Kasetsart Journal of Social Sciences xxx (2018) 1e7
ilable at ScienceDirectContents lists avaKasetsart Journal of Social Sciences
journal homepage: http: / /www.elsevier .com/locate/k jssDeterminants of maize farmers' performance in Benin, West
Africa
Cocou Jaure
s Amegnaglo a, b, *
a Department of Agricultural Economics and Agribusiness, University of Ghana, Legon, Ghana
b Faculte des Sciences Economiques et de Gestion, Universite d'Abomey Calavi, Beninarticle info
Article history:
Received 3 October 2017
Received in revised form 29 December 2017
Accepted 22 February 2018
Available online xxxx
Keywords:
Benin,
maize,
production constrains,
stochastic frontier analysis,
yield analysis* Department of Agricultural Economics and Agrib
Ghana, Legon, Ghana.
E-mail address: cocoujaures1@yahoo.fr.
Peer review under responsibility of Kasetsart Univ
https://doi.org/10.1016/j.kjss.2018.02.011
2452-3151/© 2018 Kasetsart University. Publishing
creativecommons.org/licenses/by-nc-nd/4.0/).
Please cite this article in press as: Amegnag
Journal of Social Sciences (2018), https://dABSTRACT
Increased agricultural productivity is the primary aim of all agricultural policies under-
taken in developing countries. Increased agricultural productivity involves not only the
analysis of factors limiting productivity but also efficiency because improved efficiency
leads to productivity improvement. This paper investigated the factors limiting maize
productivity in Benin based on a survey of 354 maize farmers. The mean maize yield was
1,347 kg/ha. The low level of maize yield in Benin is due to the lack of access to inputs,
capital, and the weak institutional environment in which farmers operate. Furthermore,
the efficiency model revealed that an increase in maize output of about 25 percent can be
achieved in the short run by adopting the best farming practices and by addressing socio-
economic and structural constraints. Policy should be encouraged that would facilitate
access to inputs, capital, and training, and promote the development of infrastructure in
farming areas.
© 2018 Kasetsart University. Publishing services by Elsevier B.V. This is an open access
article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/
4.0/).Introduction
Food insecurity and poverty are important matter for
farmers in Benin. Most of those reported as food insecure
and susceptible to risk of food insecurity in Benin have
agriculture as their main income-generating activity
(World Food Programme [WFP], 2014). Most farmers in
Benin live in rural areas that are characterized by a high
poverty level compared to urban areas (Institut National de
la Statistique et de l'Analyse Economique [INSAE], 2012).
Poverty reduction in Benin as in most developing countries
requires necessarily the transformation of the agricultural
sector and the improvement of farmers' conditions and
performance. From food security and food self-sufficiencyusiness, University of
ersity.
services by Elsevier B.V. T
lo, C. J., Determinants o
oi.org/10.1016/j.kjss.201standpoints, maize is a crucial crop and one of the six
strategic crops that the government of Benin has chosen to
develop intensively (Ministe
re de l'Agriculture, de l'Elevage
et de la Pêche [MAEP], 2017), with the introduction of
technological innovations and institutional support being
the options chosen by the government to achieve this
objective (Adjimoti, Kwadzo, Sarpong, & Onumah, 2017;
MAEP, 2017). However, little research has been conducted
on the appropriateness of these options and more impor-
tantly on the current level of efficient use of existing
technological innovations in Benin.
Maize is the major staple food and the most cultivated
crop in the country. Most farmers (85%) in Benin grow
maize (WFP, 2014) and about one-third of the agricultural
area harvested in Benin is devoted to maize production
(INSAE, 2013). The Government of Benin identifiedmaize as
one of the strategic crops to be intensively developed and
considerable resources, financial and technical assistance,
the introduction of new technologies, and supply of inputs,his is an open access article under the CC BY-NC-ND license (http://
f maize farmers' performance in Benin, West Africa, Kasetsart
8.02.011
2 C.J. Amegnaglo / Kasetsart Journal of Social Sciences xxx (2018) 1e7
1450
1400
1350
1300
1250
1200
1150
2011 2012 2013 2014 2015
Figure 1 Maize yield in Benin between 2011 and 2015have been invested in this sector. Despite these efforts, the
maize yield has decreased in Benin (Figure 1) from
1,422 kg/ha in 2011 to 1,281 kg/ha in 2015 (MAEP, 2017).
