ArticlePDF Available

Participation Decision and Impact of Contract Farming System on Rice Farms in Myanmar

Authors:
Participation Decision and Impact of Contract Farming
System on Rice Farms in Myanmar
Aye Moe San
Lecturer, Department of Agricultural Economics, Yezin Agricultural University
E-mail: dr.ayemoesan@yau.edu.mm
ABSTRACT
Myanmar, once known as not only “rice basket of Asia” but also “the world’s largest rice exporter”
stood as 7th largest global rice producer in 2014. Successive governments attempted priority on rice
sector in any agricultural policies, thus, rice exists as a strategic sector due to wide spread utilization,
contribution to country’s GDP, creating income and employment generation in Myanmar. Rice
contract farming system was introduced by Rice Specialization Companies in Myanmar at the end of
2008, by encouraging private sector participation for rice sector development. Some companies
practice written contracts with individual farmers while others apply written or verbal contracts with
group of farmers. Under the rice contract farming system, rice farmers can get seeds, fertilizers, credit
and technical support from contracted companies as well as they have stable market access as
compared to the traditional system. This study attempts to evaluate the smallholder households'
decision to participate into the contract scheme and the contribution of contract participation on
smallholders’ rice farm performance in Myanmar, specifically in Danuphyu Township, Ayeyarwaddy
Region and Pyay Township, Bago (West) Region. It uses full information maximum likelihood
estimation of endogenous switching regression (FIML ESR) model on a total of 403 smallholders (220
contract and 183 non-contract smallholders). The empirical results show that age and education level
of household heads, frequencies of production shocks experienced during the last five years,
participation into farmer organizations, and frequently contact with extension services are influencing
the decision of smallholders to participate in the contract farming scheme. Probability of smallholders'
contract participation also differs between two study townships. Overall findings indicate that contract
farming has positive and significant impacts on gross margin of paddy production. Contract scheme
with individual smallholders along with provision of seeds, fertilizers and credit showed more effective
ways to improve smallholder livelihoods rather than group contract arrangement with only fertilizers
provision.
Keywords: Rice, contract farming, smallholders, Myanmar, Endogenous switching regression
INTRODUCTION
Myanmar was known as not only “rice basket of Asia” but also “the world’s largest rice exporter”
during 1940s. Myanmar’s paddy production stood as 7th rank among top ten global paddy producing
countries and 3rd largest rice exporter in Southeast Asia in 2014 (FAO, 2014). During 2013-2014,
paddy accounted 35% of total crop sown areas, 40% of total gross agricultural output and contributed
13% of gross domestic product (GDP) in Myanmar. Labor engagement rate is also the highest in the
rice industry as compared to other crops and approximately three-fourths of farm household income is
derived from rice farming and related activities, especially in the main rice producing areas of
Ayeyarwaddy, Bago and Sagain Regions (World Bank, 2014). Paddy production has been increasing
more than 50% up to 2013-2014, which was two and half decades after market oriented economy was
adopted, thus country has sufficient volume for growing population and considerable surplus to export.
Share of rice export value in agricultural export value and total export value were also increased after
multi parties’ government era since 2011 (CSO, 2012).
"Rice" as being the staple food as well as a source of employment opportunities and export earnings in
national economy, remains as a strategic crop for socioeconomic development of Myanmar.
Government therefore reforms and implements various agricultural policies prioritizing the rice sector
development which includes encouraging private sector participation. Poor paddy yield, usage of poor
quality seeds, mixing large number of varieties which dilutes the quality of pre-processing paddy,
limited post-harvest infrastructures, antiquated mills, high production and marketing costs, ambiguous
and arbitrary trade policy measures, and high port and export procedure costs are major bottlenecks to
smallholder dominated rice farming and rice sector development in Myanmar (Fujii and Satyanarayan,
2015).
Private rice specialization companies (RSCs) introduced the contract farming system along Myanmars
rice value chain since 2008 monsoon season especially in major rice growing areas of the country.
Contract farming system has been considered as one of the potential business models to link
smallholders to world export markets along the stable supply chain as well as an institutional solution
in the provision of inputs, finance and technical assistance to resource poor smallholders. Rice contract
farming scheme in Myanmar is still a new phenomenon and there are limitations in studies and growing
literatures. Therefore, an empirical research is essentially and statistically needed to get a better
understanding about how this system has been empowered in smallholder rice farming and rice sector
development. This study attempts to answer: "which factors are determing the probability of
smallholders' participation in rice contract farming? How does rice contract farming influence on rice
farm performance of smallholders?" in Myanmar.
Rice contract farming system with Rice Specialization Companies (RSCs)
Gold Delta RSC, Danuphyu township in Ayeyarwaddy region and Khittayar Hinthar RSC, Pyay
township in Bago (West) region are purposively selected in this study due to their actively operating
contract farming system, wide area of paddy production and large numbers of smallholders under
contracts. Basic information and detail specification of contract farming schemes by selected RSCs are
illustrated in Table 1.
Gold Delta RSC practices informal contract model, and is working seasonal written contract with
individual smallholders and providing certified seeds along with farm inputs including seasonal credit,
and product market. Gold RSC has two types of contract farmers who are seed producers and grain
producers. There is more input provision, higher purchasing paddy price and strictly control measures
in line with following good agricultural practices for contract farmers (seed producers), as compared to
provision, purchased price and controls over cultural practices for contract farmers (grain producers).
Khittayar Hinthar RSC also uses seasonal written contract with group of farmers by providing
fertilizers and output market. It practices intermediary contract model which includes formally
contracting with village head/farmer leader who informally contracts with a number of farmers.
Table 1. Rice contract farming schemes in study areas
A. Contract company
Company name
Khittayar Hinthar RSC
Gold Delta RSC
Location
Pyay township, Bago (West) region
Danuphyu township, Ayeyarwaddy
region
Established
year
2009
2009
CF target
market
Domestic, International
Domestic, International
CF business
model
Intermediary model
Informal model
CF product
Rice seed and grain
Rice seed and grain
Farmer
selection
criteria
Own or has proper paddy land
not less than 0.40 ha and not
more than 4.05 ha
Hard working, resource
endowment, trustworthy and
based on recommendation by
village leaders/farmer-to farmer
Able and willing to adopt the
guidance/techniques of
contract company in rice
farming
Loyalty to the CF scheme
Own or has proper paddy land
not less than 0.40 ha and not
more than 4.05 ha
Hard working, resource
endowment, trustworthy and
based on recommendation by
village leaders/farmer-to farmer
Able and willing to adopt the
guidance/techniques of contract
company in rice farming
Loyalty to the CF scheme
B. Contract specifications
Type of
contract
Written contract with group of
farmers
Written contract with individual
farmers
Contract
duration
7 month seasonal contract
(June to December)
8 month seasonal contract
(June to January)
Embedded
items and
services
provided
- Fertilizer (Credit in kind):
Urea 200kg for farmer who
has < 1.93 ha paddy
land
Urea 50kg and compound
fertilizer 25kg per 0.40 ha
for farmer who has
1.93<paddy land>4.05 ha
- Knowledge transfer for improving
yield and farm management
For contract farmers (grain)
- Rice varieties: Sinthwelatt and
Hmawbi 2
-Seed: 62.7 kg per 0.40 ha
-Credit: 100,000 Kyats/0.40ha
including seed cost
For contract farmers (Seed)
- Rice varieties: Sinthwelatt and
Hmawbi 2
-Seed: 31.35 kg per 0.40 ha for
Sinthwelatt variey, 20.9 kg per 0.40
ha for Hmawbi 2 variety
-Credit: 150,000 Kyats/0.40 ha
including
seed cost only
-Fertilizers for 0.40 ha (credit in
kind):
Urea: 50 kg, Tsuper: 25 kg, Potash:
12 kg
- Knowledge transfer for improving
yield and farm management
- Access to company's facilities (eg.
access to machinery facilities for
land preparation, harvesting)
Production
practices
No strict control over land
preparation and planting method.
