Technical Efficiency Of Vietnam Rice Farms A Stochastic Frontier Production Approach.pdf

UNIVERSITY OF ECONOMICS HO CHI MINH CITY VIETNAM INSTITUTE OF SOCIAL STUDIES THE HAGUE THE NETHERLANDS VIETNAM NETHERLANDS PROGRAMME FOR M A IN DEVELOPMENT ECONOMICS TECHNICAL EFFICIENCY OF VIETNAM RI[.]

UNIVERSITY OF ECONOMICS HO CHI MINH CITY VIETNAM ·- INSTITUTE OF SOCIAL STUDIES THE HAGUE THE NETHERLANDS VIETNAM- NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS TECHNICAL EFFICIENCY OF VIETNAM RICE FARMS A STOCHASTIC FRONTIER PRODUCTION APPROACH By NGUYEN THANH DONG TRINH NGUYEN MASTER OF ARTS IN DEVELOPMENT ECONOMICS Ho Chi Minh, December 2011 UNIVERSITY OF ECONOMICS HO CHI MINH CITY VIETNAM INSTITUTE OF SOCIAL STUDIES THE HAGUE THE NETHERLANDS VIETNAM- NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS TECHNICAL EFFICIENCY OF VIETNAM RICE FARMS A STOCHASTIC FRONTIER PRODUCTION APPROACH A thesis submitted in partial fulfilment of the requirements for the degree of MASTER OF ARTS IN DEVELOPMENT ECONOMICS By NGUYEN THANH DONG TRINH NGUYEN Academic Supervisors DR NGUYEN TRONG HOAI DR PHAM LE THONG Ho Chi Minh, December 2011 CERTIFICATION "I certificate that the substance of the thesis has not already been submitted for any degree and is not currently submitted for any other degree I certify that to the best of my knowledge and help received in preparing the thesis and all sources used have been acknowledged in the thesis." Signature Nguyen Thanh Dong Trinh Nguyen Date: ACKNOWLEDGMENTS Firstly, I would like to say thank you to Dr Nguyen Trong Hoai and Dr Pham Le Thong - my academic supervisors, for their devoted recommendation Without their precious advice and instruction, I could not complete this thesis By the way, I am very proud to attend this program Every teacher sets an example of hard working for me and other students to follow And I will never forget the support from all employees of the program Their enthusiastic and friendly attitude makes me feel comfortable to study and research Moreover, I received the enormous and continue encouragement from my closed friends and my family, especially my mother Their loves have given me more strength and belief to overcome difficulties during the studying I am very grateful for everything that all of you gave me How can I pay your debt of gratitude! Nguyen Thanh Dong Trinh Nguyen ii ABSTRACT The research investigates the technical efficiency level and determinants of rice production in Vietnam The analysis employs the Vietnam Household Living Standard Survey 2008 data set and stochastic production frontier approach The mean technical efficiency level is 80% Credit approach, land policy, and experience are not significant elements of technical efficiency models while the irrigation, promotion program, education and gender of household head are significant ones Key words: technical efficiency, rice production, stochastic production frontier 111 TABLE OF CONTENT CERTIFICATION i ACKNOWLEDGMENTS ii ABSTRACT iii TABLE OF CONTENT iv LIST OF FIGURES vi LIST OF TABLES vi LIST OF ABBREVIATION vii CHAPTER I INTRODUCTION 1.1 Problem Statements 1.2 Research Objectives 1.3 Research Questions and Hypotheses 1.4 Research Methodology 1.6 Thesis Structure CHAPTER II LITERATURE REVIEW ON TECHNICAL EFFICIENCY AND CONCEPTUAL FRAMEWORK 2.1 Key Concepts 2.1.1 Technical Efficiency 2.1.2 Production Frontier 2.1.3 Stochastic Production Frontier I 2.2 Approaches to Measure Technical Efficiency 11 2.2.1 Data Envelopment Analysis 11 2.2.2 Stochastic Frontier Analysis 12 2.3 Stochastic Frontier Analysis Framework 12 2.3.1 Stochastic Frontier Model 12 2.3.2 Estimation method 14 2.4 Empirical Studies 15 2.5 Conceptual Framework 25 CHAPTER III RESEARCH METHODOLOGY FOR TECHNICAL EFFICIENCY AT THE FAMRS LEVEL 31 3.1 Data Source 31 3.2 Models Specification and Variables Definition 31 IV 2.1 Stochastic Frontier Production Function 31 3.2.2 Efficiency Model: 38 CHAPTER IV RESULTS AND DISCUSSION 45 4.1 Results of Data Analysis 45 4.1.1 Stochastic Frontier Production Function 45 4.1.2 Efficiency Model 49 4.2 Results Discussion 52 4.2.1 Discussion on Determinants of Stochastic Frontier Production Function 52 4.2.2 Discussion on Determinants of Technical Efficiency 54 CHAPTER V CONCLUSION AND RECOMMENDATION 60 5.1 Conclusion 60 5.2 Policy recommendation 61 5.3 Research limitation and further studies 63 REFERENCE 64 APPENDIX 69 v LIST OF FIGURES Figure 2.1: Technical Efficiency Figure 2.2: Production Frontier I Figure 2.3: The Stochastic Frontier Production Function 11 Figure 2.4: Conceptual Framework 26 Figure 4.1: Distribution of Technical Efficiency 56 LIST OF TABLES Table 2.1: Summary of Empirical Studies 22 Table 