Equity aspects of irrigation development: Evidence from two systems in the hills of Nepal

This article was downloaded by: [York University Libraries]On: 11 November 2014, At: 14:09Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

International Journal of WaterResources DevelopmentPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/cijw20

Equity aspects of irrigationdevelopment: Evidence from twosystems in the hills of NepalRabi K. Maskey a , Karl E. Weber b & Rainer Loof ca Research Associate, Asian Institute of Technology , Bangkok,Thailandb Professor and Dean, Asian Institute of Technology , Bangkok,Thailandc Associate Professor, Asian Institute of Technology , Bangkok,ThailandPublished online: 02 May 2007.

To cite this article: Rabi K. Maskey , Karl E. Weber & Rainer Loof (1994) Equity aspects ofirrigation development: Evidence from two systems in the hills of Nepal, International Journalof Water Resources Development, 10:4, 431-443, DOI: 10.1080/07900629408722645

To link to this article: http://dx.doi.org/10.1080/07900629408722645

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoeveras to the accuracy, completeness, or suitability for any purpose of the Content. Anyopinions and views expressed in this publication are the opinions and views of theauthors, and are not the views of or endorsed by Taylor & Francis. The accuracyof the Content should not be relied upon and should be independently verifiedwith primary sources of information. Taylor and Francis shall not be liable for anylosses, actions, claims, proceedings, demands, costs, expenses, damages, and otherliabilities whatsoever or howsoever caused arising directly or indirectly in connectionwith, in relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms

http://www.tandfonline.com/loi/cijw20

http://www.tandfonline.com/action/showCitFormats?doi=10.1080/07900629408722645

http://dx.doi.org/10.1080/07900629408722645

& Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Dow

nloa

ded

by [

Yor

k U

nive

rsity

Lib

rari

es]

at 1

4:09

11

Nov

embe

r 20

14

http://www.tandfonline.com/page/terms-and-conditions

http://www.tandfonline.com/page/terms-and-conditions

Water Resources Development, Vol. 10, No. 4, 1994 431

Equity Aspects of Irrigation Development: Evidencefrom Two Systems in the Hills of Nepal

RABI K. MASKEY1, KARL E. WEBER2 & RAINER LOOF3

1Research Associate, 2Professor and Dean and 3Associate Professor, Asian Institute ofTechnology, Bangkok, Thailand.

ABSTRACT Equity is of great relevance to developing irrigation. Its two dimensions --horizontal in regard to water distribution to farmland and vertical in terms ofproductivity differences between farm categories -- are studied in two irrigation systems,an old farmer-managed and a new government-agency-managed system. Equity in thedistribution of irrigation water differed between abundant and scarce water supplyconditions. Paddy grown during the monsoon season when water is available inabundant quantity shows that there was a reasonable degree of fairness in its distri-bution between head and tail reach farmland. However, wheat grown in the dry seasonwith limited supply of water rendered evidence of unfair distribution demanding bettermanagement of irrigation water. The analysis of vertical equity shows that small farmsare more efficient than large ones in increasing productivity through the use of irrigationfacilities.

Equity as a Perplexing Issue in Irrigation Development

Equity issues are receiving the serious attention of irrigation professionals.Certain approaches to the development of equity-orientated measures of per-formance are documented in the related literature (Lenton, 1984; Sampath, 1984,1988a, 1988b; Abernethy, 1986). Wickham & Valera (1979), in their study in thePhilippines, identified issues in the performance of an irrigation system andconcluded that not only should on-farm water use be effective but also itsdistribution should be equitable and reliable along each water delivery system.The concept of equity is considered to be both very simple and complex. It isconsidered simple because everyone is aware of it, and it is recognized ascomplex because there is hardly any degree of consensus on how best tomeasure it (Sampath, 1988a). This expert analysed the usefulness of differentmeasures of dispersion such as range, relative mean deviation, variance,coefficient of variance, standard deviation of logarithms, Gini coefficient andTheil's information measure for describing inequity. He stressed the point thatany study on the equity problem of an irrigation project should be in line withthe objectives and purposes of the project, as stated and perceived by theauthorities in charge and its beneficiaries (Sampath, 1988c). In this regard, twofairly distinct types of equity concerns can be identified. The first is the 'vertical'

Dow

nloa

ded

by [

Yor

k U

nive

rsity

Lib

rari

es]

at 1

4:09

11

Nov

embe

r 20

14

432 R. K. Maskey et al.

equity involving a 'main' view of society which examines whether an effectextends to a particular social or economic class at the expense of another (e.g. thedichotomy between small and large farms). The second type, known as 'horizon-tal' equity, reflects 'micro' considerations involving 'fairness' among individualswho are perceived to be equals in some sense (Small & Carruthers, 1991). Thisis the issue where 'fairness' is of major concern in situ.