No strict control over land
preparation and planting method for
contract grain producers.
For seed contract farmers, there is
strict control over usage of seed,
fertilizers, land preparation and
transplanting methods according to
good agricultural practices.
Product quality
criteria
Moisture content 15 % and good in
quality
Moisture content 15 %, and good in
quality
Pricing
mechanism
Current market price at delivery
time
For contract farmers (grain),
At Danuphyu 100 ton mill,
- 4,000 Kyats/46 lb (4.05
US$/46lb)
- 4,347Kyats/50 lb (4.40
US$/50lb)
- 4,521 Kyats/52 lb (4.58
US$/52lb)
At Sankin/Sakagyi villages’collection
point,
- 3,900 Kyats/46 lb (3.95
US$/46lb)
- 4,247 Kyats/50 lb (4.30
US$/50lb)
- 4,421 Kyats/52 lb (4.48
US$/52lb)
(P.S. Company will buy with current
market price if the market price at
delivery time is higher than the
contract price)
For contract farmers (seed)
At Danuphyu 100 ton mill,
- 5,500 Kyats/46 lb (5.57
US$/46lb)
Negotiation price depends on the
seed quality
Delivery of
product
Only paddy produce will be
accepted rather than cash
repayment for fertilizers, and
amount of product to be delivered
is equal to the cost of supported
fertilizer.
The deadline of product delivery is
31 December.
Minimum (50 bsk/ac) of paddy has
to sold to RSC. The value of credit in
kind and cash has to be deducted
from the value of product sold. The
deadline of product delivery is 31
January.
Regulations of
contract
Nil
Both parties have to follow all specifications of
contract. If one side breaks, the action will be
taken according to the current laws and
regulations.
Note: Exchange rate during the survey period (June to December, 2014) is 1 US$ = 987.82 Kyat according to central bank of
Myanmar.
Source: Author’s compilation based on interviews of RSCs and their related documents.
ANALYTICAL FRAMEWORK
A total of 220 contract smallholders and 183 independent smallholders from total 9 villages of two
selected townships are randomly interviewed with well-structured questionnaires focusing on
socioeconomic characteristics of households and detailed data on monsoon paddy farming activities
during 2014-2015. Participation into contracts is not only self-selection of smallholders but also non-
random selection by RSCs. Thus, participation decision could be influenced by the observed (farm and
household characteristics), and unobserved factors (motivation and management skills) of smallholders.
Full information maximum likelihood estimation of endogenous switching regression model (FIML
ESR) is used by accounting both observed and unobserved selection bias (Lokshin and Sajaia, 2004). It
calculates two separate outcome equations for contract and non-contract smallholders simultaneously
along with contract selection equation.
Contract selection: Ii =1 if Zi α + εi > 0 , Ii = 0 if Zi α + εi ≤ 0 (1)
Outcome functions: Regime 1: Y1i = β1X1i 1i if Ii =1 (2)
Regime 2: Y2i = β2X2i 2i if Ii =0 (3)
where, Ii equals 1 for contract smallholders, and 0 for independent smallholders; Y1i and Y2i are
outcomes (i.e., gross margin per hectare of monsoon paddy here) for contract and non-contract
smallholders; Zi , X1i and X2i are vectors of factors (socioeconomics and institutional characteristics); α,
β1 and β2 are the parameters to be estimated; and εi, µ1i and µ2i are the error terms. Under assumption of
trivariate normal distribution of the error terms with mean zero and covariance matrix,
σε 2 σ µ1ε σ µ2ε
Ω = cov (εi , µ1i , µ2i) = σ µ1ε σ µ12 σ µ1 µ2 (4)
σ µ2ε. Σ µ1 µ 2 σ µ22
where, Ω = variance covariance matrix to control for selection bias, σε 2 , σ µ12 and σ µ22 represent
variances of the error terms in the equations (1, 2 and 3) respectively. σ µ1ε and σ µ2ε represent the
covariance between µ1i and εi, and µ2i and εi respectively. The covariance between µ1i and µ2i, µ1 µ 2) is
unobservable as a smallholder cannot simultaneously be a contract and non-contract smallholder, thus
σ µ1 µ 2 cannot be estimated (Maddala, 1986). As the coefficient α is only estimable up to a scale factor
in the selection equation (1), the variance of σε 2 is assumed to be 1. According to Fuglie and Bosch
(1995), under given assumptions of three error terms structure, the conditional expectation of the
truncated error terms [µ1i | Ii =1] and [µ2i | Ii = 0] can be expressed as:
E [µ1i | Ii =1] = E [µ1i | εi > - Zi α ] = σ µ1ε 󰇛󰇜
󰇛󰇜 = σ µ1ε i1 (5)
E [µ2i | Ii = 0] = E [µ12i | εi - Zi α ] = σ µ2ε 󰇛󰇜
󰇛󰇜 = σ µ2ε i2 (6)
where , i1 and i2 are the Inverse Mills Ratios (IMR) computed from the selection equation (1), and
󰇛󰇜 and 󰇛󰇜 are the probability density function and cumulative distributive function of standard
normal distribution, respectively. Following Maddala (1986), a probit model can be applied to generate
i1 and i2 from selection equation (1), which can be treated as missing variables in equation (2) and (3).
In the second stage, the selection bias terms i1 and i2 are added to the outcome equations which can
then be consistently estimated by OLS, to correct for selection bias. This can be estimated using a two-
stage estimation procedure of endogenous switching regression. Therefore, equations (2) and (3) could
be specified as:
Y1i = β1X1i + σ µ1ε i1 + ζ 1i if Ii = 1 (7)
Y2i = β2X2i + σ µ2ε i2 + ζ2i if Ii = 0
(8)
where, Y1i , Y2i, β1, and β2 are as earlier defined, i1 and i2 control for bias associated with sample
selection problem, especially when smallholder within and outside contract may be different from
average smallholder with characteristics Xi and Z due to unobserved factors. The coefficients of i1 and
i2 are estimates of the covariance terms σ µ1ε and σ µ2ε, respectively. If these coefficients are non-zero
µ1ε ≠ σ µ2ε 0), there are unobserved factors associated with selection bias, it is likely the correlation
between the error terms of outcome equations and the selection equation giving the case of endogenous
switching. In this case, when σ µ1ε = σ µ2ε = 0, there exist exogenous switching regression. Typically, this
is tested by ρ1, the correlation coefficient between εi and µ1i and and ρ2, the correlation coefficient
between εi and µ2i:
ρ1= 
 and ρ2 = 
 respectively.
The residuals ζ 1i and ζ 2i in the above equations (7) and (8) cannot be used to determine the variances
of the second-stage estimates (Fuglie and Bosch, 1995). Models with self-selection or endogenous
switching can be estimated one equation at once either by two-step least square or maximum likelihood
estimation. However, both of these estimation methods are inefficient and theses approaches require the
potentially cumbersome adjustments to derive the consistent standard errors. Lokshin and Sajaia (2004)
introduced the “movestay” command in the statistical software STATA, which implements the full
information maximum likelihood (FIML) to simultaneously estimate binary selection and continuous
outcome parts of the model in order to yield the consistent standard errors. Therefore, this study applies
FIML endogenous switching regression model which estimates the selection and outcome equations
simultaneously and to yield the consistent standard errors. The model gives the test for joint
independence of the two equations, if σ µ1ε and σ µ2ε in equations (7) and (8) show non-zero and
statistically significant, there is endogenous switching, otherwise, there is exogenous switching. In
addition, a better identification requires an exclusion restriction because the FIML ESR model is
identified through the non-linearities of i1 and i2. Verbeek (2012) said that the Z variables in selection
equation (1) are expected to contain at least one variable not in X variables of outcome equations (2)
and (3). This variable directly affects smallholders’ participation decisions but does not directly affect
outcome interest (gross margin) of the smallholder households. Access to extension services is here
considered as the instrumental variable for identification of FIML ESR model and a simple falsification
test is applied to valid the selection instrument.