3.1 : Definition of Variables in Stochastic Frontier Production Function32 Table 3.2: Definition of Variables in Technical Efficiency Model 38 Table 4.1: Statistical Summary of Variables in Frontier Model 45 Table 4.2: Maximum Likelihood Estimation of Stochastic Frontier Production Function 46 Table 4.3: Statistical Summary ofVariables in Efficiency Model 49 Table 4.4: OLS -Robust Model of Technical Efficiency Determinants 51 Table 4.5: Statistical Summary of Technical Efficiency Level 55 Table 4.6: Distribution of Technical Efficiency 55 vi LIST OF ABBREVIATION COLS Corrected Ordinary Least Squares DEA Data Envelopment Analysis FAOSTAT Food and Agriculture Organization of the United NationsStatistics Division GDP Gross Domestic Production GOY government of Vietnam GSO General Statistics Office of Vietnam IPCC International Panel of Climate Change MLE Maximum Likelihood Estimation MOLS Modified Ordinary Least Squares OLS Ordinary Least Squares SBV State Bank of Vietnam TE Technical Efficiency USDA US Department of Agriculture VHLSS Vietnam Household Living Standard Survey VND Unit Currency of Vietnam- Vietnam Dong vii CHAPTER I INTRODUCTION 1.1 Problem Statements • Agriculture is an important sector in Vietnam economy Agriculture accounts for I8.2% of the country's gross domestic product in 2009 In 2008, agricultural export accounts for I2.3% of total export value of the country In 2009, the proportion of labor force in agriculture, forestry and fishery sector is 62.9% Rural population proportion was around 7I% (FAOSTAT, 20IO) and rural labor force made up 58.5 %of the total labor force (GSO, 20I 0) Rice is one of the most important crops in agricultural production with the highest cultivating area of food production Rice production is ranked at the fifth in the world Rice yield contributes 90% food production, and is related to 80% labor force of Vietnam In 20 I 0, domestic rice production of whole country was about 40 million tons from 7,500 thousands hectares of cultivated area with average yield of over tons/ha (GSO, 20IO) Vietnam annual per capita of rice consumption is very high, I69 kg/person/year (Laillou et al, 20 I 0), producing about I ,59 I calorie intake - 60% in 2007 (Timmer, 20 I 0) Rice has been a Vietnam's principal agricultural export and a great source of foreign exchange Value of exported rice accounts about 20% of agricultural and forestry products Vietnam has exported rice to I20 countries and its share in global market is about 20%, ranked at the second in the world (USDA, 20 I I) In 20 I 0, Vietnam has exported 6.88 million tons of rice worth US$3.23 billion In comparison with the year 2009, the quantity was increased I5.4 percent and the value was increased 21.2 percent (GSO, 20IO) - - - -~ the government should invest in research to create new higher yield rice varieties Technical inefficiency was used to analyze instead of technical efficiency Amor et al (20 I 0) applied stochastic production frontier approach in Cobb-Douglas form to estimate technical efficiency for vegetable, cereal and fruit-trees farming The cross-section data set taken from a national survey 2006 included 218 irrigated farmers in 11 regions in Tunisia The parameters of frontier production function and technical inefficiency model were simultaneously estimated by maximum likelihood method The technical inefficiency model was based on half-normal specification The coefficients of age, land property (dummy variable for owner), and traditional irrigation (dummy variable, traditional techniques have value of I) were significantly negative in the inefficiency model Education (dummy variable for analphabet farmer) had positive relation with inefficiency level Cereal and fruittrees cultivation was more efficiently than vegetables farming All inputs (farmyard manure fertilizer - tons, labor - days, mechanization - hours, water quantity - m3) except farm size affected output positively and significantly in frontier function Koc et al (2011) applied data envelopment approach to analyze technical efficiency of second crop maize in Turkey Then they used efficiency scores attained at the first to determine important elements affect on the scores by tobit regression Data of this study was collection information of the growing season 2004-2005 Technical efficiency in the input oriented method was 88%, suggesting that producers could use less 12% inputs on average but still have the same output level The age and formal education of farmer were not able to explain efficiency scores, even at 10% critical level The author also reviewed efficiency researches and concluded that education did not have strong relationship with efficiency in developing