In the context of the present study, the horizontal equity problem of irrigationmanagement relates to inequity in water distribution in different locations and,due to its effect, the average difference in yields between head and tail areas.Usually, the main assumption underlying equity-based measures is based onvalue judgements used in coining or selecting the corresponding definition. Forthis study, as in many other studies, equity is defined as the condition of equalamounts of water being delivered, resulting in an equal average yield per unitarea throughout an irrigation system. There are other possibilities, indeed, whichalternatively include equal degree of water stress on the crop; equal access to theright to try to obtain water (the warabandi system in India and Pakistan); andequal amount of water per household. There are many more definitions in thisrespect. In the context of vertical equity, the issue is seen in relation to the typesof farms such as the distinction between small and large, or owner and tenantfarmers.

In conducting this study, objective measures of water distribution providinginformation on the amounts of water delivered over time at specific locations(volumetric measurement) were not taken into consideration. Instead, the fre-quency of watering is taken as a proxy for actual volumetric measurement.Another indicator used is the average yield per unit of land in different locationsand by different farm categories.

Data Collection

Two medium-scale irrigation systems located in the same river valley—Chaurasikulo and Arnapurna—were studied. Chaurasi kulo is an old established farmer-managed system (FMS) serving about 200 ha of land, while Arnapurna is anewly implemented government-agency-managed irrigation system (GMS) thatcovers 300 ha of valley lowland with its five branch canals. Both systems liewithin the Yamdi watershed and have similar socio-economic and environmen-tal characteristics. The master sample of 375 farm households includes 75household heads acting as respondents sampled from the traditional irrigationsystem (FMS), 138 from the government-agency-managed irrigation system(GMS) and 37 who were enjoying the benefits of both systems. To cover theunirrigated surroundings, 35 farm household heads were sampled from contigu-ous locations within the valley which were not covered by either irrigationsystem, plus 90 respondents from ridge settlements of the same watershed. Forthe initial analysis of 'horizontal' equity, only the GMS and FMS were con-sidered, whereas for the analysis of 'vertical equity' all groups were taken intoconsideration. In the old FMS, there is a distinct difference between head and tailreaches. In the new GMS, given its layout, the head, middle and tail reaches ofthe main canal are further divided into upper and lower sections as indicated inTable 1.

Dow

nloa

ded

by [

Yor

k U

nive

rsity

Lib

rari

es]

at 1

4:09

11

Nov

embe

r 20

14

Equity Aspects of Irrigation Development 433

Table 1. Distribution of farms by upper andlower sections of head, middle and tail reachesin the government-agency-managed irrigation

system.

Reach

Head

Middle

Tail

Branch canal no.

1

2 & 3

4&4A

Section

UpperLowerUpperLowerUpperLower

No. of farms

141438381717

Horizontal Equity

Equity in the distribution of irrigation water as an operational objective of themanagement of canal irrigation systems is focused on what is conventionallytermed 'head reach' and 'tail end' and their cultivators. The topographicalreferences 'head reach' and 'tail end' refer to the distance, in gravity flowirrigation schemes, between the farmer's irrigated plot and the point at whichwater is issued into the canal system.

In discussing the issue of equity, there are two major reasons given, amongseveral others, why 'head reachers' almost always obtain more water per unit ofland than 'tail enders'. The first reason is that the hydraulics of water is such thatvolume and pressure in irrigation canals are greater at the head reach. Second,it is physically much easier for head reachers to steal water compared with tailenders. Design and management can reduce but not eliminate these causes ofinequity (Moore et ah, 1983). There are other factors, such as topography, soiltype, irrigation system layout and water management, which might have someeffects that reinforce the factor 'distance'. By and large, the equity issue is muchmore prominent in water scarcity situations compared with those of relativeabundance of irrigation water supply.