Furthermore, the current mean maize yield is far lower
than the potential achievable yields (between 3 and 5 t/ha)
(Azontonde, Igue, & Dagbenonbakin, 2010). This low per-
formance has been proposed for the country's failure to
achieve the production target of 1,900,000 t of maize by
2015. In 2015, the production deficit was 613,940 t of maize
(MAEP, 2017). However, maize consumption is increasing.
Between 1996 and 2016, The United States Department of
Agriculture (USDA) (2017) estimated that mean maize
consumption increased by 5 percent annually in Benin.
The improvement of the performance of agriculture in
Benin has become an emergency due to the rapid increase
in population and thereby the food demand. The popula-
tion of Benin is expected to double every 20 years if the
current trend continues (INSAE, 2013). It is clear that
feeding such an increasing population will require an in-
crease in food production to minimize food insecurity. To
ensure a rapid increase in food production with the aim of
meeting future food demand, it is important to have a clear
picture of the state and prospects of Beninese agricultural
production and to identify possible challenges which could
impede the reliably of supplying increasing quantities of
important food commodities. The identification of agri-
cultural inputs with the highest efficiency can help to
achieve the goal of a rapid increase in food supply. Rightly,
Farrell (1957) suggested that efficient management of
existing technologies leads to productivity improvement.
Further, efficiency can help to close the gap between actual
and potential outputs (Audibert, 1997).
Few studies have attempted to analyze the factors that
determine farmers' performance in general and maize
farmers' performance in particular in Benin. For instance,
Degla (2015) analyzed the technical efficiency of cashew
nuts in Benin and found that the mean efficiency score was
63 percent while Kpenavoun Chogou, Gandonou, and
Fiogbe (2017) reported 67 percent for the mean technical
efficiency of pineapple farmers in Southern Benin. Only the
study conducted by Toleba Seidou, Biaou, Zannou, and
Saïdou (2016) was related to maize. The authors sug-
gested that the mean technical efficiency was around 80
percent and the main efficiency determinants are sex,
formal education, access to extension services, use of agro-Please cite this article in press as: Amegnaglo, C. J., Determinants o
Journal of Social Sciences (2018), https://doi.org/10.1016/j.kjss.201chemical, access to credit, use of animals and mechanical
traction. However, that study did not analyze the de-
terminants of productivity. Houssa, Reding, and Sotirova
(2017) analyzed vegetables and rice farmers' productivity
in two regions of Benin (Mono and Couffo) and found that
capital, labor, education, and soil type significantly affected
farmers' productivity. Those authors also observed an in-
verse relationship between farm size and productivity and
a difference in productivity between female and male
producers.
This article provides additional empirical evidence on
the performance of farmers in Benin (West Africa) by
examining the following question: What are the de-
terminants of maize farmers' performance in Benin? To
answer this question, we estimated the farmers' produc-
tivity function and the technical efficiency function.
Methods
Productivity Model
Generally, agricultural productivity depends on four
factors: inputs (seeds, fertilizer, labor, capital, etc.), farm
land' characteristics (for example, quality of the soil),
weather conditions and farmers' characteristics (gender,
education, experience, etc.) (Houssa et al., 2017). In the
literature, different production functions such as constant
elasticity of substitution, translog functions, and Cobb
eDouglas functions have been used (Houssa et al., 2017).
In this study, we used a production function that theoret-
ically shows the interdependence of the four factors and
that can also incorporate the observable and unobservable
characteristics that matter for productivity. The Cobb
eDouglas production function was adopted in spite of the
assumption that the input partial elasticities of substitution
equal one in the defined input space. The CobbeDouglas
function in this case can be specified as shown in Equa-
tion (1):
FðK; L; T ;XÞ ¼ AKdLbTgXa (1)
where K is the capital used in production, L is the labor used
in production, Tis the size of the farm, X represents other
inputs including farmers individual characteristics, and A is
the technology of production. The power related to every
factor of production (d;b;g;a) is a number between zero and
one. The specification of the CobbeDouglas production
function which we chose for the econometric estimations
with the data has the form shown in Equation (2):
X
lnðYiÞ ¼ a0 þ dlnðKiÞ þ blnðLiÞ þ gðTiÞ þ akXik þ εi (2)
k
where Yi is the productivity of producer i, Ki is the stock of
capital used by producer i, Li is the quantity of labor used by
producer i, Ti is land size used by producer I, Xi represents
other explanatory variables, and a0, the constant in the
estimation, can be interpreted as total factor productivity.
