging from 260 percentat the head to 75 percent at the tail (figure14). One such test cannot verify a pattern,but it does suggest that a distribution prob-lem may exist in Saldaña at the tertiarylevel.8 The distribution arrangement at thetime of inspection was continuous flow.

We have noted above the stable orslightly rising trend in annual average dis-charge at the intake in Coello between 1977and 1993. Figure 15 shows that the annualwater supply delivered for the rice crop roseby 25 percent from about 2,000 mm/ha in1977 to about 2,500 mm/ha in 1991. How-ever, figure 16 shows a 12 percent decline inthe overall average annual volume of waterdelivered per season, from approximately1,100 mm/ha in 1982 to 970 mm/ha in 1991.This trend was influenced by two basicchanges in irrigated agriculture in Coello.

The first is the expansion in gross an-nual irrigated area (total for two seasons)in Coello from approximately 21,000 hect-ares in 1977 to between 27,000 and 37,000hectares in the late 1980s and early 1990s(figure 17). The second is the shift awayfrom rice monocropping to crop rotation asadministered by the district.9 After trans-fer, Coello district excluded the sandy areafrom rice production.10 This partial restric-

FIGURE 11.Composition of dysfunctional structures.

8Distributional inequitymay be partly the resultof the siltation problem,but it would require ad-ditional research to bearthis out.9Coello district estab-lished a zoning system,which restricted con-tinuous cultivation ofrice in certain areas.10Prior to transfer, somesandy areas were re-portedly receiving asmuch as 30,000 m3/haof water supply annu-ally for rice production.After transfer, that wa-ter was reallocated forarea expansion and forheavy soil areas thathad not been receivingenough water beforetransfer.

21

tion on cultivating rice encouraged cropdiversification in the district, with an in-crease in the area planted to cotton, sor-ghum, soybean, and other nonrice crops.

About 19,200 hectares was planted to ricein 1975, the year before turnover. The areadropped to about 16,450 hectares by 1991,a 14 percent decline (figure 18).

FIGURE 12.Discharge at intake, Coello district, 1977�93.a

*Average of 12 months.a1983 data not available.

FIGURE 13.Main canal total conveyance efficiency (ratio of water delivered to water diverted), Coello district,1982�93.

22

In contrast, figures 19 and 20 show therise during this period in area cultivated tocotton and sorghum, the main crops otherthan rice in Coello. Average water deliveriesfor these crops varied widely, with no

apparent increasing or decreasing trend. InCoello, the decrease in the area fee and therise in the volumetric fee may haveencouraged the expansion of irrigated areaand a reduction in the volume of water

FIGURE 14.Water delivery performance ratio along sample tertiary canal, Saldaña district, 15 July 1993.

Note: Distance from tertiary head increases as outlet number increases.

FIGURE 15.Annual water supply for the rice crop, Coello district, 1975�91.a

aMain growing season: March, April, May, June, and July.

23

FIGURE 16.Water delivered per hectare per season, average, Coello district, 1982�91.

FIGURE 17.Gross annual irrigated areaa before and after transfer, Coello and Saldaña districts,1953�93.b

aSummation of irrigated area for both crop seasons.bData for 1960�64 missing.

delivered per hectare. The disciplineimposed by the district to dramaticallyreduce the volume of water delivered perhectare encouraged crop diversification anda substantial growth in irrigated area. Rice

monocropping was unsuitable for Coello’ssandy soils.

In short, operational and maintenanceproblems appear to be more prevalent atthe tertiary and distributary levels than in

24

FIGURE 18.Area cultivated to rice annually and water delivered per hectare,a Coello district, 1975�91.

aPre-season average.

FIGURE 19.Area cultivated to cotton annually and water delivered per hectare,a Coello district, 1975�91.

aPre-season average.

25

the main system, as indicated by the main-tenance survey, DPR analysis, and farmerperceptions. While the problems do notappear to be severe, there is clearly roomfor improvement.

Agricultural productivity

The gradual expansion of irrigated areahalted for about 4 years at the time of turn-over, perhaps because of uncertainties andinefficiencies temporarily created by thechange in management. But the expansionresumed after this apparent learning periodand only began leveling off early in the1990s (figure 17). The rate of expansion hasbeen higher in Coello, where crop diversi-fication has occurred, than in Saldaña,where it has not.

