total productivity and technical efficiency of watermelon atfarm level

13

Raziah Mat Lin

Total productivity and technical efficiency of watermelon atfarm level(Produktiviti total dan kecekapan teknikal tanaman tembikai di peringkat ladang)

Raziah Mat Lin*

Key words: total productivity, technical efficiency, watermelon, farms

AbstractA primary survey involving 49 watermelon farmers operating in two seasons in2004/2005 was conducted with the objectives of assessing the farm productivity andefficiency. Data analysis indicated that the individual farm total productivity (TP)values ranged from 0.66 to 4.91. The mean TP for the individual sample farms was2.05. The TP for the watermelon sub-sector in 2004/2005 was 1.78. The averagecontribution of labour to income was relatively higher than the average contributionof capital. Ordinary least square procedure (OLS) undertaken revealed that youngerand higher income farmers tend to be more productive respectively than the olderand lower income farmers. The Kopp and Timmer technical efficiency (TE) computedindicated that the sample farms had a mean efficiency level of 46% with standarddeviation of 0.18. The productivity and efficiency of the watermelon sub-sectorcould be improved by increasing yield and revenue through the adoption of newtechnology, new variety and good seeds and reducing production costs. It is stronglyrecommended that irrigation system be installed in the watermelon farms to reducethe labour cost of watering and to ensure good harvests.

*Economic and Technology Management Research Centre, MARDI Headquarters, Serdang, P.O. Box 12301,50774 Kuala LumpurE-mail: [email protected]

Economic and Technology Management Review. Vol. 1 No. 1 (June) 2006: 1328

IntroductionWatermelon is a very popular short-term non-seasonal fruit and has been classified undermajor fruits by the Ministry of Agricultureand Agro-Based Industry (MOA). These majorfruits are being promoted for commercialplanting due to their potential in generatingincome to the farmers and the economy.

The total acreage of watermelon inPeninsular Malaysia has been stabilized ataround 5,000 ha since 1990s, except in theyear 1991 to 1993, whereby the total acreagewas found to be higher (7,0008,000 ha). Thismay be due to the larger acreage of farmsbeing transplanted with oil palm at that time,and watermelon is known to be a popular cashcrop during early establishment of the oil palmtrees.

In the year 2002, the estimatedwatermelon production was about 70,000tonnes. Despite its relatively small area ascompared to the other major fruits, watermelonis known to be a significant contributor toexport earning of the country. Malaysiasexport of watermelon accounted for about 2%of the worlds exports. Singapore is the majorimporter, accounting for 70% (RM29.3million) of the total export value in 2003(RM42 million). The other major importersare Hong Kong, Indonesia, Brunei and Taiwan.

The local demand for watermelon(including other melons) in 2002 was estimatedat 127,000 tonnes valued at RM170 mil.(Anon. 2006). The inadequate local supplywas supplemented by imports (mainly other

14

Total productivity and technical efficiency of watermelon at farm level

melons) from Australia, China, Japan andThailand.

The establishment of World TradeOrganization (WTO) and the rapidliberalisation of agriculture trade have openedthe agricultural sector to increasingcompetition and new market opportunities.Malaysia is a net exporter of watermelon. Thequestion currently arises as whether Malaysiacan improve or maintain its competitivenessin the glocal markets. Unless farmers arewilling to increase productivity and improveefficiency in their operations, Malaysia maylose its market shares to other competitors.More so for the watermelon sub-sector whichis totally dependent on imported seeds.

The previous study on the totalproductivity (TP) of the agriculture food sectorfocused on a few commodities, which includedvegetables, paddy, watermelon and broilerchickens. The TP of the watermelon sub-sectorin the year 2000 was found to be 1.36(Anon. 2002).

A study on the efficiency of selectedfish-based food product sectors in 1996 and1998 indicated that this sub-sector wasdominated by the Small and MediumIndustries(SMI) (Raziah 2003). The averageefficiency of the firms has decreased in 1998(0.1380) as compared to that in 1996 (0.3447).Raw materials were found to be the mostimportant factor affecting production, followedby labour and capital.

Another efficiency study on selected fruitfarms in Malaysia (Abu Kasim and Md. Yunus1994) found that the mean technicalefficiencies for starfruit, papaya, and guavafarms were 0.51, 0.60 and 0.44 respectively.This study found that starfruit and guava farmswere more labour intensive as compared topapaya farms. The operation that utilised themost labour was crop maintenance, especiallyfruit wrapping, which accounted for almosthalf of the labour input for starfruit and guavafarms.

The total productivity of the agriculturesector in 19911994 was estimated at 1.98(Mad Nasir et al. 1998), which was consideredas favourable for an active investment.