The logelog form of the econometric specification was
adopted in order to interpret the estimates from the
regression as elasticities of productivity with respect to the
given factor of production.f maize farmers' performance in Benin, West Africa, Kasetsart
8.02.011
C.J. Amegnaglo / Kasetsart Journal of Social Sciences xxx (2018) 1e7 3
Table 1
Description of variables
Variable Description
Ferti Amount of inorganic sources of plant nutrient used in maize
cultivation (kg/ha)
Capi Total monetary expenses incurred for all operations related
to maize cultivation (USD/ha)
Lab Total family and hired labor used for all operations related to
maize cultivation (labor days/ha)
Weedi Amount of active ingredient of plant protection chemicals
used in maize cultivation plot (L/ha)
Land Size of the land devoted to maize production (ha)
Seed Amount of maize seed used in cultivation plot (kg/ha)
Credit Whether has access to formal institutional credit; 1 ¼ yes,
0 ¼ otherwise
Market Whether has access to formal institutional and physical
market; 1 ¼ yes, 0 ¼ otherwise
Sex Male ¼ 1 and Female ¼ 0
Exten Whether has access to formal institutional extension
services; 1 ¼ yes, 0 ¼ otherwise
Expe Number of years the farmer is engaged in maize cultivation;
Measured in years
Yeduc Years of schooling
Hhsize Number of members in a farm family who share food from a
single source; Absolute number of members in a family
Educ Formal education received by a farmer; Categorized as
0 ¼ none; 1 ¼ primary and 2 ¼ post primary
Yield Production of maize grain per unit area (kg/ha)The empirical model used is described in Equation (3).
LnYield ¼ b0 þ b1LnFertiþ b2LnCapiþ b3LnLab
þ b4LnWeediþ b5LnLandþ b6LnSeedþ b7Credit
þ b8Market þ b9Sexþ b10Extenþ b11Expe
þ b12Yeducþ V1
(3)
where Yield is the maize production per hectare and V1 is
the equation error term. Details of the variables used in the
regression analysis along with their measurements are
given in Table 1.Efficiency Model
Two broad approaches (data envelopment analysis
(DEA) and stochastic frontier analysis (SFA)) have been
developed to measure efficiency according to Farrell's
definition. SFA is adopted because of its capacity to take
into account several exogenous factors (weather patterns,
diseases, poaching, etc) that are not under a farmer's con-
trol but can affect the farmer's efficiency. SFA also takes into
account measurement errors, statistical noise, and differ-
ential rates of adoption of technology that can affect the
farmer's efficiency (Aigner, Lovell, & Schmidt, 1977;
Meeusen & van den Broeck, 1977).
Following Farrell's (1957) conception of technical effi-
ciency as deviations from an idealized frontier isoquant,
Aigner et al. (1977) and Meeusen and van den Broeck
(1977) proposed that the production technology of a farm
is represented by a stochastic frontier production function.
The model has the form shown in Equation (4):
Yi ¼ f ðXi; bÞexpðεiÞ ¼ f ðXi; bÞexpðvi  uiÞ i ¼ 1;…;N (4)Please cite this article in press as: Amegnaglo, C. J., Determinants o
Journal of Social Sciences (2018), https://doi.org/10.1016/j.kjss.201where Yi is the observed output of farm i and fðXi; bÞ is a
function such as a CobbeDouglas or translog production
function of the vector and represents the maximum quan-
tity that can be produced with Xi (vector of inputs) and
technologydescribedby theparametersb. Furthermore, εi is
the error term that is composed of two independent ele-
ments vi and ui such that εi ¼ ðvi  uiÞ. Production can
deviate from the deterministic frontier because of random
shocks vi, which could be positive or negative, or because of
the non-negative inefficiency error term ui, which reduces
output (ui  0). Finally, vi is an iid error termwithmean zero
and constant variance assumed to be independent of ui. The
aim of the stochastic frontier model is to construct the
production frontier and the inefficiency can be estimated by
the deviations from this frontier.
The technical efficiency (TE) of the ith farm is provided
by Equation (5):

¼ ð  Þ ¼ f ðXi; bÞ 

expðvi  uiÞTEi exp ui (5)f ðXi; bÞ  expðviÞ
where TE is the ratio of actual output relative to the po-
tential output. TE takes a value between 0 and 1 with a
smaller ratio reflecting inefficiency.
The estimation of the parameters of the production
function requires the imposition of an appropriate distri-
bution concerning the inefficiency error term ui. Using the
assumption that the inefficiency effects are half normally
distributed i.e. u  iidNþi ð0; s2uÞ, the technical inefficiency
effect is defined using Equation (6):
ui ¼ Zidþ qi (6)
where Zi is a ðP  1Þ vector of explanatory variables asso-
ciated with the technical inefficiency effect such as socio-
economic, farm management, and institutional
characteristics and q is the error term of the inefficiency.