Given the trivariate normal distributions of error terms, FIML ESR model is written as follow:
󰇝
 󰇟󰇛󰇛󰇜󰇛󰇛
󰇜󰇜 󰇛󰇜󰇟 󰇛󰇜
󰇛󰇛
󰇜󰇜󰇠󰇞 (9)
where, = an optional weight for smallholders i (i=1, 2, 3,…,n),
and = the probability density and cumulative distributive functions of standard
normal distribution,
 󰇛󰇜
 (j= CF, NCF)
where ρj represent the correlation coefficients between εi and µ1i CF) and, between εi and µ2i NCF),
respectively. To ensure that the estimated ρCF and ρNCF are bounded between -1 and 1 and estimated 
and  are always positive, the maximum likelihood directly estimates lnln and atanh ρj:
where 


The estimated correlation coefficients, ρCF and ρNCF of ESR model provide the interesting insights of
the sample smallholders in choosing the contract scheme. When ρCF>0, implies “positive selection”
into choosing contract, smallholders that actually chose contract scheme, have above average gross
margin under contract. If non-contract smallholders have, in fact, chosen to join the contract, their
performance would be worse than actual contract smallholders. If ρCF<0, “negative selection” into
choosing the contract, or actual contract smallholders have below average performance under contract.
In this case, if the non-contract smallholders have, in fact, chosen to join the contract, their
performance would be above that of the contract smallholders. Conversely, ρNCF>0 implies “negative
selection” into not choosing the contract for non-contract smallholders. In other words, the non-contract
smallholders have below average performance, and if the contract smallholders have, in fact, chosen
not to contract, their performance would be above that of the non-contract smallholders. If ρNCF<0,
“positive selection” into not choosing the contract for non-contract smallholders, or smallholders who
actually choose not to enter the contract have above average performance. In the case, if the contract
smallholders have, in fact, chosen to not to engage the contract, their performance would be worse than
that of the non-contract smallholders. After parameters are estimated in ESR model, following Di Falco
et al. (2011), the conditional expectations, average treatment effects and heterogeneity effects of gross
margin of monsoon paddy in the observed and hypothetical scenarios are calculated as presented in
Table 2:
The conditional expectations of outcome interests of:
Contract smallholders with contract (observed in the sample),
YC1-1i = E (Y1i | Ii = 1, X1i) = X 1i β1+ σ µ1ε ρ1 i1 (10a)
Non-contract smallholders without contract (observed in the sample),
YC2-2i = E (Y2i | Ii = 0, X2i) = X2i β2 + σ µ2ε ρ2 i2 (10b)
Contract smallholders without contract (counterfactual),
YC2-1i = E (Y2i | Ii = 1, X1i) = X 1i β2+ σ µ2ε ρ2 i 1
(10c)
Non-contract smallholders with contract (counterfactual),
YC1-2i = E (Y1i | Ii = 0, X2i) = X 2i β1 + σ µ1ε ρ1i2 (10d)
Table 2. Treatment and heterogeneity effects
Sample
Decision stage
Treatment
effect
Not to participate
Contract smallholders
(10c)
ATT=(10a)-(10c)
Non-contract
smallholders
(10b)
ATU=(10d)-(10b)
Heterogeneity effects
BH2
TH
Note: (10a) and (10b) represent observed expected gross margin/hectare of monsoon paddy; (10c) and (10d) represent
counterfactual expected gross margin/hectare of monsoon paddy of contract and non-contract smallholders;Ii = 1 if smallholder
household participates into contract scheme, Ii = 0 if smallholder household did not participate into contract scheme; Y1i = gross
margin/hectare if the smallholder households participated, Y2i = gross margin/hectare if the smallholder households did not
participate; ATT = the average treatment effects of the treatment (i.e., participation into contract) on the treated (i.e., contract
smallholders), ATU = the average treatment effects of the treatment (i.e., participation into contract) on the untreated (i.e., non-
contract smallholders); BHi = the effects of base heterogeneity for contract smallholders (i= 1) and non-contract smallholders
(i=2); TH = (ATT –ATU) = the transitional heterogeneity.
RESULTS AND DISCUSSIONS
Household characteristics
Selected household characteristics of sample smallholders are characterized after categorizing them
into contract and non-contract groups, as in Table 3. In both regions, sample contract smallholder
household heads are younger and less rice farming experience in comparison with non-contract
smallholder heads. The sample contract smallholder household heads have above secondary level
education while non-contract smallholders attended up to secondary level education. Majority of
sample households are male headed households. Average family sizes of sample households in Pyay
were lower as compared to those in Danuphyu. Almost half of family members of sample households
in both regions contribute in agricultural activities. The average farm sizes of sample farmers in Pyay
and Danuphyu are nearly equivalent with the country’s average land holding size which is 2.65 ha per
rural household.
Table 3. Household characteristics of sample smallholder households
Item
Pyay
Danuphyu
Contract
Non-
contract
Contract
Non-
contract
Age of HH head (year)
44.70
58.47
46.21
57.08
Rice farming experiences (year)
23.41
37.68
20.69
37.30
Education of HH head (year)
8.10
6.07
9.96
5.91
Gender (%)
Male
83.51
88.17
98.37
100.00
Female
16.49
11.83
1.63
0.00
Family size (No.)
4.03
4.06
4.76
4.60
Agri-labor (No.)
1.70
1.81
1.76
1.81
Farm size (ha)
2.65
2.40
2.97
2.61
Farm characteristics
As shown in Table 4, contract smallholders have more cultivated area, yield per unit area and price of
monsoon paddy in comparison with non-contract smallholders in both regions. Urea and compound
fertilizers are used more by sample smallholders in Pyay while T super and potash fertilizers are used
more by sample smallholders in Danuphyu as compared to their counterparts. Seed rate used by sample
contract smallholders from Danuphyu look less than other smallholders. Family labors are not enough
and sample smallholder households hire labors for their monsoon paddy production. Gross return of
sample contract smallholders are higher than that of non-contract smallholders because of higher paddy
yield per hectare and price received although they incur more total production costs. Therefore, sample
contract smallholders get more profits in monsoon paddy production as compared to non-contract
smallholders, as illustrated in Table 5.