However, in this case the contribution wasmainly from oil palm.

Watermelon was chosen for this study,due to the threat of increasing competitionfrom cheaper producing countries, such asChina and Thailand especially when theAFTA/WTO becomes fully materialised in thenear future.

In this paper, the TP and technicalefficiency (TE) of the watermelon farms arediscussed, the factors affecting the productivityand efficiency of the farms are highlighted,and finally policy measures to improve theproductivity and efficiency of the sub-sectorare recommended.

Materials and methodsSource of dataPrimary surveys by the enumerators and farmrecord keeping by the respondents wereadopted to gather data in the study. Each ofthe respondents was given a book for him/herto record the daily activities involved in theproduction processes. Subsequently, theenumerators transferred the data from therecord books to the structured questionnaires.The data was for the 2004/2005 operation year.

Stratified sampling design was adoptedto identify the districts covered in the survey.The major districts producing watermelon wereselected based on the latest statistics publishedby the Department of Agriculture Malaysia(DOA). The respondents were then selected atrandom with the cooperation of the extensionagents.

The data collected were base line data onfarmers, farms background, outputs, prices ofoutputs, inputs consumed and prices of inputs.

Theoretical frameworkProductivity measurement Productionactivities can be defined as processes oftransforming various inputs into products.Productivity is the ratio of output to input.Single factor productivity also known as partialproductivity (PP) is output divided by a singlefactor. For example, labour productivity is theratio of output to labour and capitalproductivity is the ratio of output to capital.

15

Raziah Mat Lin

In a production process, many factors ofproduction or inputs are used simultaneously.Changes in the quality and quantity of otherinputs affect the productivity measure in asingle input. For example, in manufacturingprocesses, both labour and machineries areused together. In case of a firm havingsophisticated machineries, the number ofemployees still remained the same, therebyincreasing the per capita output. Thus, wewould observe an increase in labourproductivity. It is possible that labourers workless, but still produce more due to the newmachines used. In such case, the single factorproductivity fails to take into considerationthe use of other factors in the productionprocesses.

A more comprehensive measure of theefficiency of production process is to use theratio of output to aggregate inputs. This is thebasic concept behind the total factorproductivity (TFP). TFP measures theproductivity of a composite of all factors ofproduction, mainly by looking at thecontribution of capital and labour. It ismeasured relative to a base point. TFP reflectsthe efficiency with which factors of productionare jointly used to produce the output.

Total productivity (TP), on the otherhand, measures the efficiency of productionprocess by taking into consideration the totaloutput divided by the total inputs at a specifiedtime (Oguchi, N., Faculty of Commerce,Shensu Univ., pers. comm. 2004). Using thismeasurement, comparison of the efficiencylevel of firms within a commodity and acrosscommodity could be made. The efficiency ofindividual farms involved in plantingwatermelon for example, in the operation year2004 could be compared. The comparisons ofthe efficiency of the watermelon, papaya andthe pineapple sub-sectors could also be madefor a specified year.

Efficiency measurement A productionfunction (PF) is a quantitative or mathematicaldescription of the various technical productionpossibilities faced by a firm. The PF gives themaximum output (s) in physical terms for each

level of the inputs in physical term. There is afrontier setting a limit to the maximum possibleoutput which could be obtained. A firmproducing less than the maximum possibleoutput lies below the production frontier andis regarded as inefficient.

There are four popular methods used formeasuring and computing technical efficiency(Farrel 1957; Forsund et al. 1980). Most ofthe methods involve the construction of a best-practice frontier and the measurement ofinefficiency relative to this frontier. The firstmethod involves the construction of adeterministic non-parametric frontiersometimes called the pure programmingapproach (Farrel 1957). The second approachwhich was first suggested by Farrel (1957),involves the construction of a deterministicparametric frontier. It has been continued byAigner and Chu (1968) and Forsund andHjalmarsson (1974). The third approach, asproposed by Afriat (1972), uses statisticaltechniques to estimate a deterministic statisticalfrontier. Another approach of determiningtechnical efficiency involves the estimation ofa stochastic parametric frontier using specifiedfunctional forms and uses statistical techniquesto estimate the frontier (Schmidt and Lovell1979).

Model specificationTotal productivity In this study, the TPapproach is adopted to measure the individualwatermelon farm efficiency within the sub-sector in the year 2004/2005. The TP model isrepresented as

TPT2004 = Output = PiQi (1)Input pij.qij

Where,TP = total productivity;T = commodity watermelon;2004 = in the year 2004/2005;i = individual farm/firm, i = 1..49;j = input, j = 1m;P = output price;Q = output quantity;

16


p = input price; andq = input quantity.