Model Specification
A CobbeDouglas production function as shown in
Equation (7) was chosen in this study with the imposition
of monotonicity and the quasi-concavity hypothesis on the
production function. This functional form was chosen
because it is flexible, self-dual, and its returns to scale are
easily interpreted (Bravo-Ureta & Evenson, 1994).
XJ
Yi ¼ ai þ bjXji (7)
j¼1
where Yi is the output of maize (kilograms) produced in
2015 cropping season by the ith farmer; X is a set of eight
inputs: land size, labor, seed, weedicide, equipment, fer-
tilizer, b denotes the unknown parameters to be estimated,
vi denotes random shocks, and ui is the one-sided, non-
negative error representing inefficiency in the production.
The empirical model used is described in Equation (8).
LnOutput ¼ b0 þ b1LnFertiþ b2LnCapiþ b3LnLab
þ b4LnWeediþ b5LnLandþ b6LnSeedþ ε (8)
whereOutput is the output of maize (kilograms) and ε is thef maize farmers' performance in Benin, West Africa, Kasetsart
8.02.011
4 C.J. Amegnaglo / Kasetsart Journal of Social Sciences xxx (2018) 1e7
Table 2
Socio-economic characteristics of survey respondents
Variables Mean Std. Dev.
Mean maize yield (kg/ha) 1,347 484.6
Use of fertilizer (%) 56.8 49.6
Quantity of fertilizer used (kg/ha) 94.7 104.5
Weedicide applied (L/ha) 1.2 2.3
Capital (USD/ha) 34.7 39.7
Labor (labor-days/ha) 150.6 80.2
Farm size (ha) 3.9 4.4
Seed used (kg/ha) 16.5 11.2
Farm size (ha) 3.9 4.4
Sex (percentage male) 73.1 44.4
Age (years) 41.7 12.6
Maize farming experience (years) 22.2 11.7equation error term. Details of the variables used in the
regression analysis along with their measurements are
given in Table 1.
The inefficiency model of the stochastic frontier func-
tion is given by Equation (9):
u ¼ d0 þ d1Credit þ d2Market þ d3Sexþ d4Extenþ d5Expe
þ d6Dfertiþ d7Hhsizeþ d8Educþ q
(9)
where mdenotes farm specific inefficiency, d denotes a set of
parameters to be estimated, and the variables that explains
farmers' inefficiency equation are explained in Table 1.Household size 10.9 7.1
Access to extension services (%) 33.9 47.4
Access to credit (%) 52.2 50.0
Access to market (%) 89.8 25.2
Educational attainment level
No education at all 61.6 48.7
Primary School 28.5 45.2
Post Primary School 9.9 29.9Hypothesis Test
The following hypotheses were investigated:
(1) H0 : l ¼ 0, the null hypothesis specifies that inefficiency
effects are absent from the model at every level. How-
ever, l>0 means that the technical inefficiency effects
are present in the model and hence the use of a sto-
chastic frontier model is best.
(2) H0 : d1 ¼ … ¼ d8 ¼ 0, the null hypothesis that farm
specific factors do not influence the inefficiencies.
These hypotheses were tested using the generalized
likelihood-ratio statistic.Data Collection
A multi-stage, cluster-based, random sampling
approachwas used to select the respondents. The first stage
of the design consisted of the random selection of one
municipality in each of the three climatic zones present in
the country. The municipalities of Kandi, Glazoue, and Ze

were then randomly selected. In Benin, a municipality
consists of several districts. The second stage of the selec-
tion process, therefore, consisted of the random selection of
three districts per selected municipality. The third stage of
the selection process consisted of the selection of villages in
each district. Two villages were randomly selected from
each district, making up 18 villages for the nine districts
and the three municipalities. The fourth and final stage was
the actual selection of farmers to be interviewed. A struc-
tured questionnaire was used for the interview. The
optimal sample size for the number of farmers selected for
the whole study was 323. The determination of the optimal
sample size was based on Babbie (2016) dealing with
sampling from very large population sizes. Oversampling
was used and hence 396 farmers were chosen for the study.