Equity Issue in Paddy Production with Abundant Supply of Water

Paddy is grown at the time when irrigation water is in abundant supply. Thereis no difference in yield between head and tail reaches in a system, although theyield differs between these two systems (see Table 5).

The frequency was assessed at which farmers irrigated their fields. There is adifference in frequency of irrigation between the old and new systems. Thedifference is not large, however, between the head and tail reaches of the oldsystem and among head, middle and tail reaches of the new system (Table 2).Even if the head, middle and tail reaches of the new system are further dividedinto upper and lower sections, the frequencies of irrigation show no differencebetween the sections. There is, however, a significant difference between theupper and lower sections of branch canal No. 4 at the tail reach of the main canal(Table 3). The cases are further aggregated1 and presented as 'upper section' and'lower section' of all head, middle and tail reaches. The result shows that the twomeans are not significantly different (Table 4).

Dow

nloa

ded

by [

Yor

k U

nive

rsity

Lib

rari

es]

at 1

4:09

11

Nov

embe

r 20

14


Table 2. Frequency of paddy irrigation by systemand reach.

System Reach Frequency of irrigation

Old system (FMS)

New system (GMS)

HeadTail

HeadMiddle

Tail

4.374.185.795.675.53

Notes: The f-test conducted for the old system indicates that thetwo means are not significantly different at the 0.10 level.Scheffe's test conducted for the new system indicates that notwo groups are significantly different at the 0.10 level.

Table 3. Frequency of paddy irrigation in the govern-ment-agency-managed irrigation system by reaches and

sections.

Reach

Head

Middle

Tail

Branch canal no.

1

2 & 3

4&4A

Section


Frequency of irrigation

6.005.575.605.716.00*5.06'

Note: *The Mest result indicates that the two groups are significantlydifferent at the 0.10 level.

Table 4. Frequency of paddy irrigation in upper and lowersections of all head, middle and tail reaches of the main canal

in the government-agency-managed irrigation system.

No. of Frequency ofLocation cases irrigation

Upper section of head, middle and tail reaches 69 5.78Lower section of head, middle and tail reaches 69 5.78

Note: The f-test result indicates that the two means are not significantly differentat the 0.10 level.

Similar results were obtained for the overall yield differences between thesevarious locations. Table 5 shows the paddy productivity in the different reaches.Although the overall yield seems to be higher in the head reaches of bothsystems, the difference between those yields is not statistically significant. Evenif the branch canals of the head, middle and tail sections of the new system weredivided into upper and lower sections, the corresponding yield differences werenot statistically significant (Table 6). Finally, the productivity differences of the

Dow

nloa

ded

by [

Yor

k U

nive

rsity

Lib

rari

es]

at 1

4:09

11

Nov

embe

r 20

14


Table 5. Paddy productivity in different reaches by system.

System

Old system (FMS)

New system (GMS)

Reach

HeadTail

HeadMiddle

Tail

Productivity(muri/ropani)2

2.98a

2.82a

2.68b

2.50b

2.28b

Productivity(kg/ha)

29742814267524952275

CV

30.0434.8525.0126.4424.51

Notes: CV = coefficient of variation.a The f-test result shows that the productivity of paddy in the head and tailreaches does not vary significantly at the 0.05 level.b Analysis of variance indicates that no two groups are significantly different atthe 0.05 level.

Table 6. Paddy productivity in the government-managed irrigationsystem by reaches and sections of the branch canals.

Branch

Head

Middle

Tail

Branchcanal no.

1

2 & 3

4&4A

Section


No. ofcases

141438381717

Productivity(muri/ropani)

2.70872.65492.58022.41822.33762.2228

Productivity(kg/ha)

270326502575241323332218

CV

29.820.026.226.722.227.2

Notes: CV = coefficient of variation.Scheffe's test conducted under analysis of variance to observe the mean differences of paddyyields in different sections of each branch canal indicates that no two groups are significantlydifferent at the 0.05 level.

Table 7. Difference in paddy productivity between upper and lower sectionsof the head, middle and tail reaches of the main canal in the government-

agency-managed irrigation system.