Table 4. Farm characteristics of sample smallholder households for monsoon paddy
production
Item
Pyay
Danuphyu
Contract
Non-
contract
Contract
Non-
contract
Cultivated area of monsoon paddy
(ha)
2.63
2.40
2.91
2.60
Gross paddy yield (ton/ha)
3.93
3.61
4.09
3.48
Effective paddy yield (ton/ha)
3.41
3.15
3.72
2.84
Marketing charges (‘000Ks/ton)
4.95
1.77
4.34
0.17
Effective paddy price (‘000Ks/ton)
208.97
201.86
212.37
203.11
Seed rate (Kg/ha)
147.68
133.74
116.88
136.51
Urea (Kg/ha)
104.65
95.34
111.25
87.95
T super (Kg/ha)
3.81
5.97
51.70
49.90
Potash (Kg/ha)
-
-
7.78
1.37
Compound (Kg/ha)
87.21
70.11
29.36
43.33
Organic manure (Cart/ha)
7.41
7.41
7.41
7.41
Pesticide and Herbicide (Litter/ha)
1.72
1.68
1.44
1.81
Fuel (Gallon/ha)
7.41
7.41
7.41
7.41
Family labor (md/ha)
28.35
31.69
35.80
36.10
Hired labor (md/ha)
81.03
84.98
68.07
57.99
Table 5. Cost and return of monsoon paddy production
Benefit and Cost (‘000Ks/ha)
Pyay
Danuphyu
Contract
Non-
contract
Contract
Non-
contract
Gross Benefit
712.27
636.85
790.89
577.56
Total material input cost (Cash)
132.09
110.27
143.22
123.03
Total material input cost (Opportunity)
63.70
71.34
52.11
66.73
Total family labor cost
53.78
57.92
69.38
65.75
Total hired labor cost
156.91
155.03
165.88
137.81
Total interest on cash cost
1.30
1.19
1.39
1.17
Total variable cost
407.78
395.75
431.98
394.50
Total variable cash cost
290.30
266.49
310.49
262.02
Gross Margin or Profit
304.49
241.10
358.91
183.06
Benefit-Cost ratio
1.75
1.61
1.83
1.46
Description of variables used in endogenous switching regression model
As presented in Table 6, the mean difference of gross margin of monsoon paddy of contract and non-
contract smallholders is about 119,913 Ks/hectare and significant at 1% significance level. It implies
contract smallholders earn more profit than their non-contract counterparts. Contract smallholder
household heads are younger, more educated and richer in comparison with non-contract household
heads, which are all significant at 1% level of significance. Contract smallholders have significantly
larger farm size but their family members who worked in agricultural activities are lesser as compared
to non-contract smallholders. Paddy price is also significantly indicating that contract smallholders gets
12,020 Ks/kg more than non-contract smallholder counterparts. There are no significant differences in
cost of seed, pesticides, herbicides and fuel between two groups of smallholders, however, significant
higher fertilizer cost is spent by contract smallholders than non-contract ones. Significant more hired
labor cost is paid by contract smallholders while non-contract smallholders pay significant higher
family labor cost than their counterparts. During the past half of the decade of monsoon paddy
production, climatic attacks were significantly experienced by non-contract smallholders more than
contract smallholders. Both smallholder groups faced more or less similar production shocks such as
yield loss, pest and disease damages and low quality of paddy during the last 5 years of monsoon paddy
production. There is no significant difference between numbers of non-farm income jobs between two
groups of smallholders. Access to extension services and participation in local farmer based
organization by contract smallholders are significantly higher in comparison with non-contract group.
Although there is no significant difference for market distance between two farmer groups, contract
smallholders are located close to output markets compared to non-contract smallholders, indicating that
more opportunity to get contact with RSCs for initial adoption.
Table 6. Description of variables used in endogenous switching regression model
Variables
Description
Mean (SD)
t
statisti
cs
Contract
(N=220)
Non-
contract
(N=183)
Selection (treatment) variable
Contract farming 1= if household participates in contract farming scheme in 2014
monsoon paddy production season, 0= otherwise
Outcome variables:
Gross margin
Gross margin per hectare of
monsoon paddy (Ks/ha)
393,389
(115,945)
273,476
(103,614)
10.84*
**
Explanatory variables
Age
Age of household head (year)
45.55(9.10)
57.79 (6.94)
15.31*
**
Gender
1= if HH head is male
0= if HH head is female
0.92 (0.27)
0.94(0.24)
0.85
Education
Completed schooling years of HH
head (year)
9.14 (3.01)
6.00 (2.14)
12.19*
**
Family size
Total family member in HH (No.)
4.44 (0.10)
4.33 (0.10)
0.78
Agri-labor
Share of agricultural labor in HH
(%)
41.03(15.28)
43.89
(15.67)
1.84*
Farm size
Total land holding size of HH (ha)
2.84 (0.07)
2.50 (0.06)
3.52***
Paddy price
Sold price of monsoon paddy
(Ks/Kg)
215.48(21.0
3)
203.46
(19.41)
5.96***
Asset value
Value of all assets by HH (000Ks)
15,878
(8,441)
12,312
(4,715)
5.09***
Seed cost
Cost of paddy seeds (000Ks/ha)
40.59
(12.59)
41.06 (7.74)
0.46
Fertilizer cost
Cost of fertilizers (000Ks/ha)
73.75
(20.36)
63.25
(17.79)
5.52***
Pesti/Herbi &
Fuel cost
Cost of pesticide, herbicide and
fuel (000Ks/ha)
47.85
(13.83)
48.32 (1.59)
0.36
Family labor
cost
Costs of total family labors
(000Ks/ha)
66.15(22.68)
72.07
(24.19)
2.51**
Hired labor
cost
Costs of total hired labors
(000Ks/ha)
166.95
(48.62)
150.12
(44.49)
3.63***
Demo shock
HH experienced from
illness/dead/ birth in the past 5
years (No.)
0.75 (0.88)
0.76 (0.89)
0.06
Climate
shock
HH experienced from
drought/flood in monsoon paddy
production in the past 5 years
(No.)
0.92 (1.07)
1.11 (1.03)
1.83*
Production
shock
HH experienced yield loss, pest
and disease damage, low product
quality in the past 5 years (No.)
1.80 (0.85)
1.78 (0.75)
0.18
Nonfarm
source
Nonfarm income activities of
HH(No.)
0.60 (0.58)
0.56 (0.55)
0.84
Extension
1= if HH got extension services,
0= otherwise
0.99 (0.10)
0.82 (0.39)
5.86***
Farm
organization
1= if HH participate in any local
farmer based organization,
0= otherwise
0.78 (0.42)
0.03 (0.18)
23.97*
**
Market
distance
Distance from farm to selling
points (miles)
8.68 (4.95)
9.33 (5.08)
1.29
Region
1= Pyay, 0= Danuphyu
0.44 (0.50)
0.51 (0.50)
1.35
Participation into contract system and its impacts on monsoon paddy profit per hectare
The estimation results to assess the impact of contract participation on gross margin (profit) of
monsoon paddy per hectare, as in Table 7, provide the driving forces that motivate the smallholder
farm households to participate into contract system and the estimates of outcome regression equations
for contract and non-contract smallholders. The Wald χ2 test statistics indicates that the selected
covariates provide good estimates determinants to apply the model and they are jointly and statistically
significant.
Determinants of participation into contract farming system
As shown in the first column of Table 7, the dependent variable for selection equation is binary,
showing the value 1 if the smallholder households engage with Rice Specialization Companies (RSCs)
under contract in 2014 monsoon paddy production activities and value 0 (zero) otherwise. The results
show that the decision to participate into contract system is influenced by a number of farm and
household characteristics, different shock experiences in paddy production, and other supported
services. Sample smallholder households headed by younger and more educated household heads, who
received higher market price for monsoon paddy and spent higher costs for fertilizers, participated in
local farmer based organizations, got more frequently contact with extension agents, faced less
production shocks during last five years in monsoon paddy production have more probability to work
together with RSCs under contract system. Smallholders who lived in Pyay Township also have high
probability to join contract system as compared to those lived in Danuphyu Township.