This oversampling was done due to the possibility of some
farmers refusing to participate in the study. The survey was
conducted in 2015. For each village, the approximate
number of farmers was provided by the village chief. The
farmers selected were those known to be available in the
village at the time of the study. Twenty-two (22) farmers
were randomly selected from each village. Amegnaglo,Please cite this article in press as: Amegnaglo, C. J., Determinants o
Journal of Social Sciences (2018), https://doi.org/10.1016/j.kjss.201Anaman, Mensah-Bonsu, Onumah, and Amoussouga Gero
(2017) presented details on the sampling techniques and
the study areas.Data Analysis
About 38 percent of farmers interviewed came from
Kandi municipality, 32 percent from Ze
 and 30 percent
from Glazoue. About 73 percent of the respondents were
male and one-third of the farmers were young (18e35
years). The mean age of sampled farmers was 41.7 years
(with 12.6 as the standard deviation) with the youngest
being 18 and the oldest 85 years (Table 2). Respondents
with no schooling were predominant (61.6%), while pri-
mary school leavers were the second largest class of re-
spondents (28.5%). The mean farming experience regarding
maize production is 22 years (with 11.7 as standard devi-
ation). The mean farm size for the whole group was about
3.9 hectares (with 4.4 as the standard deviation). One-third
of farmers interviewed had interacted with extension ser-
vices during the last three farming seasons (2012e2014).
About 57 percent of farmers had applied fertilizer during
the last cropping season. On average, a farmer used 94.7
(±104.5) kg of fertilizer, 1.2 (±2.3) L of weedicide, 16.5
(±11.2) kg of seed, 150.6 (±80.2) labor/day and USD 34.7
(±39.7) as capital per hectare for the production of an
average of 1,347(±484.6) kg of maize per hectare (Table 2).
Results and Discussion
Determinants of Productivity
The results of the estimation suggested that the quantity
of labor, fertilizer, capital, and seeding rate had significant
and positive effects on the productivity of maize farmers
(Table 3). The use of fertilizer significantly increased land
productivity. Furthermore, WFP (2014) showed that
depletion of land fertility and land degradation are thef maize farmers' performance in Benin, West Africa, Kasetsart
8.02.011
C.J. Amegnaglo / Kasetsart Journal of Social Sciences xxx (2018) 1e7 5
Table 3 Table 4
Productivity model results Maximum likelihood estimates of Stochastic CobbeDouglas production
frontier of maize
Variable All variables Only significant
variables Variable Coefficient t- value
Coefficient t-value Coefficient t-value LnCapi .09530*** 5.46
LnFerti .1805*** 4.06
LnFerti .0521*** 4.82 .0546*** 5.22
LnLab .0992*** 2.80
LnCapi .0953*** 4.56 .0952*** 4.66
LnWeedi .0061 .190
LnLab .0985*** 2.95 .0982*** 2.79
LnLand .7406*** 13.56
LnWeedi .0293 1.25
LnSeed .0399 1.21
LnLand .3481*** 6.83 .3475*** 6.96
Constant .2318*** 4.27
LnSeed .1199*** 3.17 .1215*** 3.42
  Log-likelihood function 115.7352Credit .0231 .64
Sigma square) d2( .1821 6.54
Market .1380** 2.16 .1455** 2.29
Lambda) l( 1.2045*** 15.75
Sex .0204 .54 2
d .3283 5.8442
Exten .0919** 2.15 .0893** 2.28 u2
dv .2726 11.3545Expe .0002 .12
Yeduc .0017 .35 ***p < .01
Constant 5.7604*** 28.78 5.1435*** 20.80
Number of obs ¼ 354, Number of obs ¼ 354,
F(12, 341) ¼ 29.89, F(7, 346) ¼ 48.13,
Prob > F ¼ .0000, R- Prob > F ¼ .0000, R-
squared ¼ 0.4335 squared ¼ 0.4295
**p < .05, ***p < .01main problems faced by farmers in Benin, making access to
fertilizer indispensable for maize production in the
country.
Farm size significantly and negatively affected farmers'
productivity. The inverse sizeeproductivity relationship
has been observed by several researchers in Africa and this
relationship is explained in the literature by factors like
market imperfections, lack of data on soil condition, and
measurement error (Ali & Deininger, 2015; Carletto,
Savastano, & Zezza, 2013; Houssa et al., 2017).
Access tomarket significantly increasedmaize yield. It is
argued that commercialization can positively affect yield
through specialization (better resource allocation), inten-
sification (increased use of inputs), and the reduction of
loss of perishable harvests (Jaynes, 1994; Nonvide, 2017).