Section

Upper section of head,Lower section of head,

middlemiddle

andand

tailtail

reachesreaches

No. ofcases

6969


2.54652.4181

Productivity(kg/ha)

25412413

CV

26.425.7

Note: The f-test result indicates that the tw o means are not significantly different at the 0.10 level.

upper and lower sections of head, middle and tail reaches of the new systemwere analysed. The result further supports the preceding findings that there isno significant difference in the paddy yield between these sections (Table 7).

These results show that there is a reasonable degree of fairness in thedistribution of water between the head and tail reach farmers during themonsoon season, or when water is available in abundant quantity. The resultsdo not, however, hold true for the wheat crop. The following analysis for wheat

Dow

nloa

ded

by [

Yor

k U

nive

rsity

Lib

rari

es]

at 1

4:09

11

Nov

embe

r 20

14


shows that there is a difference in performance between various sections. As formonsoon paddy, its better performance in relation to the equity factor is mainlydue to the good precipitation during the surveyed year, bringing sufficientamounts of water for irrigation into the canals. It also happens to be the areawith the highest precipitation in the whole country. Since both systems are of therun-of-the-river type, a good amount of water flowing through the canal ensuresa fair amount of water input into its different reaches and sections. It should beemphasized that in situations of water abundancy, low physical water useefficiencies may be economically efficient. In such a situation, it is economicallydesirable to substitute low-value water for high-cost management inputs. It isnot correct to conclude that a system is performing poorly simply because it haslow water-use efficiency, in the traditional engineering sense (Barker et al. 1984).Even with the good amount of irrigation water, the application of water wasstatistically different between GMS upper and lower sections of branch canalNo. 4. This, however, did not show up in the yield analysis, where the yielddifferences were not statistically significant. Martin & Yoder (1987), in theirstudies in Nepal, concluded that at the extremes, where water is either veryscarce or extremely abundant, increased management efforts through a strongerorganization are either unproductive or unnecessary. Water abundance due tohigh precipitation may be one of the reasons for farmers' poor participation inthe irrigation system studied. Burns (1993), however, cautioned that untimely orinsufficient rainfall in the wet season may bring drought and there might be anemergency when farmers destroy their own systems in a desperate attempt toextract water to save their crops. This had reportedly happened earlier in thestudy area, especially when drought had prevailed during the transplantingperiod. Burns compared this situation with a desert irrigation system which iseasier to design and less troublesome to operate because irrigation is defined bythe amount of water available, rainfall-dependent standing crops are absent, andall farmers are aware of their water allocation and rights.

Equity Issue in Wheat Production with Scarce Supply of Water

Wheat is sown in the dry season (November-December) throughout the studyarea. However, winter irrigation is feasible only in the area covered by the new,government-agency-managed system. This crop, similar to paddy, is mostlygrown in khet or lowland areas. The area covered by the wheat crop is smallerthan that grown with paddy. Wheat, unlike paddy, is not grown at the timewhen irrigation water is in abundant supply, for its need for water is low. Thevolume of water in the system is lower compared with the monsoon season.Even in this situation, there is no statistical difference in the frequencies at whichthe crop is irrigated in the head, middle and tail reaches (Table 8). If the head,middle and tail reaches of the new system are further divided into upper andlower sections, the frequencies of irrigation do show a statistical differencebetween its middle and tail reaches (Table 9). The analysis was carried furtherby aggregating these cases into upper and lower sections of head, middle andtail reaches. The result shows that the difference between the two means isstatistically significant, confirming the fact that there is a difference in the supplyof water to different sections of the system (Table 10).

Similar results were obtained when wheat yields of different sections werecompared. Simple aggregation of the new system into head, middle and tail

Dow

nloa

ded

by [

Yor

k U

nive

rsity

Lib

rari

es]

at 1

4:09

11

Nov

embe

r 20

14


Table 8. Frequency of wheatcrop irrigation in the govern-ment-agency-managed system

by reaches.

Reach Frequency of irrigation

Head 3.21Middle 3.32Tail 2.94

Note: The analysis of variance resultindicates that no two groups aresignificantly different at the 0.10 level.