The results show that the younger and the more educated the smallholder household heads, the higher
the probability of contract participation. This implies that the older household heads tend to be risk
adverse and might avoid the risks by the introduction of contract farming arrangements. Education
level of household head is another important determinant of participation probability into contract as
shown by significant positive effect. That is consistent with the conventional economic theory on the
role of literacy in improving conceptualization of information and making economically viable
decisions in financial markets. These findings are corroborated by empirical results from earlier
contract farming studies by Chang et al., 2006; Cai et al., 2008; Musara et al., 2011; and Mercy et al.,
2013. The region dummy variable expresses correlation with contract participation, likely reflecting
unobservable spatial and ecological differences. Smallholder households who live in Pyay township are
more likely to join contract farming arrangement as compared to smallholders from Danuphyu
township.
The economic incentive such as product price could have a positive significant effect on participation
decision. The policy implication by introducing contract farming system in rice sector is the facilitating
and disseminating of market information, including guarantee product price and assured market, might
facilitate interest to participate in contract arrangement. This result is verified by the findings of rice
contract farming in Cambodia by Cai et al. (2008) and in Thailand by Sriboonchitta and
Wiboonpoongse (2008). On the other hand, the fertilizer costs are also positively associated with
contract participation decision of smallholder households. Smallholders, who are spending more
expenses on chemical fertilizers in paddy production, would like to join with RSCs under contract
because they could achieve supports of fertilizers as a credit in kind. Similar finding revealing positive
correlation between fertilizer cost and adoption of System of Rice Intensification in Timor Leste was
also confirmed by Noltze et al. (2013).
Production shocks such as high price of certified seeds, fertilizers, scarcity or high wage for hired
labors, insufficient credit, pest and disease attacks and poor product quality which are experienced
during last five years of monsoon paddy production are significant and negatively associate with the
contract participation probability of smallholder households. Smallholders who faced high production
shocks in past five years might have the unsecured conditions to meet the contract terms and
agreements such as minimum product quantity to be sold to RSCs and the product quality standard of
RSCs.
The participation into local farmer based organizations at village tract level and having frequently and
regularly contacts with extension services are positive and significantly associate on the contract
participation of smallholder households. The main function of local farmer based organizations is to
provide season loan for monsoon paddy production as well as sharing agricultural production and
market information among each other. The collective action might enable smallholders to attain better
bargaining power, economies of scale and reduce transaction costs. Therefore, participation into such
farmer based organizations could provide information about contract farming, which supports to high
probability of contract participation of smallholders. Similar finding confirmed by Sharma (2008) is
that membership in farmers group/association/cooperatives significantly determined participation in
contract farming. Farm households who come to know contract farming system via contact with
extension staffs from RSCs were more likely to make contract with RSCs, as compared to farm
households who are informed by other agricultural information dissemination pathways. These results
also validate with recent studies that revealed access to extension services is one of the most important
factors to enhance the technical capability for better yield and increased the likelihood of smallholders’
contract participation (Adam Kephas, 2011; Ogeto et al., 2012; Moyo, 2014; Abdulai, 2016; Azumah et
al., 2016).
Table 7. FIML ESR estimates: Contract participation decision and functional relationship
between smallholders’ characteristics and monsoon paddy profit
Explanatory variable
Participation
decision
Gross margin (profit) per hectare (ln)
Contract
smallholders
Non-contract
smallholders
Age (year)
-0.09*** (0.02)
0.01** (0.00)
0.001 (0.01)
Gender
0.20 (0.44)
0.28** (0.10)
0.22 (0.25)
Education (year)
0.16** (0.05)
0.03** (0.01)
-0.001 (0.03)
Family size (No.)
-0.17 (0.11)
-0.11*** (0.02)
-0.07 (0.05)
Agril-labor (%)
-0.01 (0.01)
-0.001 (0.00)
-0.01 (0.00)
Farm size (ha)
0.07 (0.21)
0.12** (0.04)
0.10 (0.13)
Asset value (ln)
0.54 (0.50)
0.11 (0.08)
0.04 (0.28)
Paddy price (Ks/kg)
0.01** (0.01)
0.01*** (0.00)
0.01** (0.00)
Seed cost (ln)
0.08 (0.51)
-0.34*** (0.08)
-0.25 (0.29)
Fertilizer cost (ln)
0.69** (0.36)
-0.39*** (0.08)
-0.63** (0.20)
Pesti/HerbiFuelcost (ln)
-0.63 (0.51)
-0.13 (0.10)
-0.69** (0.25)
Family labor cost (ln)
0.19 (0.23)
0.05 (0.05)
-0.27** (0.14)
Hired labor cost (ln)
0.51 (0.40)
-0.29** (0.09)
-0.36* (0.20)
Demo shock (No.)
-0.10 (0.13)
-0.05* (0.03)
-0.06 (0.07)
Climate shock (No.)
0.04 (0.14)
-0.01 (0.03)
-0.03 (0.06)
Production shock (No.)
-0.40** (0.17)
-0.01 (0.03)
-0.01(0.09)
Nonfarm source (No.)
0.28 (0.26)
0.04 (0.05)
-0.23** (0.12)
Farm organization
2.54*** (0.32)
0.04 (0.08)
0.47 (0.69)
Market distance (Mile)
0.00 (0.04)
0.01 (0.01)
0.01 (0.02)
Region
0.62* (0.34)
0.08 (0.09)
0.26 (0.16)
Extension
2.03** (0.65)
Constant
-10.03* (5.87)
5.89***(1.09)
6.66** (3.12)
ln δCF, ln δNCF
-1.04***(0.05)
-0.31*** (0.05)
ρCF , ρNCF
0.01**(0.19)
0.04 (0.48)
Number of observations
403
Wald χ2
235.08***
Log pseudo-likelihood
-361.46
Likelihood ratio test for independent equations χ2
7.38**
Note: *, **, and *** denotes significance at 10, 5, and 1% levels. Values in parentheses represent robust standard errors.
Factors influencing on gross margin (profit) per hectare of monsoon paddy
The estimates of two outcome equations for contract and non-contract smallholder households, as in 2nd
and 3rd columns of Table 7, show that product price, fertilizer and hired labor costs significantly affect
gross margins of both contract and non-contract smallholders. An increase in cost of fertilizer and hired
labor creates a decline in gross margin per hectare while high product price contributes positively to the
profit per hectare. Monsoon paddy profit of non-contract smallholders are more elastic in response of
fertilizer and hired labor cost as compared to contract smallholders. Paddy production in Myanmar is
still labor intensive farming especially in transplanting, weeding and harvesting period. Paddy yield and
quality could be affected at the end if these production activities are delayed because of unavailability
or insufficient usage of labor (ACDGI, 2011). Therefore, smallholder households need to allocate more
labor even the hired labor wage rate might be expensive at the peak labor requirement periods in order
to protect losses in yield and quality damage.
There are differences in some coefficient estimates which determined the gross margin of monsoon
paddy among contract and non-contract smallholder households. These notable differences confirmed
that the switching regression framework is more appropriate than data pooling in one regression. Age,
gender, education level of household heads, family size, farm size, seed costs and demo shocks faced
during last five years are significantly associated with the gross margin of monsoon paddy of contract
smallholders; however, the effects are insignificant among the non-contract smallholders. The facts that
less family size, larger farm size, and less seed costs, less frequency experienced in demo shocks
suggest that the contract smallholder households might be implementing input saving activities.
Positive relationship between farm size and profit shows large farm has higher profit per hectare. The
contract smallholders try to efficiently use the amount of seeds per hectare because the seed prices of
recommended rice varieties by RSCs are relatively expensive, and 1 % increase in seed cost would
significantly cause 34 % decline in gross margin.