Area expansion was primarily a resultof two factors. First, the tertiary networkwas extended and improved over time. Sec-ond, as farmers gained more experiencewith irrigation and their livelihoods im-proved, they expanded the area irrigatedwithin their farms. Most of the change oc-curred when modern, fertilizer-responsive

varieties were introduced in the 1960s and1970s. Average rice yields increased dra-matically from approximately 2.5 t/ha inthe mid-1950s to approximately 6 t/ha in1976, at the time of transfer (figure 21). Bythe 1990s, average rice yields were between6.5 t/ha and 7.0 t/ha. Most of the increasein yields occurred before transfer, but highyield levels were sustained afterwardsthrough the early 1990s, with a slightlyincreasing trend. We conclude that thetransfer did not have any noticeable detri-mental impact on yields.

Both the cost and value of rice produc-tion declined moderately during the 11-yearperiod from 1984 to 1994. The cost fellfrom about US$380/ha in 1984 to aboutUS$320/ha in 1994 (figure 22). Average netincome for rice production varied widely,from zero to about US$105/ha during theperiod, peaking in 1989 and falling to aboutUS$45/ha in 1994.

The total cost of water relative to thecost of rice production dropped from ap-proximately 4.4 percent during the 1950s,before turnover, to between 3.1 percent (inSaldaña) and 3.3 percent (in Coello), largely

FIGURE 20.Area cultivated to soybean annually and water delivered per hectare,a Coello district,1975�91.

aPre-season average.

26

FIGURE 21.Average annual rice yields before and after transfer, Coello and Saldaña districts. 1953�93.

FIGURE 22.Cost and valuea of rice production, Coello and Saldaña districts, 1984�94.b

aIn constant 1988 U.S. dollars (US$1 = Col$333).bBased on total production data for both systems

27

FIGURE 23.Gross value of output (GVO) in relation to net area and water use, Coello district, 1983�91.a

FIGURE 24.Irrigation cost as a percentage of gross value of output (GVO) and cost of irrigation per hectare tofarmers, Coello district, 1983�91.a

aIn constant 1988 U.S. dollars (US$1 = Col333).

aIn constant 1988 U.S. dollars (US$1 = Col333).

28

due to rising cost of production. However,during the post-transfer period it has beenrising, from 2 percent in 1984 to 3.3 percentin 1993 (in Coello).

In Coello, under post-transfer manage-ment during the 1980s, the total cost of irri-gation11 remained essentially constant atabout US$50/ha in real terms. However, thetotal gross value of output per hectare for allirrigated crops rose over fourfold during thesame period, from US$944 to US$4,300 perha (figure 23). The cost of irrigation as a per-

centage of the gross value of output was rela-tively small and dropped still further, from 5.4percent to only 1.2 percent by 1991 (figure 24).

Coello district also achieved impressivegains in gross value of output per unitof water, which quadrupled fromUS$2.35/100m3 in 1983 to US$9.35/100m3

in 1991 (figure 23).12 This primarily reflectsrises in crop prices, enhanced value per unitof water due to crop diversification, and,to some extent, lower average water con-sumption per hectare.

FarmersInterest in turnover. The initiative for theturnover of management came from thewater users rather than from the govern-ment. Farmers assessed the implications ofturnover and gave their collective approvalin the general assembly meetings in Sep-tember 1976. By the time of turnover, farm-ers were already financing most of the costof O&M and expected that they would beable to keep the irrigation fees from rising,or even reduce them.

Role of government. In 1976, farmers agreedthat HIMAT should continue to provideoversight and advice to the farmers’ asso-ciations about management of the irrigationdistricts. But it soon became apparent thatHIMAT’s role in the districts after turnoverwas more than just oversight. Many farm-ers saw HIMAT as restricting their abilityto further reduce staff and budgets, as theassociations had wanted. Therefore, farm-ers perceived the transfer as being less thanenough to give them full control.