Table 9. Frequency of wheat irrigation in the govern-ment-agency-managed irrigation system by reaches

and branch canal sections.

Reach

Head

Middle

Tail

Branch canal no.

1

2 & 3

4&4A

Section


Frequency of irrigation

3.073.363.62a'b

3.03a

3.23C

2.60"*

Notes: a The analysis of variance result indicates a significantdifference at the 0.10 level.bThe f-test result indicates that the two means are significantlydifferent at the 0.05 level.cThe f-test result indicates that the two means are significantlydifferent at the 0.05 level.

Table 10. Frequency of wheat irrigation in the upper and lowersections of the head, middle and tail reaches of the main canal

in the government-agency managed system.

Section Frequency of irrigation

Upper section of head, middle and tail reaches 3.4Lower section of head, middle and tail reaches 3.0

Note: The f-test result indicates that the two groups are significantly different atthe 0.05 level.

reaches did not show any yield difference (Table 11). When these were aggre-gated further into upp*er and lower sections within these reaches, yields werefound to be statistically different between the upper and lower sections of themiddle and tail reaches (Table 12).

Dow

nloa

ded

by [

Yor

k U

nive

rsity

Lib

rari

es]

at 1

4:09

11

Nov

embe

r 20

14


Table 11. Wheat productivity in the govern-ment-agency-managed irrigation system by

reach.

Productivity ProductivityReach (muri/ropani) (kg/ha) CV

HeadMiddleTail

0.8841.0541.058

121414471452

29.4536.9736.25

Note: The result of the analysis of variance indicates thatno two groups are significantly different at the 0.05 level.

Table 12. Wheat productivity in the government-agency-managed irrigationsystem by reach and branch canal section.

Reach

Head

Middle

Tail

Branch canal no.

1

2 & 3

4&4A

Section


No. ofcases

141434341715


0.94650.82111.17643

0.93243

1.1709b

0.9316b

Productivity(kg/ha)

129911271615128016071279

CV

31.425.534.236.632.138.9

Notes: CV = coefficient of variation.a The f-test result indicates that the two means are significantly different at the 0.05 level.b The f-test result indicates that the two means are significantly different at the 0.10 level.

Table 13. Wheat productivity in the upper and lower sections of the head,middle and tail reaches of the main canal in the government-agency-managed

irrigation system.

Section

Upper sectionLower section

of head,of head,

middle andmiddle and

tailtail

reachesreaches

No. ofcases

6565


1.12540.9074

Productivity(kg/ha)

15451246

CV

33.935.4

Note: The /-test result indicates that the two means are significantly different at the 0.001 level.

Table 13 also shows the statistical difference in yields between the upper andlower sections, aggregating all head, middle and tail reaches.

hi a water abundance situation, the concern about equity is insignificant,although paddy requires large quantities of water. In a water scarcity situation,however, unfair distribution of water was observed. It necessitates better man-agement of irrigation water. This situation prevails at the time when wheat isgrown in an area that is smaller than the paddy area. If the area under wheatis to be expanded, there is an obvious need for better water management.Several studies (such as those by Wickham & Valera, 1979; Bhutta & van der

Dow

nloa

ded

by [

Yor

k U

nive

rsity

Lib

rari

es]

at 1

4:09

11

Nov

embe

r 20

14


Velde, 1992) show that better operational procedures at the distributary level cansubstantially improve water supply conditions in the tail reaches.

Vertical Equity

Complementarity to horizontal equity, vertical equity examines how an irri-gation system affects a particular social or economic structure, in this study theeffects on different categories of farms. It also addresses the differences betweenirrigated and non-irrigated farms. To analyse the situation and evaluate selectedindicators, it is necessary to define the term 'equity' in the social and economiccontexts. In principle, each farm household has a right to its share of the totalirrigation water available in proportion to the size of its total landholding. Thisdefinition itself is biased when seen from the perspective of vertical equity. If theshare of water is to be proportional to the landholding size, the larger farms arebenefiting more. There are other systems through which the irrigation waterdistribution is regulated such as on the basis of issuing shares to farmers inproportion to their contribution to the initial investment costs of the system(Martin & Yoder, 1986; Molin & Paudyel, 1992). There are yet other systemswhich give emphasis to equal amounts of water per household; and some stressequal access to the right to try to obtain water such as under the warabandisystem. There is a most remarkable institutional innovation in Bangladesh wherethe irrigation assets are entrusted to cooperatives of the landless who operate theirrigation devices and equipment and sell water, in some instances for cash andin others for a share of the crop, to landowners within their command area(Wood, 1984).