A case in point is the costs for pesticide, herbicide and fuel, family labor costs and the number of non-
farm income sources are significantly influenced the profit of monsoon paddy of non-contract
smallholders, while these factors are insignificant in the contract regime. Regular weeding and
protection of pests/diseases in monsoon paddy production are recommended by Department of
Agriculture (DOA) in order to achieve high yield, but not always follow by smallholders especially
independent smallholders. In addition, they used to apply only family labor for land preparation and
threshing process. Chemical weed control and high use of family labor in their paddy farming
significantly affect monsoon paddy profit of non-contract smallholders. At the other side, an additional
source of non-farm activities could lead to reduce time for paddy farming and negatively associates
with monsoon paddy profit of non-contract smallholders.
Impact of contract farming on gross margin (profit) per hectare of monsoon paddy
The significance of likelihood ratio test criterion for independence of equations, as in lower part of
Table 7, shows that there is joint dependence between contract selection equation and outcome
functions for contract and non-contract smallholders. The sign and significance of ρCF and ρNCF for
contract and non-contract smallholders report the presence of selection bias. ρCF >0 and ρNCF >0 show
that contract smallholders would have higher profit whether they decide to participate into contract or
not. The positive and significant ρCF indicates “positive selection” bias and shows that there is self-
selection among contract smallholders, and both observed and unobserved factors affect the contract
participation decision and gross margin. In particular, the positive sign of ρCF indicates that the contract
smallholders would have above average gross margin per hectare whether they chose to join the
contract or not. They have an “absolute advantage’ or they could have better farms in general. On the
other side, ρNCF >0 shows that non-contract smallholders have lower average gross margin whether they
decide to participate into contract or not. Thus, it could say that their paddy production is not as good as
that of contract smallholders. The findings of positive selection bias is similar with many other impact
evaluation studies, which indicated that more productive or progressive farm households were usually
the first to try new technologies or new farming systems (e.g; Fuglie and Bosch, 1995; Abdulai, 2016,
Haile, 2015, Tambo, 2015).
The expected gross margins per hectare of smallholders under actual (aand b) and counterfactual
simulations (c and d) are presented in Table 8. Under observed conditions, the expected monsoon
paddy profits per hectare by contract smallholder households are higher than that of non-contract
smallholder households for each and both regions; and these differences indicate big gap. Therefore,
the simple comparison between two smallholder groups can be misleading and drive the researcher to
conclude that on average contract smallholders earned more than non-contract smallholders.
The average treatment effect on treated (ATT) measures the difference between the actual average
gross margin of contract smallholders and the counterfactual one what they would have earned if they
do not participate into contract. The average treatment effect on untreated (ATU) indicates the
difference between actual average gross margin of non-contract smallholders and the counterfactual
one what they would have earned if they do participate into contract. The last column of table 8
presents the percentage changes of average treatment effects on smallholder in order to provide the
realistic interpretations of the treatment effects.
Table 8. Average expected monsoon paddy profit per hectare, treatment and heterogeneity
effects of contract systems
Item
Decision stage
Treatment effect
Effec
t1 in
%
To participate
Not to
participate
Pyay township
CF smallholders
(a) 5.58 (0.03)
(c) 5.27 (0.05)
ATT =
0.31***(0.06)
36.3
4
NCF
smallholders
(d) 5.50 (0.04)
(b) 5.01 (0.04)
ATU =
0.49***(0.05)
63.2
3
Heterogeneity
effects
BH1=0.08 (0.05)
BH2=0.26***(0.06
)
TH = -
0.18***(0.05)
Danuphyu township
CF smallholders
(a) 5.65 (0.04)
(c) 5.15 (0.04)
ATT = 0.50
***(0.05)
64.8
7
NCF
smallholders
(d) 5.39 (0.03)
(b) 4.79 (0.05)
ATU = 0.61**(0.06)
84.0
4
Heterogeneity
effects
BH1=0.25***(0.05
)
BH2=0.38***(0.06
)
TH = -0.11**(0.05)
Total (Pooled sample from both twonships)
CF smallholders
(a) 5.62 (0.02)
(c) 5.20 (0.03)
ATT =
0.42***(0.04)
52.2
0
NCF
smallholders
(d) 5.45 (0.03)
(b) 4.90 (0.03)
ATU =
0.55***(0.04)
73.3
3
Heterogeneity
effects
BH1=0.17***(0.04
)
BH2=0.30***(0.04
)
TH = -
0.13***(0.04)
Note: *, **, and *** denotes significance at 10, 5, and 1% levels. Values in parentheses represent robust standard errors.
1
The
percentage changes in average treatment effect was derived based on 100 (eTE -1), where “e” is the exponential “e” (e = 2.718)
and “TE” is the average treatment effects provided by the analysis of the log-transformed variable.
As ATT and ATU show positive and statistically significant results, indicating that working together
with Rice Specialization Companies (RSCs) under contract arrangements provide significant positive
impact on gross margin per hectare of smallholders in both regions. The results reveal that contract
farming arrangements significantly increase gross margin per hectare 52.20% for contract smallholders
and 73.33% for independent smallholders if they would join into contract system. Contract
smallholders in Pyay and Danuphyu would achieve 36.34% and 64.87% less profit per hectare
respectively if they do not participate in contract scheme. The impact of contract farming scheme is
more important for non-contact smallholders, where they would have achieved 63.23% and 84.04%
more profit for Pyay and Danuphyu townships, respectively if these independent smallholders would
join to RSCs through contract system. These results reveal that participation in contract system
significantly increase gross margin per hectare for all smallholders; in particular, the effects are more
important for non-contract smallholders because they would have benefited more by contract
participation. The contract farming system in Danuphyu township is more likely to be effective than
that in Pyay township.
The base heterogeneity effects of non-contract smallholders (BH2) show that the contract smallholders
would still have statistically significant higher gross margin even without contract participation (c) than
the observed gross margin of non-contract smallholders (b), in each townships and both regions. The
positive and significant BH2 indicate that there are some important sources of heterogeneity that
position contact smallholders to generate more profit, which also likely influence the contract
participation. These sources of heterogeneity are also important in comparing (a) and (d) in order to
identify the base heterogeneity effects of contract smallholders (BH1). Due to not significant (BH1)
value, non-contract smallholders in Pyay township could earn more or less similar gross margin as
contract smallholders with actual contract condition, if they would join RSCs through contracts. (BH1)
values in Danuphyu township show positive and statistically significant base heterogeneity effects.
However, non-contract smallholders in Danuphyu township, although the non-contract smallholders
would have benefited when they decide to participate into contract arrangements, they could not
generate significant equal gross margin like as the actual contract smallholders. The transition
heterogeneity effects (TH) in each township show negative and statistically significances, indicating the
effects of contract farming on gross margin are significantly lower for contract smallholders as
compared to that of non-contract smallholders, though working together with RSCs via contracts could
increase all smallholders’ profits.
CONCLUSIONS AND POLICY RECOMMENDATIONS
This study contributes for the essential requirement of initial assessment study of rice contract farming
system at farm level as well as for growing literature on the determinants of participation in contract
farming system and its potential outcome effects on economic condition of smallholder farm
households in Myanmar. Household characteristics (age and education level of household head), farm
characteristics (paddy price, fertilizer cost, production shocks experienced during five years ago in
paddy production), institutional characteristics (farmer organizations and extension access) have
positive relations on the probability of contract participation. Smallholder households who live in Pyay
township have more likelihood to work together with RSC via contract system compared to
smallholders of Danuphyu township.