countries The number of irrigation methods was insignificant at 5% level The area was positively significant at 5% 21 The table 2.1 shows the summary of empirical studies in this literature review above Table 2.1 : Summary of Empirical Studies Author Data Variables Results Kompas (2002) farm SIZe Unbalanced Farm size, ratio of land Small panel cultivated by tractor constrained efficiency growth, ratio of land cultivated by tractor impacted on efficiency positively Hi en (2003) Crosssection Land size, variety, IPM, sowing technique, access, credit education, market access 22 Technical efficiency was positively affected by land size, variety, IPM adoption, sowing technique together with availability of credit Education and market access were negative related to efficiency Shortcomings I Advant age Analysis at province level not farm level Rios (2005) Small size, education, Crossratio of titled land, section, survey 209 credit, irrigation pipe length, coffee farmers Small farms were less efficient than large farm For small farms, education higher appears to reduce efficiency Access to credit and security of land tenure were not found to be significant factors Johansso Unbalanced Farm size n (2005) panel Small farms were most efficient and the medium farms had lowest efficiency level Only concentr ated on farm size Tijani (2006) Crosssection of Application traditional preparations, off-farm mcome, extension contact with officer, family size, age, education and experience of farmers The application of traditional preparations and offfarm income increased level of efficiency Extension contact with officer reduced efficiency Singh (2007) Crosssection Education, extension All factors had COLS is services, price of significant effects on not inputs, business attitude efficiency suitable for efficienc y analysis Idiong (2007) Crosssection Farm size, education, age, experience, size, household membership of cooperative/farmer 23 Education, communication with social organizations, credit accessibility significantly affected associations, extension efficiency level contact, credit access, sex Msuya (2008) Crosssection extension All variables affect on Education, credit, efficiency services, distance from home to land plot, gender, hand hoe, agro-chemicals Onoja (2008) Crosssection Education, experience, All variables were m extension, insignificant services technical inefficiency family size AIHassan (2008) variables extension All Education, Crosscontact, gender, age significant influenced section, on efficiency survey 732 and family size rice farms Kompas (2009) Crosssection, primary, VHLSS, province data Average farm size, proportion of land with tractor, no of plots, irrigated, education, ratio of land with use extension rights, services Khai (2009) Crosssection, Primary survey experience, Experience and area Area, of policy increased efficiency dummy recognition, Policy had negative impact on technical Am or (2010) Crosssection Primary survey of 218 farms m 11 regions Age, land property, education, dummy for traditional irrigation technique Khan (2010) Crosssection Age, experience • fragmented, Using Less size, data larger land services, sets extension land use right, greater capital for cultivated increased land efficiency Age, land ownership, traditional irrigation reduced inefficiency; education had opposite impacts education, Education reduced inefficiency High efficiency proposed research new rice 24 variety Koc (2011) Crosssection Age, education, no of All variables were DEA IS irrigation applications, insignificant at even sensitive of pesticide 10% with no application errors 2.5 Conceptual Framework This research will follow the conceptual framework described in the Figure 2.4 The study applies stochastic frontier approach, and two-step maximum likelihood estimation method to analyze technical efficiency Firstly, the stochastic frontier production function is estimated by maximum likelihood estimation method The input factors investigated in production function is showed in the first left rectangle in the figure The technical inefficiency component which is derived from the stochastic frontier production function is used to estimate technical efficiency level Secondly, technical efficiency level is used as dependent variable in technical efficiency model to examine its' determinants Technical efficiency model is estimated by ordinary least square method The second left rectangle consists of the independent variables which will be examined their effects on technical efficiency Finally, result of the analysis helps the author give policy recommendations on the issues in the last rectangle 25 Figure 2.4: Conceptual Framework