A stratified random sample of 93 farm-ers (44 in Coello and 49 in Saldaña) wasdrawn in 1994, taking half from the upperthird area (head enders) and half from thelower third area (tail enders) in each sys-

tem. It was found that in Coello, only 29percent of the farmers sampled wanted thegovernment to withdraw completely fromworking with the district; 48 percentwanted the government to continue to bepartially involved in assisting the irrigationdistrict; and 21 percent stated that the gov-ernment should take over managementagain (figure 25). In Saldaña, only 14 per-cent favored complete government with-drawal; 68 percent favored continuing par-tial government involvement; and 16 per-cent favored government takeover.

The most commonly mentioned rolesthat sample farmers said they would likethe government to continue to play in theirrigation districts were to provide techni-cal guidance, settle disputes among farm-ers, regulate water allocation in the riverbasin, manage the intake and main canal,and help rehabilitate the system.

Ownership. Sample farmers were also askedwho should own the irrigation infrastruc-ture. In Coello, 76 percent thought the farm-ers’ association should own it; in Saldaña,80 percent thought the farmers should ownit. Only 19 percent in Coello and 18 percentin Saldaña thought that the governmentshould own the structures.

11Total revenues from allwater charges dividedby net area irrigated peryear.12Unfortunately, compa-rable data were notavailable for Saldaña.

Perspectives of Stakeholders

29

Outcomes of turnover. Sample farmers wereasked, “Has the 1976 transfer of manage-ment for the irrigation district from the gov-ernment to the farmers’ organization im-proved, worsened, or not changed much ofthe management of the irrigation district?”In Coello, 53 percent responded that it hadnot changed much, 40 percent said it hadimproved, and only 7 percent said it hadworsened. In Saldaña, 39 percent said man-agement had improved after turnover, 36percent said it had not changed much, and25 percent said it had worsened. In Saldaña,7 of the 11 farmers who stated managementhad worsened were tail enders. In Coello,there was no significant difference betweenhead and tail enders.

The most common ways in which man-agement had improved, according to thefarmers of both systems were (1) in com-munication between district staff and farm-ers, (2) in responsiveness of district staff tofarmers, and (3) in water distribution.About 70 percent of the sample farmers inCoello and 91 percent in Saldaña stated thatthey had attended a district associationmeeting within the last year.

Impacts on maintenance. Sample farmerswere asked, “Has the functional conditionof the secondary canal that delivers waterto your field improved, worsened, or stayedabout the same over the past 10 years?” InCoello, 81 percent said it had stayed aboutthe same, 17 percent said it had improved,and only 2 percent said it had worsened. InSaldaña, 73 percent said it had stayed inabout the same condition, 15 percent saidit had worsened, and 13 percent said it hadimproved (with no significant difference be-tween head and tail enders).

Impacts on operations. Sample farmers wereasked, “Over the last 2 years was the irri-gation water delivered to your farm alwaysenough for your crop water requirement,enough most of the time, not enough mostof the time, or never enough?” In Coello,45 percent said it was always enough and32 percent said it was enough most of thetime. Only 20 percent said it was not enoughmost of the time, and 2 percent said it wasnever enough. In Saldaña, 59 percent saidwater was always enough, and 31 percentsaid it was enough most of the time. Only 4

FIGURE 25.Farmer perspective about withdrawal of the irrigation agency.

Note: N=44 in Coello and 48 in Saldaña.

30

percent said it was not enough most of thetime, and 2 percent said it was neverenough. In Coello, a surprising 96 percentsaid water was delivered to their fields ontime all or most of the time. In Saldaña, 90percent said water was delivered on timeall or most of the time. This question didnot address the issue of change, but it diddemonstrate that since turnover, wide-spread satisfaction exists among farmersabout water distribution.

Regarding water theft or disputes overwater, only six sample farmers (14%) inCoello stated that they were aware of inci-dents in the last 2 years. Of the sample farm-ers, 86 percent were not aware of any suchcases. In Saldaña, only two sample farmersstated that they were aware of any theft ordispute over the last 2 years. Forty-sevensample farmers (96%) were not aware ofany such occurrence.

Impacts on agricultural productivity and prof-itability. Farmers did not indicate that man-agement turnover had had a significantimpact on either the productivity or theprofitability of agriculture.