Equity Issue in Respect of the Size of Landholding

In the valley locations under study, each farm household has a right to its shareof the total irrigation water available in proportion to the size of its landholding.In this situation, farmers with larger landholdings have a greater opportunity touse water and spin-off inputs, especially capital, in such a way as to obtainlarger production volume compared with small farms. This has often beenreported and criticized as the negative impact of the 'Green Revolution', from anequity perspective. The 'Green Revolution' strategy stresses the increased use ofmodern machinery, pesticides, fertilizers and irrigation to which larger farmshave better access.

The issue of land productivity—farm size in relation to returns to scale—istested by estimating the coefficient of land in the following crop function:

ylog — = b0 + bi

A;

where Y is the total crop output; b0 is the constant term; b, is gross elasticity ofland; and X,- is the net sown area.

This equation is the log transformation of the Cobb-Douglas productionfunction used in the estimation of input elasticity by the ordinary least-squaremethod. This form of equation was used because of the ease and uniqueness ofinterpretation of regression parameters in bringing out the relationship betweensize of holding and productivity. Here, the value of crop output per ropani isassumed to be a function of the net sown area in both paddy and wheat

Dow

nloa

ded

by [

Yor

k U

nive

rsity

Lib

rari

es]

at 1

4:09

11

Nov

embe

r 20

14


Table 14. Relationship between area size and productivity in paddy cultivationby location.

Farmland location

Whole watershedOld system (FMS)New system (GMS)Overlap between systemsArea without irrigationRidge cultivation area

Constant

0.43350.36550.46680.53200.3087

- 0.0130

Elasticity (&,-)

- 0.4680a

0.1080-0.1134a

- 0.1547a

- 0.04810.1125a

J-value

-4.8531.404

-3.257-2.176-0.663

2.280

d.f.

37373

136353388

R2

0.060.030.070.120.010.06

F value

23.55a

1.9710.61a

4.74a

0.445.20a

Note:a Indicates that the corresponding parameters are statistically significant at the 0.05 level.

Table 15. Relationship between size of wheat area and productivity of wheat.

Farmland location

Whole watershedOld system (FMS)New system (GMS)Overlap between systemsArea without irrigationRidge cultivation area

Constant

0.0196- 0.0475

0.15110.0601

- 0.1405- 0.3082

Elasticity (b,)

- 0.3330a

- 0.1660- 0.3466a

- 0.1528- 0.0291 lb

- 0.1014

f-value

-6.417-1.267-4.894-1.235-1.670-1.258

d.f.

2755283292479

R2

0.130.030.220.050.100.02

F value

41.17a

1.6123.95a

1.532.79b

1.58

Notes:a Indicates that the corresponding parameters are statistically significant at the 0.05 level.b Indicates that the corresponding parameters are statistically significant at the 0.10 level.

cultivation. This is used to estimate the elasticity of crop output per ropani withrespect to land. The sign for coefficient b, is called the 'gross elasticity', becauseland alone is the independent variable which indicates the relationship betweenland productivity and farm size. The elasticity parameter %', if negative,indicates an extreme form of equitable distribution, in which the level ofproductivity declines as the size of holding increases; a zero value for %'indicates a lack of association between farm size and irrigation water distri-bution. A positive value indicates some degree of inequitable distribution, inwhich productivity increases with the size of holding.

The analysis was carried out to observe the effect of landholding size onproductivity of paddy and wheat, where productivity is taken as the indicatorfor the efficient use of irrigation water, together with other inputs, to increaseproductivity by different farm categories. Here, it is hypothesized that if largefarms are benefiting from irrigation more intensively, then this group shouldhave a positive relationship between farm size and productivity. However, inthe cases of both paddy and wheat cultivation, there is an inverse relationshipbetween farm size and productivity (Tables 14 and 15).