Technical efficiency The deterministicstatistical frontier proposed by Afriat (1972)was used to determine the efficiency of thefirms. This approach involves the assumptionsof functional forms (usually Cobb-Douglas)for the frontier and estimations using thecorrected ordinary least square (COLS)procedure. The functional form was firstestimated using ordinary least square procedure(OLS), and then the constant term wascorrected by shifting it up until no residualwas positive and at least one was zero. Theextent of a particular observation inefficiencywas measured by the ratio of actual output topotential output, with the latter given by thefrontier itself.

The frontier production function Mathe-matically, the non-linear Cobb-Douglasfunction is written as

Y = f (x) eu O (2)

To estimate the function using OLS, thefunction has to be transformed into linear form,

nn Y = + i n Xi i O (3)

i=1

Where, Y = output; = intercept;i = estimated coefficients;Xi = set of input variables; and i = one-sided residual.

In this study, the output was specified as thegross income (INC) of the individual farms inone crop cycle during the 2004/2005 operationyear. The input variables were the cost ofcapital (CAP), the cost of labour (LABOUR)and the cost of other inputs (OTH) whichincluded costs of seeds, fertilizers, insecticides,herbicides and plastic covers in the specifiedperiod of operation.

Measure of technical efficiency Koppsmeasure of technical efficiency (TE) comparesthe actual level of input used with the levelthat could be used if a firm is operating on thefrontier, given the actual output generated bythe firm using the same ratio of input.Timmers measure of TE for a specific firm isdefined as the ratio of actual output to potentialoutput, given the levels of input used by thatparticular firm. This measure indicates theamount of extra output that could be obtainedif a firm is operating on the production frontier.

To generate Kopp and Timmers TEmeasures, denotes CAP*, LABOUR*, andOTH* as the optimal usage of inputs CAP,LABOUR and OTH respectively. For eachfirm, for the given output INC,

n CAP* = n INC 1 1 n LABOUR 2 n OTH CAP CAP

nbii=1

Where,1 = + U maxn LABOUR* and n OTH* can be calculatedin a similar way.

Kopp TE = CAP* = LABOUR* = OTH* 1 (4)CAP LABOUR OTH

The technically efficient firm will have CAP*= CAP, LABOUR* = LABOUR, andOTH* = OTH, while other firms tend to useinput more than required.

Timmer TE = n INC n INC* = INC 1 (5)INC*

The technically efficient firm will meet theconstraint INC* = INC. The other firms tendto produce output less than their potentials.

Results and discussionThe data used in this study consisted ofproduction information on samples of 49 farmsinvolved in producing watermelons (twoseasons) in the operating year 2004/05.

17

Raziah Mat Lin

Background of the respondents The 49respondents involved in the study were selectedfrom the districts of Kluang, Kota Tinggi, Muarand Pasir Putih. Most of the respondents wereMalays in the age group of 4060s and mostof them had completed their secondaryeducation. The gross monthly income for mostof the respondents was between RM500RM1,000. A majority of the respondents had49 family members living with them, andmost of them owned 0.4 1.6 ha of land(Table 1).

Background of the farms The most popularvarieties planted were New dragon followedby Hitam Manis, Quality and other varieties.Out of the 49 farms, 8 farms were categorisedas big farms with the planted acreages of 8 haand more. The rests were small farms with anarea of 0.41.6 ha. Most of the respondents(65%) rented the lands from the owners whilethe others were using their own land (Table 2).

Capital About 22% of the respondentsindicated having tractors to help in theoperation specifically for land preparation. Themost common equipment available wassprayers which were owned by about 76% ofthe respondents. Irrigation system was criticalin watermelon cultivation especially duringhot season and more than 55% of therespondents were having them either on theirown or subsidised by the DOA. Other capitalitems that contributed to the watermelonfarming activities were buildings andtransports. The buildings were mostly farmshades and the transports were mostlymotorcycles. The value of the capital itemsare depicted in Table 3.

Labour The summary statistics of thenumbers and costs of labour in watermelonproduction based on the sample farms areshown in Table 4. The labour work force wasclassified into family labour, wage labour andcontract labour. In the small farms, contractlabour was usually employed during the landpreparation. Both family labour and wage

labour were equally important in watermeloncultivation especially during the harvesting.

Input costs The inputs relevant to thewatermelon production are seeds, fertilizers,pesticides, herbicides and plastic covers forweed control. The summary statistics for thetotal input costs is shown in Table 5.

Other costs The other costs which are notdirectly involved in watermelon cultivationare cost of land (land tax or rental), insuranceand tax (vehicles), fuel (vehicles andmachinery), repair (vehicles, machinery, etc),depreciation and others not classifiedelsewhere. However, these expenses wereincluded in the calculation of TP (Table 6).