The results show that the participations in contract farming and gross margin per hectare of monsoon
paddy of sample smallholders are jointly depending. Smallholder households with better gross margin
above the average level of sample households, regardless of joining to contracts, are more likely to
work together with RSCs. There are some important factors that skill the actual contract smallholders
have better conditions than the non-contract smallholders even if they are without contract system and
these important factors could have also influenced on the decision of smallholders to participate in
contract farming system. There are obviously differences in some characteristics of smallholders which
are determining their paddy profit of contract and non-contract smallholder groups. Such differences
make smallholders into different likelihood of participation into contract and bear different extents of
participation impacts on economic livelihoods of smallholders. Participation in contract arrangements
shows significant positive impacts to enhance smallholders' gross margin (profit) from growing
monsoon paddy. About 73% and 52% increase monsoon paddy profits for non-contract and contract
smallholders respectively would be achieved under contract schemes. Smallholders who work with
Gold Delta RSC through individual contracts achieve higher positive impacts on gross margin in
comparison with those who work with Khittayar Hinthar RSC through contracts with group of
smallholders.
Overall results imply that the contract farming system would have broader impacts on economic
conditions of smallholders if appropriate strategies could be improved in reducing production shocks in
monsoon paddy production and in supporting sufficient inputs especially chemical fertilizers with
reasonable prices, community relationship with farmer based organizations and very often access to
extension services at farm household level. These findings would suggest policy makers, authorized
executives from Myanmar Rice Federation (MRF) and RSCs in the promotion of rice contract farming
system in rice value chain of Myanmar. The determining factors that influence the contract
participation decision of smallholders (eg; product price and fertilizer cost, access to extension
services,…) could be considered as the entry points to reorganize the existing contract arrangements as
well as to promote the contract system. Informal contract farming model along with supporting farm
inputs including credit shows more effective way as compared to intermediary contract model with
only fertilizer provision. Comparison between two RSCs in two regions indicates that the contract
system of Gold Delta RSC in Danuphyu township is more efficient than that of Khittayar Hinthar RSC
in Pyay township. Therefore, rather than contracting with group of smallholders, the arrangement such
as contracting with individual smallholder should be considered in order to promote the contract
farming system. In addition, credit seems to appear essential capital input to resource poor smallholders
and informal contract farming model including credit provision should be considered in expansion of
contract farming network to other rice growing area in Myanmar.
REFERENCES
Abdulai, Y. (2016). Effects of contract farming on small-holder soybean farmers’ income in the eastern
corridor of the northern region, Ghana. Journal of Economics and Sustainable Development.
ISSN 22222-1700 (Paper) ISSN 22222-2885 (Online). Vol 7, No.2, 2016.
ACDGI (ASH CENTER FOR DEMOCRATIC GOVERNANCE AND INNOVATION). (2011). Myanmar
agriculture in 2011: Old problems and new challenges. Harvard kennedy School. 79 John F.
Kennedy street, Cambridge, Massachusetts 02138.
Adam Kephas, S. (2011). Economic impact of contract farming on income of smallholder farmers:
Case study of Paprika contract farming in Mbozi district, Tanzania, Master Thesis, Economics
of Sokoine University of Agricultrue. Morogoro. Tanzania
Azumah, S.B.,Donkoh, S. A. and D. S. Ehiakpor. (2016). Examining the determinants and effects of
contract farming on farm income in the northern region of Ghana. Ghana Journal of Science,
Technology and Development.
Cai, J., Ung, L., Setboonsarng, S., and Leung, P. S. (2008). Rice contract farming in Cambodia:
Empowering farmers to move beyond the contract toward independence ADB Institute
Discussion paper No. 109.
Chang, C. C., Chen, C. C., Chin, M. C. and Tseng, W.C. (2006). Is contract farming more profitable
and efficient than non-contract farmingA survey study of rice farms in Taiwan. Selected
paper prepared for presentation at the American Agricultural Economics Association annual
meeting, Long Be ach, California, July 23-26, 2006
CSO. (2012). Statistical year book (Various issues). Central Statistical Organization, Ministry of
Economic Development in collaboration with department of Agricultural Planning, MOAI,
Yangon, Myanmar.
FAO (2014). Agriculture production, consumption and trade statistics. Retrieved Nov 8, 2014, from
FAOSTAT database. http://www.faostat3.fao.org. Food and Agricutlrue Organization, Rome,
Italy, Updated 2014.
Fuglie, K. O. and Bosch, D. J. (1995) Economic and environmental implications of soil nitrogen testing:
a switching regression analysis. American Journal of Agricultural Economics 77:891900
Fujii, Y. and Satyanarayan, R. (2015). Myanmar’s untapped opportunity: How agribusiness can
transform its economy. KPMG. www.kpmg.com.sg.
Haile, A. G. (2015). Integrated rainwater harvesting practices: Smallholder farmers’ decisions, impact
on livelihood and environmental services in Azgo watershed, Ethiopia, 2015 XII + 172 pp.
ISBN 978-3-8236-1704-4.
Lokshin, M. and Z. Sajaia (2004). Maximum likelihood estimation of endogenous switching regression
models. Stata Journal 4(3):282-289.
Maddala, G.S. (1986). Disequilibrium. self-selection and switching models. In Z. Griliches and D. I.
Michael (eds). Handbook of Econometrics, North-Holland, Elsevier Pp: 1633- 1682.
Mercy, M., Judith, O, Patience, M. and S Mwanarusi. (2013). Does contract farming improve
smallholder farmers income? The case of Avocado farming in Kenya. Invited paper presented
at the 4th International conference of the African Association of Agricultural Economists,
September 22-25, 2013, Hammamet, Tunisia.
Moyo, M. (2014). Effectiveness of a contract farming arrangement: A case study of tobacco farmers in
Mazowe district in Zimbabwe, Master Thesis. Stellenbosch University. http://www.fao.org/
uploads/media/moyo_effectiveness_%20%202014.pdf
Musara. J. P, Zivenge. E, Chagwiza. G, Chimvuramahwe. S., and Dube. P (2011) Determinants of
smallholder contract farming participation in recovering economy: Empirical results from
Patchway District, Zimbabwe, Journal of Sustainable Development in Africa (Volume 13,
No.4, 2011), ISSN:1520-5509, Clarion University of Pennsylvania, Clarion, Pennsylvania.
Noltze, M., Schwarze, S. and Qaim, M. (2013). Impacts of natural resource management technologies
on agricultural yield and household income: The system of rice intensification in Timor
Leste.Ecological Economics, 85, 59-68.
Ogeto, R. M., Mshenga, P., Cheruiyot, E. and C. N. Onyari. (2012). Influence of institutional factors on
sorghum production in Nakuru county, Kenya. Journal of Agricultural Economics and
Development Vol. 1(6), pp. 130-137, December 2012
Sharma, V. P. (2008). India’s agrarian crisis and corporate-led contract farming: Socio-economic
implications for smallholder producers. International food and agribusiness management
review. Volume 11, Issue 4, 2008.
Sriboonchitta, S. and Wiboonpoongse, A.(2008). Overview of contract farming in Thailand: Lessons
learned, ADBI, Discussion Paper No.112, Tokyo: Asian Development Bank Institute, 2008.
Tambo, J. A. (2015). Farmer innovation in rural Ghana: Determinants, impacts and identification. Ph.D
thesis, Department of Economics and Technological Change, Center for Development
Research (ZEF), University f Bonn, Bonn, Germany.
Verbeek, M. (2012). A guide to modern econometrics. Fourth edition John Wiley & Sons. World Bank.
(2014) Myanmar: Capitalizing on rice export opportunities, Economic and Sector Work.
Report Number: 85804, Document of the World Bank, Southeast Asia Sustainable
Development Unit, East Asia and Pacific Region, February 28, 2014.