Access to extension services significantly and positively
increased maize productivity. Extension workers can pro-
vide information on and explain the best agricultural
technologies and practices. Farmers who benefit from
extension visits are more likely to use the recommended
technologies appropriately (Onumah, Brümmer, &
Ho€rstgen-Schwark, 2010).Table 5
Estimated inefficiency of maize farmers in Benin
Inefficiency model Coefficient t- value
Constant 2.9525*** 3.42
Credit .5831 1.59
Market .8283* 1.80
Sex .6325* 1.71
Exten 1.2558*** 2.88
Expe .0296** 2.09
Hhsize .0347* 1.95
Educ
Primary .7796** 2.34
Post primary 1.5237*** 3.11
*p < .10, **p < .05, ***p < .01Efficiency Model
The estimated parameters of the stochastic frontier
model (Table 4) and the inefficiency model (Table 5) are
presented below as Frontier and Inefficiency models. The
results of the estimation showed that the test of absence of
inefficiency effects in the model was rejected. Furthermore,
the test specifying that the inefficiency effects are not sto-
chastic was strongly rejected. The estimated lambda was
significantly greater than zero, implying that the traditional
average for the ordinary least squares, (OLS) function is not
an adequate representation for the data. The hypothesis
stating that the intercept and the coefficients associated
with all variables in the technical inefficiency model are
zero was strongly rejected. This suggested that the exoge-
nous variables jointly explained the technical inefficiencyPlease cite this article in press as: Amegnaglo, C. J., Determinants o
Journal of Social Sciences (2018), https://doi.org/10.1016/j.kjss.201effects. Therefore, it allows the identification of relevant
policy variables.
The expected coefficients of all the factors were positive
but the coefficients of seedling rate andweedicide were not
significant. The highest elasticity was related to the land
size, implying that a 1 percent increase in land size devoted
to maize culture will increase the production by 0.74
percent. Compared to the results obtained earlier (Table 3),
land had a positive effect on maize output but a negative
impact on maize yield. Farmers with a small-sized maize
farm may apply better monitoring on their farms (field
maintenance, fertilizer application on time) and use tech-
nologies (such as application of manure) which are not
convenient on large-sized maize farms. Subsequently, they
obtained better yield.
The output elasticity for fertilizer was 0.18 percent,
indicating that an increase in fertilizer use by 1 percent will
increase maize production by 0.18 percent. The elasticities
of output with respect to capital and labor were 0.09
percent and 0.10 percent, respectively. The return to scale
was 1.11. This finding suggests that if all the inputs were
multiplied by 1 percent, the mean production would be
multiplied by 1.11 percent, thus revealing the existence of
economies of scale.
Estimated parameters in the technical inefficiency
model revealed that access to market, access to extension
services, and gender significantly reduced farmers' tech-
nical inefficiency while education, household size, and
farming experience significantly increased farmers' tech-
nical inefficiency (Table 5).f maize farmers' performance in Benin, West Africa, Kasetsart
8.02.011
6 C.J. Amegnaglo / Kasetsart Journal of Social Sciences xxx (2018) 1e7
50
45
40
35
30
25
20
15
10
5
0
< 0.50 0.51-0.60 0.61-0.70 0.71-0.80 0.81-0.90 > 0.91
 
Figure 2 Frequency distribution of technical efficiencyThe coefficient estimated for the gender dummy was
significantly negative, indicating that male farmers oper-
ated less inefficiently than their female counterparts.
Women generally lack access to financial and human re-
sources to conduct appropriately their farming operations
(Toleba Seidou et al., 2016). Furthermore, the women's
domestic role does not give them enough time to spend on
the farm, contributing to inefficiency of production
(Onumah et al., 2010). Access to extension services
improved farmers' technical efficiency as found by Toleba
Seidou et al. (2016) as such access equips farmers with
the requisite knowledge on the best inputs to employ and
their appropriate use.
The coefficient of market access was negative and
significantly related to technical inefficiency. Farmers who
had access to an output market were more efficient
compared to those who did not have access to markets.
Access to market can positively affect efficiency through
specialization, intensification, and reduction of losses
(Jaynes, 1994; Nonvide, 2017).
Experience in farming was estimated to be significantly
positive, indicating that the more-experienced farmers
were more technically inefficient in their production than
new farmers who were more willing to implement new
production systems (Onumah et al., 2010). The coefficient
of education in this study was surprisingly positive, sug-
gesting that farmers with a high level of formal education
(at least post primary) operated inefficiently in their pro-
duction. Though education is an important factor influ-
encing efficiency, Kalirajan and Shand (1985) argued that
illiterate farmers can understand modern production
technology as well as their educated counterparts when
they are trained and the technology is communicated
properly. Furthermore, maize is produced in Benin using
traditional methods and the education of farmers might
not play a role in the optimal combination of inputs. This
result was substantiated by the finding of Onumah et al.