Yield and revenue Depending on the farmsize, the gross yield from each farm rangedfrom a low of 6,800 kg to a high of 633,200 kgper crop cycle. The average yield per ha wasabout 21,700 kg. The average yield per ha forboth the big and small farms was consistent(Table 7).

Depending on the variety and quality ofthe watermelon produced, the farm-gate priceranged from a low of 20 sen per kg for theseeded variety to a high of 80 sen per kg forthe seedless variety. The average income per hawas about RM10,900. However, the averagemargin per ha for the small farms was aboutRM6,700, which was relatively higher thanthat of the big farms which has an averagemargin of about RM4,500 (Table 8).

Contribution of labour and capital toincome For the overall sample farms, theaverage contribution of labour to income wasrelatively higher than the average contributionof capital, which was 55% and 45%respectively. The average contribution ofcapital to income for the big farms wasrelatively higher (53%) than labour (47%).However, the average contribution of labourto income for the small farms as expected washigher (57%) than the contribution of capital(43%). The investment in capital in the bigfarms as expected was higher than that of the

18


Table 1. Summary of the socio-economic variables for the sample of 49 respondentsinvolved in watermelon cultivation in 2004/05

Variables Per cent by categories Total no. of respondentsDistricts Kluang = 10.2% 49

Kota Tinggi = 4.08%Muar = 2.04%Pasir Puteh = 83.67%

Age (years) 30 = 4.65% 433140 = 20.93%4150 = 39.53%50 = 34.88%

Ethnic Malay = 83.67% 49Chinese = 14.28%Others = 2.04%

Education SPM/STPM = 42.22% 45PMR = 26.66%Sekolah rendah = 28.83%Others = 2.22%

Gross monthly 500 = 4.25%] 47income (RM) 5011,000 = 53.19%

1,0012,000 = 25.53% 2,000 = 17.02%

Family members 3 = 8.33% 48(stay together) 46 = 47.92%

79 = 41.66% 10 = 2.08%

Land ownership 0.4 = 10.25% 39(ha) 0.41.2 = 56.40%

1.252.0 = 20.51% 5 = 12.82%

Table 2. Summary of the varieties planted, area planted and land tenured in a sampleof 49 watermelon farms in 2004/05

Variables Per cent by categories Total no. of respondentsVarieties New dragon = 30.61% 49

Hitam manis = 20.40%Quality = 12.24%Others = 36.73%

Area planted* 0.4 0.8 = 55.10% 49(ha) 0.85 1.20 = 24.48%

1.25 1.60 = 4.08% 8 = 16.32%

Land tenure Own land = 24.48% 49Rent = 65.30%Others = 10.20%

* 8 ha = Big farms< 8 ha = Small farms

19

Raziah Mat Lin

Table 3. Status of capital in a sample of 49 watermelon growers in 2004

Capital item Estimated value (RM) Percentage of respondentshaving the item (%)

Building 29,500 24.48Transport 1,122,610 73.46Machinery and equipment 229,583 75.71Irrigation system 124,945 55.10

Table 4. Summary statistics of the labour and its costs in a sample of 49 watermelonfarms in 2004/05

Variables Summary statistics Total no.of respondentsMin. Max. Mean

Family labour (no.)Big farms 0 0 0 8Small farms 0 40 3.29 41Overall farms 0 40 2.76 49

Wage labour (no.)Big farms 0 10 7.6 8Small farms 0 5 1.76 41Overall farms 0 10 2.71 49

Contract labour (no.)Big farms 0 40 6.5 8Small farms 0 1 0.05 41Overall farms 0 40 1.06 49

Labour costs (RM)Big farms 6,540 28,180 18,590 8Small farms 578 4,465 1,593 41Overall farms 578 28,180 4,368 49

small farms (Table 9). The contributions ofboth capital and labour were equally importantin the watermelon sub-sector.

Total productivity The TP of an individualfarm ranged from a low of 0.66 to a high of4.91 (Table 10). The mean productivity forthe watermelon farms was 2.05. The smallfarms were found to be relatively moreproductive than the big farms with the meanproductivity of 2.10 and 1.80 respectively(Table 11).

Most of the farms were operating at theTP>23 (43%), followed by the TP>12(38%). The rests were either having high TPor not productive (Table 12). The total outputdivided by the total input for all the 49 samplefarms, gave a TP of 1.78 for the watermelon

sub-sector, which was considered as high foran active investment. However, the TP ofwatermelon in the year 2004 was relativelyhigher than the TP in 2000 (1.36). This couldbe due to better crop husbandry, good varietyand quality seeds. As has been mentionedearlier, water is critical in watermeloncultivation and more than 55% of the farmswere irrigated using drip system. This factorcontributed to improving TP.