Date submitted: November 11, 2019
Reviewed, edited and uploaded: December 2, 2019
Conference Paper
Full-text available
This study was carried out in Myanmar New Aye Yar Agriculture Co., Ltd., which is located in No. (16-A) V-4 road Dakkinathiri Township, Hotel Zone (3), Nay Pyi Taw Council Area, Myanmar. This research examined the input supply chain to farmers from the perspective view of the contract farming company. The purpose of this research was to study the linkage of the input-supplying process between farmers and contract farming company and to understand the opportunities, and threats that occur in this supply chain. Data were collected by face-to-face interviewing by using well-structured questionnaire to the General Manager from Myanmar New Aye Yar Co., Ltd. The contract farming company bought different kinds of agricultural input from the selected agricultural input company especially Myanmar Awba Company. After buying the inputs, the contract company provided farmers as fertilizers, pesticides, herbicides, and fungicides with the current market price. Moreover, the contract company supplied contracted farmers with high-quality seeds (registered seeds) and farm machinery at the current market price. For both summer paddy and monsoon paddy production, the company provided agricultural inputs and farm machinery varieties, and the amount and prices were the same. Due to supplying quality inputs to farmers, the contract company got the desired amount of paddy from the farmers. The contract company supplied agricultural techniques and training services to farmers by doing practical fieldwork from the contract company's extension workers. Moreover, this research also studied the opportunities and risks in the supply chain of contract farming systems. As the company provided inputs with affordable prices, farmers' trust on the company became stronger because they did not need to invest a lot of money at the start of production. Moreover, it also attracted the willingness of farmers to cooperate more with the company. On the other hand, the contract company also found it difficult to determine the price of supplied agricultural inputs because of price fluctuation.
Article
Full-text available
Purpose: Smallholder farmers' agriculture in developing countries particularly in low and lower-middle-income countries is known for poor production and productivity levels which has been related to the inadequate use of improved agricultural inputs and marketing systems, and in this case, contract farming. Therefore, this study aimed to investigate the various factors which affect contract farming among smallholder farmers in developing countries. However, trying to investigate those contract farming determinants by a single analytical method leads it to have limited findings, narrowed generalizability and difficulty in investigating major determinants that determine contract farming success. Therefore, a joint approach using factor analysis and meta-analysis can give a more comprehensive understanding of contract farming factors. Methodology: The data was gathered through a Systematic Literature Review, finding a total of 3007 studies from SCOPUS, PubMed, PubAg, and EMBASE. The PRISMA method was applied to, and finally, 35 peer-reviewed articles in English between January 1990 to September 2023 were selected. Findings: The meta-analysis showed interesting insights into factors influencing contract farming participation. Education and household size exhibited a negative association, suggesting that farmers with higher education levels and larger families may be less likely to participate in contract farming arrangements. This could be due to a preference for independent decision-making or the need for family labor in alternative income-generating activities. On the other hand, farmers with larger landholdings are more inclined to participate. To encourage broader participation, policymakers, and program designers could consider targeted outreach and support services for these specific demographics. Originality and value: Our new approach, joining meta-analysis and factor analysis, sheds innovative light on contract farming determinants. While the expected determinants like farm and household size, along with education, remained significant, our analysis showed unexpected nuances. Age of the household head emerged as potentially favoring younger, less experienced farmers. Moreover, access to extension services played an important role, while large household size might have a more complex influence depending on age composition. This comprehensive approach offers valuable insights for targeted outreach programs and collaboration with extension services to optimize contract farming adoption and success in developing countries.
Article
Full-text available
Due to the recent economic recession and a decade of economic stagnation, anomalies have been observed in the cotton sector. An examination of the determinants of farmer participation in cotton contract farming was carried out on 100 smallholder farmers. The case was Patchway in Kadoma. Purposive sampling was used to select the study villages. Snowballing was then used to identify respondents. Questionnaires and Focus Group Discussions were used to collect data. Observations augmented data collection. Data were analyzed using logit regression model and Friedman rank test. Most parameters were observed to significantly influence participation. Non flexibility of the contractual arrangements was the major problem faced as well as price inconsistency. Though opportunities such as policies to counter side marketing exist, stringent implementation and monitoring regimes need to be put in place. This baseline analysis will aid in the development of sustainable strategies in restoring viability of the enterprise.
Article
Full-text available
A simultaneous equations, or "switching-regression," model is developed to assess the impact of soil nitrogen (N) testing on N fertilizer use, crop yields, and net returns in corn growing areas of Nebraska. The results indicate that when there is uncertainty about the quantity of available "carry-over" N, N testing enables farmers to reduce fertilizer use without affecting crop yields. However, the value of information from N tests depends critically on cropping history and soil characteristics. These findings have implications for environmental and technology transfer policies designed to reduce nonpoint source water pollution.
Chapter
This chapter reviews the recent literature on disequilibrium and selectivity models. The cornerstone of the disequilibrium models is the minimum condition. One of the most disturbing points in the empirical applications is that the disequilibrium models have been mechanically applied with no discussion of what disequilibrium is due to and what the consequences are. In spite of all the limitations, the model has been the model with the most empirical applications. For instance, Sealy used the model to study credit rationing in the commercial loan market. It is used to estimate demand for money and savings functions in centrally planned economies and to study the demand for consumption goods in centrally planned economies. The main application of the methodology of disequilibrium model is to regulate markets and centrally planned economies, where there are price and quantity regulations. The chapter discusses the case of controlled prices and shows the way the analysis can be applied to credit markets with interest rate ceilings.
Article
Natural resource management (NRM) technologies, such as the system of rice intensification (SRI), have been proposed to tackle agricultural challenges such as decreasing productivity growth and environmental degradation. Yet, the benefits of NRM technologies for farmers are often debated. Impacts seem to be context-specific, which are especially relevant in the small farm sector with its large degree of agroecological and socioeconomic heterogeneity. This was not always considered in previous research. We analyze the impacts of SRI adoption on rice yield and household income among smallholder farmers in Timor Leste. Heterogeneity is accounted for in an endogenous switching regression framework. Comparing mean yield and income levels, we find no significant differences between SRI adopters and non-adopters. This is due to negative selection bias; SRI seems to be adopted more on plots and by farmers with less than average yields. Controlling for this bias reveals significant yield and income gains. Poor and non-poor households benefit from SRI adoption; small farms benefit more than larger farms. The results also suggest that in the context of Timor Leste SRI may not be beneficial when compared to conventional rice grown under favorable conditions. Some implications for future research are discussed.
Article
This article describes the movestay Stata command, which implements the maximum likelihood method to fit the endogenous switching regression model. Copyright 2004 by StataCorp LP.
Effects of contract farming on small-holder soybean farmers' income in the eastern corridor of the northern region
  • Y Abdulai
Abdulai, Y. (2016). Effects of contract farming on small-holder soybean farmers' income in the eastern corridor of the northern region, Ghana. Journal of Economics and Sustainable Development. ISSN 22222-1700 (Paper) ISSN 22222-2885 (Online). Vol 7, No.2, 2016. ACDGI (ASH CENTER FOR DEMOCRATIC GOVERNANCE AND INNOVATION). (2011). Myanmar agriculture in 2011: Old problems and new challenges. Harvard kennedy School. 79 John F. Kennedy street, Cambridge, Massachusetts 02138.
Economic impact of contract farming on income of smallholder farmers: Case study of Paprika contract farming in Mbozi district
  • Adam Kephas
Adam Kephas, S. (2011). Economic impact of contract farming on income of smallholder farmers: Case study of Paprika contract farming in Mbozi district, Tanzania, Master Thesis, Economics of Sokoine University of Agricultrue. Morogoro. Tanzania