Data from the Vietnam Household Living Standard Survey 2008 is used to estimate the stochastic frontier production model and the efficiency model Among them, there were 4,691 households involved in rice production Their information serves as the main data source for this technical efficiency analysis Explanatory variables in the stochastic production frontier model are information of input factors, which building up the cost of production, and farmers could control flexibly the inputs used at some extent These variables are: Working hour of family labor Amor (2010) measured labor in working days This thesis use the working hour to measure the family labor Because the data set used in the existing thesis included information about the number of months, average days per months, number of working hours per day So, it is more accurate to calculate the working hour of family labor Expenditure on hiring labor As the hired labor was very important resource in farming, Onoja (2008) separated labor to family labor and hired labor, and 26 measured both in days In Vietnam hired labor is also a considerable human resources, the hired labor should be included in the production analysis and measured by their wage in the year Land size Land size - an important component of production function was presented in almost researches of technical efficiency like Kompas (2002), Tijani ; (2006), Onoja (2008) and etc Amount of family seed Idiong (2007), Chirwa (2007), and Onoja (2008) weighed quantity of used seed in kilogram This component is very necessary in production model This analysis also weighs seed of sowing in kilogram Expenditure on buying seed There are some farmers use seed bought from outside, so the paper needs to consider this inputs factor as the family seed and examine it in the production frontier Amount of fertilizers Fertilizers were measured in kilograms in researches of Chirwa (2007), Onoja (2008), Khai (2009) The amount is sum all kinds of fertilizers used by each farm, and is measured in kilograms Expenditure on herbicides Amount of herbicides used was considered as an component in the production function in the research of Rios (2005), Khai (2009) in mililiter The data set of this paper also supplies information of expenditure on herbicides used in currency unit of Vietnam Expenditure on insecticides Pesticides was measured in currency unit (Khan, 201 0), (Hien, 2003), and in volume (Rios, 2005) The data set of this paper also supplies information of expenditure on insecticides used in currency unit of Vietnam 27 Expenditure on energy In the current research, this item includes expenditures on fuel, electricity, oil, gasoline Energy consumption was considered as one of input factors in the stochastic model of Johanson (2005) Value of capital assets Capital input in currency unit (naira) was utilized in the research of Idiong (2007) This item is sum of original value of some capital assets like tractor, water pumping machine, plough machine, plucking machine, and cattle Expenditure on hiring machines Khai (2009) did employ hired machines (in days) in the production model Hien (2003) examined the expenditure on hiring machines per hectare Kompas (2009) measured it by expenditure on it for the whole year 2004 In this current research, the author utilizes the expenditure on hiring machines available in the data set Expenditure on hiring cattle There are farmers use cattle from hiring, so the expenditure on it should be taken into the analysis The independent variables in technical efficiency model are elements maybe out of the control of farmers There were many factors found having effect on technical efficiency In this study, the author just investigates factors related to the research questions Estimating technical efficiency model will allow the author to test the statistically significant of all independent variables and main hypotheses of the study and answer the research questions a) Testing the significance of the variables: time of land use right, average size of land parcel, ratio of titled land will answer the question whether land policy affects on technical efficiency b) Testing the significance of the variables: ratio of land irrigated by machines, naturally or manual will answer the question whether irrigation manners affect technical efficiency 28 c) Testing the significance the variable of credit attainment will answer the question whether the credit affect technical efficiency Beside factors investigated to answer the research questions, the technical efficiency model should examine some common factors such as characteristics of household head and participation in agricultural promotion programs So, all