The elasticities calculated for most cases show a decline in productivity witha 1% change in land input. Four locations in Table 14 and three in Table 15 havestatistically significant relationships in terms of the 'F ratio' at 95% and 90%confidence levels, respectively. The R2 values in these cases were found to bevery low. In other words, there are certain factors which could better explain theproductivity of these crops rather than the differences in farm size alone. It isworth noting that the farm size effect in these regression equations for most

Dow

nloa

ded

by [

Yor

k U

nive

rsity

Lib

rari

es]

at 1

4:09

11

Nov

embe

r 20

14


cases is negative. Even with the availability of irrigation, small farms were foundto be efficient in that they were producing higher yields. In general, thisphenomenon is observed in many developing countries of Asia and LatinAmerica that are characterized by widely differing natural and climatic condi-tions, types of soil, agrarian structures and cropping patterns. This is in contrastwith the finding reported by Sampath (1992) who, with his vast number ofstudies on equity aspects, concluded for his study in India that canal irrigationdoes not play any role in reducing the degree of inequity in the distribution ofwater across different farm-size categories.

All location-specific groups have a negative elasticity, indeed, for land size inrespect of wheat productivity (Table 15). There was, however, a positive elastic-ity found in the old system and ridge cultivation area for the paddy crop (Table14). In the old system, there is only one large farm. There, medium-size farmstended to have higher yields than small farms. In ridge cultivation, larger farmshad higher yields compared with small farms. A possible explanation is that thevery small farms, where the scale of operation is too minute as to allow for anyform of efficiency, probably had a total social factor productivity that was lowerthan those of any other farm categories. This is true for ridge cultivators wherethe quality of land is so poor that there is a need for larger farm size so as tomeet the local scale of operation, especially for the efficient production of anycrops.

Especially with irrigation, small farms were found to be much more efficientcompared with large farms. This rejects the hypothesis that larger farms aremore efficient in increasing productivity with irrigation facilities than smallones. It is not only the productivity but also the cropping intensity which wasfound to be higher among smaller farms irrespective of their location for, withan increase in farm size, there was a decline in cropping intensity.

Discussion and Conclusions

Most of the time water would be distributed on a continuous basis in both theGMS and FMS under study. In the FMS, there were no effective controls andvirtually no devices installed in the distribution canals and turnouts. It wasfound that paddy was only cultivated when water was abundant in supply.Even in this situation there was great competing demand for water to pushactivities like seedbed preparation and transplanting. Water distribution rulesand regulations are necessary as the circumstances highlight. Moreover, thereshould be clear-cut rules for unfavourable situations when precipitation isinsufficient. In winter, there is usually a scarcity of water. Hence, rules andregulations for the distribution of water are a must, not only from the efficiencypoint of view but also in an equity perspective. The principles of waterdistribution may be based on any one or a combination of the followingmethods: (i) fixing proportionate allotments by volume or by duration underrotation; (ii) a priority system based on location (such as near head or tail reachof a canal), farm characteristics, or economic value of the crop grown; and/or(iii) on a demand basis. Rotational irrigation can result in water savings of20-30% (Wang, 1973). The water thus saved can be used to supply or sup-plement resources in areas where there is a shortage of water. In addition,rotational irrigation encourages plant growth, leads to economies in fertilizeruse, and almost eliminates water disputes. In some of Nepal's successful

Dow

nloa

ded

by [

Yor

k U

nive

rsity

Lib

rari

es]

at 1

4:09

11

Nov

embe

r 20

14


irrigation systems, a wooden notched weir is used to distribute water in a timelyrotation called sancho (Martin & Yoder, 1986). This could be replicated in othersystems in the hills to ensure the fair and efficient delivery of water.

The analysis of the issue of vertical equity in the two irrigation systems understudy rejects the hypothesis that large farms are more efficient in increasingproductivity with the use of irrigation facilities than small ones. If the sizes oflandholdings within an irrigated area vary grossly, it is likely, indeed, that thebenefits from irrigation will likewise be distributed in a highly unequal fashion,if and wherever there is a direct link between land rights and effective waterrights. There are several examples of how to tackle these problems. In somecases, water rights are separated from land rights. In some communal irrigationsystems of Nepal, water rights were given according to the contribution pro-vided in the form of labour at the initial construction stage of the system (Martin& Yoder, 1986). This method could be suited to solving the vertical equityproblems rather than the providing of irrigation water on the basis of the size oflandholding.