(2010) but it is contrary to the findings of Toleba Seidou
et al. (2016) and Battese, Malik, and Gill (1996).
Household size had a positive relationship with tech-
nical inefficiency indicating that farmers with a larger
household sizewere less efficient than farmers with a small
household size. Farmers with a larger household size rely
more on family labor, however Kloss and Petrick (2014)
suggested that hired labor is more productive than family
labor. Furthermore, larger households in developingPlease cite this article in press as: Amegnaglo, C. J., Determinants o
Journal of Social Sciences (2018), https://doi.org/10.1016/j.kjss.201countries may put more pressure on farm income and
thereby reduce the available income for investment in
agriculture (acquisition of productivity-enhancing inputs).
Technical Efficiency of Maize Farmers
The technical efficiency of maize farmers ranged from
0.22 to 0.94 (Figure 2). The predicted mean technical effi-
ciency was estimated to be 0.75 and this means that about
25 percent of the technical potential output was not
realized.
Therefore, increasing maize farming production by an
average of about 25 percent can be achieved in the short
run by adopting good maize production practices.
Conclusion
This paper examined the determinants of maize
farmers' performance in Benin based on a sample of 354
farmers across the three main climatic zones in Benin. The
results revealed that the mean maize yield in Benin was
around 1,347 kg/ha. The increase in maize yield in Benin
requires production intensification (capital, fertilizer, labor,
and seed) rather than increased land area. Improvement of
access to extension services and markets would also
contribute to yield improvement.
Another avenue for yield increase is improved farmer
efficiency. The results indicated that there was significant
variation in technical efficiency among maize farmers, as
the estimated efficiency ranged from 0.22 to about 0.94
with mean technical efficiency levels of 75 percent. This
indicates that maize production could be increased by 25
percent through better use of available resources such as
land, labor, seed, and fertilizer given the state of technol-
ogy. Farmers were operating at an increasing return to
scale. Male maize farmers who had not had a formal edu-
cation, with lower farming experience, and a smaller
household size, but with relatively good access to markets,
extension services, and the use fertilizer achieved higher
levels of technical efficiency in maize production in Benin.
Policy Recommendation
Based on the findings, the policymakers could assist
improving production and yield through better and reliable
access to key inputs such as fertilizer, labor, seeds, andf maize farmers' performance in Benin, West Africa, Kasetsart
8.02.011
C.J. Amegnaglo / Kasetsart Journal of Social Sciences xxx (2018) 1e7 7equipment. It is important for the government to create an
institutional environment that facilitates reliable access to
markets and extension services to farmers. Roads should be
constructed that improve market access. Due to the fact
that farmers operate below the frontier line, the policy-
makers need to organize training and educational pro-
grams to improve crop management practices and thereby
the farmers' technical efficiency. Policies should be imple-
mented that aim to empower female farmers through
better access to economic resources, education, informa-
tion, and decision-making.
Conflict of Interest
There is no conflict of interest.
References
Adjimoti, G. O., Kwadzo, G. T. M., Sarpong, D. B., & Onumah, E. E. (2017).
Input policies and crop diversification: Evidence from the Collines
region in Benin. African Development Review, 29(3), 512e523. https://
doi.org/10.1111/1467-8268.12286.
Aigner, D., Lovell, C. A. K., & Schmidt, P. (1977). Formulation and esti-
mation of stochastic frontier production function models. Journal of
Econometrics, 6, 21e37.
Ali, D. A., & Deininger, K. (2015). Is there a farm size-productivity rela-
tionship in African agriculture? Evidence from Rwanda. Land Eco-
nomics, 91(2), 317e343. https://doi.org/10.3368/le.91.2.317.
Amegnaglo, C. J., Anaman, K. A., Mensah-Bonsu, A., Onumah, E. E., &
Amoussouga Gero, F. (2017). Contingent valuation study of the ben-
efits of seasonal climate forecasts for maize farmers in the Republic of
Benin, West Africa. Climate Services. https://doi.org/10.1016/
j.cliser.2017.06.007.
Audibert, M. (1997). Technical inefficiency effects among paddy farmers
in the villages of the ‘Office du Niger’, Mali, West Africa. Journal of
Productivity Analysis, 8, 379e394. https://doi.org/10.1023/A:
1007767508848.
Azontonde, A. H., Igue, A. M., & Dagbenonbakin, G. (2010). Carte de fertilite
des sols du Benin par zone agroecologique du Benin. Benin: Rapport
Final. [in French]
Babbie, E. R. (2016). The practice of social research (14th ed.). Boston, MA:
Cengage Learning.