Socio-economic factors influencingproductivity OLS procedure was used todetermine factors that affect the TP of thewatermelon farms. The dependent variable wasspecified as productivity indices (PVITIT) andthe independent variables were the respondentage (AGET), area planted (ACRET) and gross

20


Table 5. Summary statistics of the input costs involved in a sample of 49 watermelonfarms in 2004/05

Variables Summary statistics (RM) Total no.of respondentsMin. Max. Mean

Big farms 8Seed costs 1,500 7,250 3,417Fertilizer costs 4,000 57,210 26,576Pesticide costs 2,690 43,000 13,107Herbicide costs 0 3,360 1,250Plastic costs 0 37,440 10,811

Total 16,830 101,190 55,162Small farms 41

Seed costs 100 900 421Fertilizer costs 272 3,080 1,172Pesticide costs 78 551 228Herbicide costs 0 338 69Plastic costs 312 1,225 511

Total 1,406 4,495 2,403Overall farms 49

Seed costs 100 7,250 910Fertilizer costs 272 57,210 5,319Pesticide costs 78 43,000 2,331Herbicide costs 0 3,360 1,262Plastic costs 0 37,440 2,193

Total 1,406 101,190 11,017

Table 6. Summary statistics of other costs (RM) involved in a sample of49 watermelon farms in 2004/05

Variables Summary statistics Total no.of respondentsMin.* Max. Mean

Land nil 8,250 462 49Insurance and tax nil 3,137 236 49Fuel nil 9,596 665 49Repair nil 7,000 279 49Depreciation nil 4,433 998 49Others nil 460 28 49*Indicates that some respondents did not pay for the variables

monthly income (INCT). The estimatedregression model is shown in the equation (6).The figures in parenthesis indicated thet-values of the estimated coefficients.

PVITIT = 2.4720 0.0093 AGET 0.0353 ACRET + 0.0001 INCT (6)(7.256) (1.335) (2.523) (2.729)

Where,

PVITIT = productivity indices for individualfarm;

AGET = age of individual farmer;ACRET = acreage of individual farm;INCT = gross monthly income for

individual farmer;R2 = 0.1911r = 0.4371, p = 0.05

21

Raziah Mat Lin

Table 7. Summary statistics of the gross yield, yield and price of a sample of 49watermelon farms in 2004/05

Variables Summary statistics Total no.of respondentsMin. Max. Mean

Gross yield (kg)Big farms 170,000 633,200 313,518 8Small farms 6,800 38,134 19,801 41Overall farms 6,800 633,200 67,755 49

Yield (kg/ha)Big farms 11,900 32,200 22,000 8Small farms 8,525 47,100 21,654 41Overall farms 8,525 47,100 21,700 49

Price (sen/kg)Big farms 25 75 50 8Small farms 20 80 50 41Overall farms 20 80 50 49

Table 8. Summary statistics of the gross income, income per ha and margin per haof a sample of 49 watermelon farms in 2004/05


Gross incomeBig farms 69,600 314,790 145,020 8Small farms 3,100 25,210 10,234 41Overall farms 3,100 314,790 32,240 49

Income per haBig farms 4,292 17,344 9,941 8Small farms 3,830 30,900 11,100 41Overall farms 3,830 30,900 10,900 49

Margin per haBig farms 1,480 10,986 4,515 8Small farms 904 24,987 6,706 41Overall farms 1,480 24,987 6,348 49

The farmer's age and the farm size tend tohave negative influence on productivity.Younger farmers and small farm sizes werefound to be more productive than the olderfarmers and the big farms as seen in the survey.Younger farmers tend to be more responsivetowards new technology and innovation.Unlike the older farmers, they usually havethe energy required to undertake farmactivities.

Watermelon is a short-term and risky cropthat needs intensive crop husbandry. Smallfarms would be easier to look after to ensuretheir productivity than big farms. However,

the gross monthly income of the farmerspositively influenced the productivity. Thefarmers with higher gross monthly income tendto be more productive than those with lowergross monthly income. Those with betterincome would have the resources needed toacquire sufficient inputs in their farmoperations, thus increasing their productivity.The R2 for the estimated model was 0.19(r = 0.4371) which is considered as acceptablefor a cross-sectional data. This means that19% of the variations in the productivityindices could be explained by the three

22


Table 9. Summary statistics of costs for capital, labour and their contributions to income

Variables Summary statistics Total no. ofrespondentsMin. Max. Ave.