District staffDistrict managers expressed concern thatthe strong farmer disposition toward cost

reduction was causing some decline in ser-vice and that this would eventually resultin the visible deterioration of the system.They said experienced senior personnel hadbeen replaced by younger, inexperiencedstaff, key technical positions had been elimi-nated or merged, and little or no expendi-ture was being made in training or replace-ment of equipment or structures. Somenoted occasional undue influence by large-scale farmers over field operations staff inthe distribution of water.

Agency staffAt first, members of the HIMAT staff at thedistrict and higher levels generally were re-sistant to the transfer. They perceived thatjobs would be lost and the role and powerof the agency would diminish as a result ofmanagement turnover, first in Coello andSaldaña, and eventually in other districtsas well. For several years after turnover, theagency pressured the farmers’ associationsin Coello and Saldaña against releasing staffand reducing budgets. This resulted in law-suits between the farmers and the agency,mainly over the issue of releasing staff. Af-ter the land development law of 1993 wasenacted, the government granted full au-thority over district staff and budgets tofarmers’ associations.

Conclusion

Perception

Most farmers see turnover as having pro-duced a more responsive and cost-efficientmanagement. Most, however, favor con-tinuing a limited role for the agency, pri-marily in providing technical advice and inhelping with resolving disputes. The ma-jority believe that the association shouldown the irrigation infrastructure. However,most farmers appear satisfied with the per-

formance of O&M tasks. Many believe thatmanagement performance, especially costefficiency, would have improved even morehad the users been granted full control overstaff and budgets after turnover. Associa-tion board members perceived that the par-tial turnover brought only partial benefits.

Professional staff in the districts are lesssanguine about the results, expressing con-cern that cost-cutting measures are compro-mising the quality of O&M. The agency was

31

concerned about the implications of turn-over on agency staff and budgets.

Main resultsThe following are the study’s main conclu-sions about performance changes after turn-over.

1. Management turnover achieved thegovernment’s objective of discontinu-ing subsidies and making the districtsfinancially self-reliant for O&M. Thedelegation of administration, however,did not result in full turnover of author-ity to the farmers’ associations. Thegovernment agency continued to exer-cise partial influence over budgets andstaffing. Nevertheless, after turnover,the districts began to gradually reducestaff, while continuing virtually thesame level of management intensity asbefore. Most sample farmers felt thatcommunications with district staff andtheir responsiveness to farmers had im-proved after turnover.

2. The districts have been only partiallysuccessful in containing costs. Staff lev-els have been reduced 35 percent sincetransfer. However, the cost of irrigationremained relatively constant for a de-cade after turnover. Coello district hasbeen financially solvent ever since turn-over, with a decreasing margin of bud-get surplus over time. It has also diver-sified its revenue sources beyond wa-ter charges. Saldaña, however, has hadcontinuing problems balancing its bud-get, but has made progress toward sol-vency with growth in revenues outpac-ing growth in expenditures over time.Both districts raised irrigation fees forrice over time and costs of irrigation tofarmers rose in real terms—although asa percentage of the total cost of rice pro-duction, or gross value of output, thecost of irrigation dropped substantially.

In Coello, financial viability has beenachieved by spreading the cost of irri-gation among more farmers throughexpansion of area, by increasing volu-metric fees for rice, and by diversifica-tion of revenue sources.

3. Nineteen years after the transfer, only2 percent of total canal length in Coelloand 8 percent in Saldaña were dysfunc-tional (mostly in tertiary canals). Of allwater control and measurement struc-tures, only 15 percent in Coello and 12percent in Saldaña were dysfunctional.The vast majority of dysfunctionalstructures were field outlet measure-ment structures (which were not nor-mally used). We conclude that the dis-tricts have been able to sustain preven-tive maintenance so far. And owing tostatements by sample farmers, we con-clude that system maintenance has notyet been adversely affected by turnover.The intensive and costly maintenanceinvestment the districts have been ableto support, relative to the serious silt-ation problem, has been impressive.

4. However, because the government re-tained ownership of the scheme assets,farmers insist that the governmentshould finance future rehabilitationand modernization. Neither associationis raising a capital replacement fund. Itis apparent that this arrangement is pre-venting the associations from achiev-ing complete local financial sustainabil-ity. Although the systems have beenwell maintained until the present, thismay lead to some deferred mainte-nance in the future.