Battese, G. E., Malik, S. J., & Gill, M. A. (1996). An investigation of technical
inefficiencies of production of wheat farmers in four districts of
Pakistan. Journal of Agricultural Economics, 47(1), 37e49. https://
doi.org/10.1111/j.1477-9552.1996.tb00670.x.
Bravo-Ureta, B. E., & Evenson, R. E. (1994). Efficiency in agricultural pro-
duction: The case of peasant farmers in eastern Paraguay. Agricultural
Economics, 10(1), 27e37. https://doi.org/10.1016/0169-5150(94)
90037-X.Please cite this article in press as: Amegnaglo, C. J., Determinants o
Journal of Social Sciences (2018), https://doi.org/10.1016/j.kjss.201Carletto, C., Savastano, S., & Zezza, A. (2013). Fact or artifact: The impact of
measurement errors on the farm size-productivity relationship.
Journal of Development Economics, 103(1), 254e261. https://doi.org/
10.1016/j.jdeveco.2013.03.004.
Degla, P. (2015). Technical efficiency in producing cashew nuts in Benin's
Savanna Zone, West Africa. Quarterly Journal of International Agricul-
ture, 54(2), 117e132.
Farrell, M. J. (1957). The measurement of productive efficiency. Journal of
the Royal Statistical Society. Series A (General), 120(3), 253. https://
doi.org/10.2307/2343100.
Houssa, R., Reding, P., & Sotirova, A. (2017). Financing for MSMEs in the
agricultural sector. Part II. A case study on projects of the Belgian
Technical Cooperation in Benin (Working Paper). Belgium: Belgium
Policy Research Group on Financing for development.
INSAE. (2012). Enquête Modulaire Integree sur les Conditions de Vie des
menages 2
eme Edition (EMICoV 2011). Cotonou. [in French]
INSAE. (2013). Resultats provisoires du rgph4. Cotonou. [in French]
Jaynes, T. S. (1994). Do high food marketing costs constrain cash crop
production? Evidence from Zimbabwe. Economic Development and
Cultural Change, 42(2), 387e402. https://doi.org/10.1086/452086.
Kalirajan, K. P., & Shand, R. T. (1985). Types of education and agricultural
productivity: A quantitative analysis of Tamil Nadu rice farming. The
Journal of Development Studies, 21(2), 232e243. https://doi.org/
10.1080/00220388508421940.
Kloss, M., & Petrick, M. (2014, September). The productivity of family and
hired labour in EU arable farming. Goettingen, Germany: Paper pre-
sented at the 54th Annual Conference.
Kpenavoun Chogou, S., Gandonou, E., & Fiogbe, N. (2017). Mesure de
l'efficacite technique des petits producteurs d'ananas au Benin. Ca-
hiers Agricultures, 26(2), 1e6. https://doi.org/10.1051/cagri/2017008.
[in French]
MAEP. (2017). Plan Strategique de Developpement du Secteur Agricole
(PSDSA) 2025 et Plan National d' Investissements Agricoles et de Securite
Alimentaire et Nutritionnelle PNIASAN 2017-2021. Cotonou. [in French]
Meeusen, W., & van den Broeck, J. (1977). Efficiency estimation from
Cobb-Douglas production functions with composed error. Interna-
tional Economic Review, 18(2), 435. https://doi.org/10.2307/2525757.
Nonvide, G. M. A. (2017). Effect of adoption of irrigation on rice yield in
the municipality of Malanville, Benin. African Development Review, 29,
109e120. https://doi.org/10.1111/1467-8268.12266.
Onumah, E. E., Brümmer, B., & Ho€rstgen-Schwark, G. (2010). Productivity
of hired and family labour and determinants of technical inefficiency
in Ghana's fish farms. Agricultural Economics, 56(2), 79e88. https://
doi.org/10.1111/j.1477-9552.2012.00363.x.
Toleba Seidou, M., Biaou, G., Zannou, A., & Saïdou, A. (2016). Evaluation Du
Niveau D'efficacite Technique Des Syste
mes De Production A Base De
Maïs Au Benin. European Scientific Journal, 12(27), 276e299. https://
doi.org/10.19044/esj.2016.v12n27p276. [in French]
USDA. (2017). Statistical database. Retrieved from https://quickstats.nass.
usda.gov/.
WFP. (2014). Analyse Globale de la Vulnerabilite, de la Securite Alimentaire
et de la Nutrition (AGVSAN). Roma, Italy: Republique du Benin. [in
French]f maize farmers' performance in Benin, West Africa, Kasetsart
8.02.011