Capital cost (RM)Big farms 7,000 40,500 22,389 8Small farms 0 55,085 4,278 41Overall farms 0 55,085 7,235 49

Labour cost (RM)Big farms 6,540 28,180 18,590 8Small farms 578 4,465 1,593 41Overall farms 578 28,180 4,368 49

Share of capital in income (%)Big farms 0.37 0.74 0.53 8Small farms 0 0.97 0.43 41Overall farms 0 0.97 0.45 49

Share of labour in income (%)Big farms 0.26 0.63 0.47 8Small farms 0.03 1.00 0.57 41Overall farms 0.03 1.00 0.55 49

Table 10. The total productivity of a sample of 49 watermelon farms in 2004/05

Code no. Input (RM) Output (RM) Total productivityT2 119235 93675 0.79T3 34226 69600 2.03T4 68595 82800 1.21T5 67377 74700 1.11T6 132138 86850 0.66T7 105685 245685 2.32T8 65646 192067 2.93T9 93891 314790 3.35T10 7120 11850 1.66T11 8476 25210 2.97T12 5105 5989 1.17T13 8055 19120 2.37T14 3428 5858 1.71T15 4138 6804 1.64T16 5727 11660 2.04T17 4928 11462 2.33T18 6669 8050 1.21T19 5749 7180 1.25T20 5054 9750 1.93T21 4226 9820 2.32T22 3054 4757 1.56T23 3108 5330 1.71T24 4414 4348 0.99T25 3298 5770 1.75T26 3385 9180 2.71

(Cont.)

23

Raziah Mat Lin

T27 2924 9420 3.22T28 4088 10439 2.55T29 4786 10188 2.13T30 6739 8602 1.28T31 4021 8730 2.17T32 5691 15700 2.76T33 7272 20722 2.85T34 4156 9314 2.24T35 3553 7100 2.00T36 6305 10000 1.59T37 4658 10011 2.15T38 6839 21447 3.14T39 5100 25017 4.91T40 4641 5420 1.17T41 2634 4600 1.75T42 5710 9775 1.71T43 4226 8672 2.05T44 3135 4617 1.47T45 4197 3100 0.74T46 3095 8759 2.83T47 4853 10435 2.15T48 3639 8946 2.46T49 4825 8700 1.80T50 4918 17747 3.61Minimum 2634.00 3100.00 0.66Maximum 132138.00 314790.00 4.91Mean 18055.97 32240.14 2.05Std. deviation 32740.76 61996.29 0.8248

Total productivity = Total output = 1,579,766.70 = 1.78Total input 884,742.50

Table 10. (Cont.)Code no. Input (RM) Output (RM) Total productivity

Table 11. Summary statistics of the total productivity of a sample of 49 watermelonfarms in 2004/05


Big farms 0.66 3.35 1.80 8

Small farms 0.74 4.91 2.10 41

Overall farms 0.66 4.91 2.05 49

24


independent variables, where the F-value is26.40.

Efficiency analysis Using the primarysurvey data collected from 49 watermelonfarms, the average Cobb-Douglas productionfunction was estimated using OLS (7).

n INC = 2.3534 + 0.6055 n LABOUR +(5.099) (4.952)0.1056 n CAP + 0.2371 n OTH (7)

(2.829) (2.131)

Where,

INC = gross income for eachindividual farm in one cropcycle;

LABOUR = total labour costCAP = cost of capital for each

individual farm in one cropcycle;

OTH = cost of other inputs for eachindividual farm in one cropcycle;

R2 = 0.8691;F value = 7.5056; max = 0.8124.

The letters n in front of each variablerepresent natural logarithm and max denotesthe largest positive estimated residual recordedby one of the sampled farms. Figures inbrackets indicate the t-values of the regressioncoefficients.

All the independent variables (CAP,LABOUR and OTH) were found to besignificant in explaining the variation in thedependent variable (INC). However, labourwas found to be the most important factor

Table 12. Productivity level of a sample of 49 watermelon farms in 2004/05

Productivity level Index Percentage of farms Total no. of respondentsNot productive

25

Raziah Mat Lin

In general, the sample farms were using inputshigher than required, thus reducing thetechnical efficiency.

The average efficiency of about 46%would indicate that there is room for improvingits average efficiency level to at least 50%efficient, in order to be competitive in theglocal markets.

The average efficiency for themanufacturing sector from 19922001 wasabout 56% (Nor Aini 2004), whereas theaverage efficiency for the apparel industry forthe same period was about 56% (Mohd. Anuarand Faridah 2004).

Conclusion and recommendationsThis study revealed that the TP of thewatermelon sub-sector in the year 2004/05was 1.78 which was relatively higher than theTP of the crop in the year 2000 at 1.36, anincrease of about 30%. The return of 78% isconsidered favourable for an active investment.The sample farms had a mean efficiency levelof 46% which was considered as acceptablefor a high risk agriculture investment.