5. After turnover, the farmers’ associa-tions soon established new crop rota-tion and irrigation scheduling arrange-ments designed to permit extension ofirrigated area while decreasing the av-erage amount of water delivered per

32

hectare. The Coello district was able tosubstantially expand its area irrigatedby steadily delivering less water perhectare and diversifying cropping.Saldaña, which has heavier soils, con-tinued to irrigate only for rice, thoughit staggered planting dates to spread ir-rigation demand over the year. How-ever, the study indicates there is a mod-erate problem of inequitable water dis-tribution to tail enders, which is duepartly to siltation and some lack of con-trol at the tertiary level.

6. It is apparent that the transfer did notinhibit long-term expansion of the areairrigated or the ability of irrigated ag-riculture to sustain high levels of riceyields. Despite rising costs of agricul-tural production and a decline in cropprices, yields and area irrigated roseslightly after transfer and then re-mained stable.

7. Perhaps the most important finding ofthe study is that increases in the grossvalue of output per hectare and per unitof water increased dramatically, whilethe cost of irrigation to farmers re-mained roughly the same after turn-over. Irrigation constituted a relativelysmall and declining proportion of thetotal cost and value of production. Im-provements in economic performanceafter turnover can be only partially at-tributed to broader factors such as cul-tivation improvements and crop prices.After turnover, new district policies torestrict rice production in sandy areasand reduce average volume of waterdelivered per hectare supported cropdiversification and improved the valueof irrigated output. Cost-containmentpolicies such as reductions in staff andcessation of flow of funds outside theschemes undoubtedly helped preventincreases in the cost of irrigation tofarmers.

Literature Cited

Garcés-Restrepo, Carlos, and Douglas L. Vermillion. 1994. Results of irrigation management transfer inCoello and Saldaña Irrigation Districts, Colombia. Paper presented at the Internal Program Review,International Irrigation Management Institute, November, 1994, Colombo, Sri Lanka.

HIMAT (Instituto Colombiano de Hidrologia, Meteorologia y Adecuación de Tierras). 1991a. Distrito deRiego Coello. Subdireccion de Adecuación de Tierras. Bogota, Colombia.

HIMAT. 1991b. Distrito de Riego Saldaña. Subdireccion de Adecuación de Tierras. Bogota, Colombia.Ministerio de Agricultura. 1993. Ley de Adecuación de Tierras. Bogota, Colombia.Plusquellec, Herve. 1989. Two irrigation systems in Colombia: Their performance and transfer of management to

user’s associations. Policy, Planning and Research Working Paper Series No. 264. Washington, D.C.: WorldBank.

Vermillion, Douglas L. 1992. Irrigation management turnover: Structural adjustment or strategic evolu-tion? IIMI Review 6(2): 3–12.

Vermillion, Douglas L., and C. Garcés-Restrepo. 1994. Irrigation management transfer in Colombia: A pilotexperiment and its consequences. Short Report Series on Locally Managed Irrigation, No. 5. Colombo, SriLanka: International Irrigation Management Institute.

33

Research Reports

1. The New Era of Water Resources Management: From “Dry” to “Wet” Water Savings. David Seckler,1996.

2. Alternative Approaches to Cost Sharing for Water Service to Agriculture in Egypt. C.J. Perry, 1996.

3. Integrated Water Resource Systems: Theory and Policy Implications. Andrew Keller, Jack Keller, andDavid Seckler, 1996.

4. Results of Management Turnover in Two Irrigation Districts in Colombia. Douglas L. Vermillion andCarlos Garcés-Restrepo, 1996.

Research Report

Results of Management Turnoverin Two Irrigation Districtsin Colombia

Douglas L. VermillionandCarlos Garcés-Restrepo

INTERNATIONAL IRRIGATION MANAGEMENT INSTITUTEP O Box 2075 Colombo, Sri Lanka

Tel (94-1)867404 • Fax (94-1) 866854 • E-mail IIMI@cgnet.comInternet Home Page http:/ /www.cgiar.org

ISSN: 1026-0862ISBN: 92-9090-329-5

4