Notes

1. Upper sections of all the branch canals are grouped into one category and, likewise, their lowersections are grouped into another.

2. 1 muri paddy = 49.90 kg; 1 muri wheat = 68.64 kg; 1 ropani = 0.05 ha.

References

Abernethy, C.L. (1986) Performance Measurement in Canal Water Management: A Discussion, ODI-IIMI,Irrigation Management Network Paper 86/2d (London, ODI).

Barker R., Coward Jr., E.W., Levine, G. & Small, L.E. (1984). Irrigation Development in Asia: Past Trendsand Future Directions, Cornell Studies in Irrigation No. 1 (Ithaca, NY, Cornell University).

Bhutta, M.N. & van der Velde, E.J. (1992) Equity of water distribution along secondary canals inPunjab, Pakistan, Irrigation and Drainage Systems, 6, pp. 161-177.

Burns, R.E. (1993) Irrigated rice culture in monsoon Asia: the search for an effective water controltechnology, World Development, 21 (5), pp. 771-789.

Lenton, R., (1984) A note on monitoring productivity and equity in irrigation systems in: NiranjanPant (Ed.) Productivity and Equity in Irrigation Systems (New Delhi: Asish).

Martin, E.D. & Yoder, R. (1986) Institutions for Irrigation Management in Farmer Managed Systems:Examples from the Hills of Nepal (Kathmandu: International Irrigation Management Institute).

Martin, E.D. & Yoder, R. (1987) Organizational structure for resource mobilization in hill irriga-tion systems, in: Irrigation Management in Nepal, Research Papers from the National Seminar(Kathmandu, IIMI).

Molin, N. & Paudyel, D.P. (1992) Community irrigation opportunities and challenges, in: Water,Environment and Management Conference Pre-Prints, 18th WEDC Conference (Kathmandu: WEDCand NEA).

Moore, M.P., Abeyratne, F., Amarakoon, R. & Farrington, J. (1983) Space and the generation of socioeconomic inequality on Sri Lanka's irrigation schemes, Marga, 7 (1), pp. 1-32.

Sampath, R.K. (1984) Income distribution impacts of irrigation water distribution policy, WaterResources Research, 20 (6), pp. 647-654.

Sampath, R.K. (1988a) Equity Measures for irrigation performance evaluation, Water International, 13(1), pp. 25-32.

Sampath, R.K. (1988b) Some Comments on Measures of Inequity in Irrigation Distribution, ODI/IIMIIrrigation Management Network Paper 88/2f (London, ODA).

Sampath, R.K. (1988c) Measures of managerial performance of irrigation system, in: Proceedings ofInternational Conference on Irrigation System Evaluation and Water Management, Vols I and II (Wuhan,Wuhan University of Hydraulic and Electric Engineering).

Sampath, R.K. (1992) A farm-sizewise analysis of irrigation distribution in India, The Journal ofDevelopment Studies, 29 (1), pp. 121-147.

Dow

nloa

ded

by [

Yor

k U

nive

rsity

Lib

rari

es]

at 1

4:09

11

Nov

embe

r 20

14


Small, L.E. & Carruthers, I. (1991) Farmer-Financed Irrigation: The Economics of Reform (Cambridge,Cambridge University Press).

Wang, Y.T. (1973) Development impact and management of irrigation in Taiwan, in: RegionalWorkshop on Irrigation Water Management (Manila, Asian Development Bank).

Wickham, T.H. & Valera, A. (1979) Practices and accountability for better water management in:Taylor, D.C. & Wickham, T.H. (Eds), Irrigation Policy and the Management of Irrigation Systems inSoutheast Asia (Bangkok, Agricultural Development Council).

Wood, G.D. (1984) Provision of irrigation services by the landless -- an approach to agrarian reformin Bangladesh, Agricultural Administration, 17 (2), pp. 55-80.

Dow

nloa

ded

by [

Yor

k U

nive

rsity

Lib

rari

es]

at 1

4:09

11

Nov

embe

r 20

14

Documents

Equity aspects of irrigation development: Evidence from two systems in the hills of Nepal