There are opportunities to maintain andimprove the farms productivity and efficiencyeither by reducing the production costs and/orincreasing the yield and revenue.

The farm-gate price for watermelon wasfound to be fluctuating depending on demandand supply. For the seeded variety, the pricecould be as low as 20 sen/kg. Watermelonsare high in demand during hot seasonespecially during the fasting month ofRamadan. Demand for watermelons will bevery much affected during heavy fruit seasonsand during rainy season. Thus, the farmerstogether with the assistance of the extensionagents should plan in advance the appropriatetime to produce the crop in order to get abetter price and high revenue, thus increasingtheir farms' productivity level.

Watermelon can be categorised as labourintensive and input driven sub-sector withrelatively low investment in capital assets.Labour and fertilizer constituted 37% and 30%respectively of the variable costs. Thus, theproductivity of the sub-sector could be

improved further by increasing capitalutilisation, particularly by the use of theirrigation system. The use of the irrigationsystem, in general could contribute to higherfarm productivity as the labour cost forwatering the plants especially during early cropestablishment could be reduced significantly.More than 50% of the farmers surveyed werehaving the equipment either throughgovernment subsidy or self acquisition. Thegovernment should thus extend its supportprogramme to all watermelon growers whoare in need of the facilities by providing acredit scheme to acquire the irrigation system.Priority should be given to younger farmerswith higher education level who were moreproductive based on the study. Considerationshould also be given to small scale growerswhich were more productive than the largescale growers as shown in the survey. Thisrecommendation is in line with the governmentpolicy to increase the participation of younggraduates in agricultural activities consideringthe scarcity of suitable land resource.

There is an indication of excessive usedof material inputs particularly fertilizer in thewatermelon cultivation. Fertilizer applicationat an optimum level could improve theproductivity and technical efficiency ofwatermelon at the farm level.

The average yield of watermelon(21,700 kg/ha per season) was considered lowas compared to the average yield of the cropin China, our close competitor (about33,360 kg/ha per season). From the survey,the maximum attainable yield by one of therespondent was about 46,950 kg/ha, whichwas much higher than the average yield of thecrop in China. There is a great possibility thatthe yield of watermelon in this country couldbe improved further by adopting superiorvariety and quality seeds. At this moment, thefarmers are dependent on imported seeds.There are many varieties to choose and thefarmers usually rely on recommendations fromseed suppliers. The R&D personnel shouldalso be working together with the extensionagents in identifying the most viable andprofitable variety for higher farm productivity.

26


Table 13. Measures of the technical efficiency in a sample of 49 watermelon farms

Rank no. Kopp TE Timmer TE1 1.0015 1.00132 0.9052 0.90983 0.8850 0.89064 0.7659 0.77665 0.7591 0.77006 0.6788 0.69257 0.6673 0.68158 0.6560 0.67059 0.6366 0.6516

10 0.6351 0.650211 0.5856 0.602112 0.5612 0.578313 0.5402 0.557714 0.5154 0.533415 0.5147 0.532716 0.5146 0.532617 0.5093 0.527418 0.4938 0.512219 0.4907 0.509120 0.4786 0.497221 0.4713 0.490122 0.4612 0.480123 0.4524 0.471424 0.4448 0.463825 0.4445 0.463526 0.4286 0.447927 0.4271 0.446428 0.4269 0.446229 0.3835 0.403130 0.3833 0.402831 0.3525 0.372032 0.3524 0.371933 0.3486 0.368234 0.3482 0.367735 0.3481 0.367636 0.3370 0.356537 0.3311 0.350638 0.3151 0.334539 0.3099 0.329340 0.2958 0.315141 0.2898 0.309042 0.2790 0.298143 0.2733 0.292344 0.2732 0.292245 0.2652 0.284146 0.2578 0.276647 0.2134 0.231248 0.1587 0.174649 0.1457 0.1610Mean 0.4615 0.4785Standard deviation 0.1899 0.1865

27

Raziah Mat Lin

Table 14. Actual and frontier usage of inputs in a sample of 49 watermelon farms

Rank no. Actual usage (RM) Frontier usage (RM)ID INC LABOUR CAP OTH LABOUR* CAP* OTH*

1 T39 25017 2058 81 2735 2060 81 27382 T27 9420 578 81 1705 523 73 15433 T9 314790 28180 4208 42030 24939 3724 371974 T46 8759 950 10 1746 727 7 13375 T8 192067.5 21522 2785 23112 16338 2114 17544

45 T6 86850 15650 4308 101190 4151 1142 2684346 T2 93675 19174 4433 91675 4944 1143 2363847 T40 5420 1690 269 2026 360 57 43248 T24 4348 1898 223 2108 301 35 33449 T45 3100 1336 100 2497 194 14 447

The production technology forwatermelon is available. This is indicated bythe high productivity and efficiency achievedby some of the growers. Perhaps the successfulfarmers should become models for others toemulate.