variables investigated in technical efficiency model are: Ratio of titled land on the total cultivated area Rios (2005) utilized the ratio of titled land on the total cultivated area to represent for the land tenure security This research also concerns in the relationship between the land ownership and the efficiency So, the author applies the ratio of titled land on the total cultivated land in estimating the efficiency model Average of land parcel Kompas (2009) used the average land size of farm in province data to evaluate the impact of land fragmentation on technical efficiency This paper also uses the average size of land parcel for the same purpose Time of land use right In the studies of literature review, the author find there was not any research studied relation between the time of land use right and the efficiency This paper is interest in effect oftime limitation of land use right on the efficiency So it utilizes the average time that the owner had registered for their land parcels in the efficiency model Ratio of land watered manual, ratio of land watered by machines, ratio of land watered by common irrigation systems Am or (20 10) used dummy variable for traditional irrigation techniques Rios (2005) found that number of water pumpers and the length of pumpers were correlated with higher efficiency In this paper, the author calculates the ratio of land area irrigated by three different manners over the total cultivated area to assess the influence of irrigation methods on the rice farming efficiency 29 Credit attainment Hien (2003) employed amount of credit achieved by farmers Effect of accessibility to agricultural credit on technical efficiency was also investigated by dummy variable as Msuya (2008), Raphael (2008) This research employs dummy variable of accessibility to credit in technical efficiency of rice farming Participation in promotion program Researchers also investigated the effect on technical efficiency of participation in governmental programs as IPM - Integrated Pest Management (Hien, 2003), BIMAS of Indonesia (Brazdik, 2006) through dummy variables This thesis also applies an dummy variable of participation in any agricultural promotion programs to find out the statistically significant of these program in improving technical efficiency of rice farmers Characteristics of household head (gender, age, education, experience) Charactersitics of household head were interested in many technical efficiency studies as: Dummy variable of gender (Msuya, 2008); Age of the farmers (Amor, 201 0), (Idiong, 2007); Number of formal education (Onoja, 2008), (Idiong, 2007); Number of experienced years (Tijani, 2006), (Amor, 201 0), (ldiong, 2007) etc Geographic Msuya (2008) used dummy variable presenting for differences between agro-ecological and environmental region Kompas (2009) utilized dummy variable to discriminate quality of soil of province in Red River Delta and Mekong River Delta with other regions In this thesis, two dummies variables RRD and MRD are used to distinguish all natural specifics of the Northern Delta, the Southern Delta and other rice-cultivating regions in Vietnam 30 CHAPTER III RESEARCH METHODOLOGY FOR TECHNICAL EFFICIENCY AT THE FAMRS LEVEL This chapter introduces data set used in analysis, and displays the stochastic frontier production function and technical efficiency model Estimation methods applied to analyze two functions as well as variables of these functions are explained in detail 3.1 Data Source The paper uses the data from Vietnam Household Living Standard Survey 20072008 (VHLSS 2008) This research also employs the data about rice production of the whole country available online by General Statistics Office of Vietnam The VHLSS has been conducted by General Statistics Office of Vietnam for every twoyear with the financial and technical support from Word Bank This is a very large data set in term of both the number of observations and the detail information level of population and household It provides sufficient information related to the purposes ofthis research The data set includes information about 9,189 households Among them, there were 4,691 households involved in rice production Their information serves as the main data source for this technical efficiency analysis 3.2 Models Specification and Variables Definition 3.2.1 Stochastic Frontier Production Function 3.2.1.1 Model Specification and Variable Definition Applying formula (I) in the section 2.3.1, the logarithm model of the stochastic frontier function is written as follows: Ln(Yi) = Po + P1 *ln(FLAi) + P2*ln(HLAi) + P3 *ln(AREAi) + P4 *ln(FSEEDi) + Ps*ln(BSEEDi) + P6*ln(FERTi) + P7*1n(INSECTi) + P8*ln(HERBi) + P9 *ln(FUELi) 31 + P10*1n(HIRED_MACHINE;) + P1 1*ln(HIRED_CATTLE;) + P12*ln(CAPITAL;) + The variables in stochastic frontier production function are defined in the following i table: Table 3.1: Definition of Variables in Stochastic Frontier Production Function EXPECTED VARIABLE DEFINITION · UNIT SIGN Dependent Variable y Rice output Kg Independent Variable FLA Working time of family labor Hour HLA Expenditure on hiring labor AREA Cultivated area m~ FSEED Amount of family seed Kg BSEED Expenditures on buying seed FERT Amount of fertilizers 1,000 ' + + + VND Expenditure on Kg buying 1,000 + + VND insecticide Expenditure HERB + VND 1,000 INSECT + on buying 1,000 VND herbicide 32 + - - - 1,000 FUEL Expenditure on energy + VND HIRED MACHINE Expenditure on hiring machines LOOO 1,000 HIRED CATTLE + VND Expenditure on hiring cattle + VND ; 1,000 CAPITAL Value of capital assets + VND Rice output (Y) Rice output is total amount of all varieties of rice harvested by each household in the year The varieties of rice include winter-spring ordinary rice, summer-autumn ordinary rice, autumn-winter ordinary rice, glutinous rice, ordinary rice, ordinary rice on burnt-over land, special rice Family labor (FLA) Family labor is measured by total hours that all members in family working for their family The labor hours of family member is the product of the average number of working months in the year multiplied by the average number of working days in a month, multiplied by the average number of working hour per day For every household, family labor is the summation of total hours in the year of all members who take part in the production Tải FULL (86 trang): https://bit.ly/3c8PPsh Dự phòng: fb.com/TaiHo123doc.net Hired labor (HLA) Hired labor is measured by wage paid for hired labor by each family in the year Area (AREA) Area is the sum of squared meter of all land parcels belonged to household which are under rice cultivation of all varieties in the period of twelve months 33 Family seed (FSEED) Family seed measures the quantity of paddy that the farm harvested in the previous year and kept it for the sowing in the year 2008 Out-sourced seed (BSEED) Some households use seed which was bought from the market or institutions for sowing This variable measures the expenditure on buying seed Fertilizer (FERT) This variable is total amount of active ingredients as nitrogen, phosphate, potassium, NPK, and other fertilizers Insecticides (INSECT) This is the expenditure on buying the used insecticides Herbicides (HERB) This is the expenditure on buying the used herbicides Fuel (FUEL) This is the expenditure on fuel, energy, gas, lubricant, electricity and so forth Tải FULL (86 trang): https://bit.ly/3c8PPsh Dự phòng: fb.com/TaiHo123doc.net Hired machine (IDRED_MACHINE) This variable is measured by expenditure that a farm paid for renting machine used in rice farming Hired cattle (IDRED_CATTLE) This variable is measured by expenditure that a farm paid for renting oxen or buffaloes to draw plough Capital (CAPITAL) Capital is the total original value (the value at buying date) of some common means of production such as tractor, water pumping machine, plough machine, plucking machine, and cattle Because the data set does not state the capacity of cattle and machine, the analysis uses the monetary value to present for capital assets In two researches of Kompas in 2002 and 2009, all coefficients of stochastic frontier production function are positive In this research, the author also assumed that all inputs still have positive coefficients 34 3.2.1.2 Estimation Method The stochastic frontier production function in this research is estimated by maximum likelihood method Aigner et al (1977) established the following maximum log-likelihood function for the half normal distribution Ui - N+(O, cru): (4) LnL(Yia,p, :t,cr2 ) = Nln V2rr + Nlncr- + If ln[1- (EiA/a) a Where: (.) and (.) = The standard normal density function and the standard normal distribution function evaluated at (EA./cr) Point estimate of TEi with half-normal distribution (!li = 0) was defined as: 35 6669976 ... determinants of rice production in Vietnam The analysis employs the Vietnam Household Living Standard Survey 2008 data set and stochastic production frontier approach The mean technical efficiency. .. applies stochastic frontier approach to analysis technical efficiency of rice farming 2.3 Stochastic Frontier Analysis Framework 2.3.1 Stochastic Frontier Model Aigner et a! (1977), Meeusen and Broeck... the total cultivated land in estimating the efficiency model Average of land parcel Kompas (2009) used the average land size of farm in province data to evaluate the impact of land fragmentation