AcknowledgementThe author would like to thank the Director ofthe Economic and Technology ManagementResearch Centre, Y.M. Tengku Mohd. AriffTengku Ahmad and the Director of Rice andIndustrial Crops, Mr Ariffin Tawang forsupporting and reviewing this paper. Specialthanks were due to Mr Mohd Farid Hj.Mahmud Sabri, Mr Zulkifli Chik and theDepartment of Agriculture officials atPasir Putih, Kluang and Kota Tinggi districtsfor their assistance in the survey. The projectwas funded by the Ministry of Agricultureand Agro-Based Industry (Research GrantNo. 8977)

ReferencesAbu Kasim, A. dan Mohd. Yunus, J. (1994).

Kecekapan ladang buah-buahan di SemenanjungMalaysia kertas kerja dibentang di MesyuaratMajlis Sains MARDI ke 81, Serdang

Aigner, D.J. and Chu, F.S. (1968). On estimating theindustry production function. AmericanEconomic Review VXIII: 82639

Anon. (2002). Laporan kajian produktiviti sektorpemakanan terpilih 2000, KementerianPertanian dan Industri Asas Tani p. 12. (mimeo)

(2006). Myfruit.org. (http://www.mardi.my)Afriat, S.N. (1972). Efficiency measurement of

production functions, Int. Economic ReviewXIII: 56898.

Farrel, M.J. (1957). The measurement of productiveefficiency. Journal of Royal Statistical Society,series A (General) CXX: 25381

Forsund, F.R. and Hjalmarsson (1974). On themeasurement of productive efficiency. SwedishJournal of economics, LXXXVI: 14153

Forsund, F.R., Lovell, C.A.K. and Schmidt, P. (1980).A survey of frontier production functions andof their relationship to efficiency measurement.J. econometrics XIII: 525

Mad Nasir, S., Alias, R. and Fred, L. (1998). Trendand performance of agriculture productivity,Jurnal Produktiviti 16(1): 541

Mohd. Anuar, A.K. and Faridah, T.J. (2004). Technicalefficiency measurement of Malaysian ApparelIndustry. Paper presented at the Seminar ontotal factor productivity measurement andanalysis, 2326 Aug. 2004, Hilton Hotel,Petaling Jaya, Malaysia. Organiser: APO/NPC

Nor Aini, M.A. (2004). Efficiency and productivityat firm level, the Malaysian experience. Paperpresented at the Seminar on total factorproductivity measurement and analysis, 2326Aug. 2004, Hilton Hotel, Petaling Jaya,Malaysia. Organiser: APO/NPC

Raziah, M.L. (2003). Penilaian kecekapan firmapemprosesan produk makanan berasaskan ikandi Malaysia, 19961998. J. Trop. Agric. andFd. Sc. 31(2): 27380

28


Schmidt, P. and Lovell, C.A.K. (1979). Estimatingtechnical allocative inefficiency relative tostochastic production and cost functions.J. econometrics IX: 34366

AbstrakTemu bual bersemuka yang melibatkan 49 penanam tembikai bagi dua musimpenanaman dalam tahun 2004/2005 telah dilaksanakan dengan objektif menilaiproduktiviti dan kecekapan ladang. Analisis data menunjukkan produktiviti total(TP) bagi tiap-tiap ladang mempunyai nilai antara 0.66 dan 4.91. Produktiviti totalpurata bagi sampel ladang ialah 2.05 dan bagi sub-sektor tembikai pada 2004/2005ialah 1.78. Secara purata, sumbangan buruh kepada pendapatan lebih tinggiberbanding dengan sumbangan kapital. Prosedur ganda dua terkecil biasa (OLS)yang dibuat menunjukkan petani muda dan yang berpendapatan lebih tinggi, masing-masing lebih produktif daripada petani yang berumur dan berpendapatan rendah.Kecekapan teknikal (TE) ukuran Kopp dan Timmer yang dikira menunjukkansampel ladang beroperasi pada tahap kecekapan purata 46% dengan sisihan piawai0.18. Ruang bagi membaiki produktiviti dan kecekapan sub-sektor ini boleh dicapaidengan meningkatkan hasil dan pendapatan melalui penggunaan teknologi baru,varieti baru dan benih yang baik dan mengurangkan kos pengeluaran. Pemasangansistem pengairan di ladang tembikai sangat disyorkan bagi mengurangkan kosburuh menyiram tanaman dan juga bagi memastikan pengeluaran hasil yang baik.

total productivity and technical efficiency of watermelon atfarm level

Documents