J. Agric. Environ. Sci. Vol. 6 No. 1 (2021) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Journal of the College of Agriculture & Environmental Sciences, Bahir Dar University 9
Determinants of Technical Efficiency in Agricultural Production among Sub Saharan
African Countries
Davis Bundi Ntwiga
School of Mathematics, University of Nairobi
Corresponding author: dbundi@uonbi.ac.ke
Received: October 16, 2020 Accepted: January 9, 2021
Abstract: Climate change has led to a decline in agricultural production due to erratic weather patterns,
compromised crop yields and population pressure on arable land. Sub-Saharan Africa is most vulnerable to
climate change due to its geographical location, increase in population, destruction of the forests and other
agricultural malpractices. This is a threat to livelihoods, food systems, and increase in malnutrition and shocks
in food prices. This study examines the influence of climatic factors on the technical efficiency of agricultural
production in Sub Saharan Africa using time series data for 25 years from 1991 to 2015 selected from nine
countries. The data envelopment analysis estimates technical efficiency with input variable as agricultural land
and output variable as agricultural value-added. The panel data analysis response variable is the technical
efficiency scores. Predictor variables were population, forest area, temperature, rainfall, and greenhouse gases.
In the last 25 years, there has been an increase in population, agricultural land, temperature, and greenhouse
gases with a decrease in forest area and rainfall. Temperature, forest area, and greenhouse gases showed
significant influences on the technical efficiency of agricultural production. The intricate nature of climate
change requires significant efforts to reverse the trend being observed and boost agricultural production
efficiency.
Keywords: Climate change, greenhouse gases, panel analysis, Sub-Saharan Africa, technical efficiency
This work is licensed under a Creative Commons Attribution 4.0 International License
1. Introduction
The last two decades have witnessed a drastic
change in climatic conditions that have led to a
decline in food production. Climate change is
taunted to increase the frequency and intensity of
disasters, disruption of food production and
livestock rearing. This has raised concern among
public policymakers and interest groups due to
uncertainty in food security and agricultural
sustainability (Eniko et al., 2018). Climate change
is one of the several changes affecting food
systems and this varies between regions and the
different social groups within a region. The threat
of global food shortage is evident due to a number
of factors: population pressure, water scarcity, land
degradation, frequent droughts, declining soil
fertility, lack of credit facilities, poor agronomic
practices, poor seed quality, pests, weeds and
incidence of diseases (Abera et. al., 2018; Nsiah
and Fayissa, 2019; Popp et. al., 2019). African is
already overburdened with food insecurity, poverty
and low adaptive capacity but climate change is
projected to increase the vulnerability and burden
(Muller et. al., 2011). The models that derive the
relationship between environmental conditions and
production systems project a continued decline in
crop yields due to climate change (Ray et. al.,
2019) with traditional rain-fed agriculture facing
more climate-related risks (Eniko et al., 2018).
Smallholder systems in Africa will be the most
compromised in agriculture production due to little
adaptive capacity to climate change (Muller et. al.,
2011). In Asia and Latin-America, there is
improved food security that has also reduced the
prevalence of undernourishment (Popp et. al.,
2019). Smallholder farmers account for 80% of all
the farms in SSA employing million directly
and 70% of the smallholder farmers being women.
Subsistence farming is a key employer with 76 %
of the population in Botswana, 85% in Kenya and
90% in Malawi depending on agriculture (AGRA,
2014). The future impacts and exposure to climate
variability and extremes are expected to increase
with time (FAO, 2018). Food systems are changing
rapidly due to globalization and urbanization with
an increasing population (Gregory et al., 2005;
Nsiah and Fayissa, 2019). This is compounded by
J. Agric. Environ. Sci. Vol. 6 No. 1 (2021) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Journal of the College of Agriculture & Environmental Sciences, Bahir Dar University 10
the expected growth of in the SSA population
by the year projected to be 1.5 billion (Nsiah
and Fayissa, 2019) and an additional of 2.4 billion
people in the world by 2050 (Islam and Wong,
2017). The rate of population growth exceeds the
agricultural production due to declining land size,
climate change and other vulnerabilities facing
farmers (Nsiah and Fayissa, 2019). Worldwide,
million people are undernourished, with
million in SSA and million are in Africa
(Islam and Wong, 2017; FAO, 2018).
The risks for African agriculture and food
production are due to anthropogenic climate change
with statistical, process-based and econometric
models indicating negative and positive impacts on
agriculture. The underlying assumptions in the
climate change projections and its impact on food
production are greenhouse gas emissions,
biophysical and socio-economic conditions.
Climate change has increased the global mean
annual air temperature by 0.74 °C and atmospheric
greenhouse gases during the last 100 years
(Tokunaga, et al., 2015). The vulnerability of the
African continent to the effects of climate change
are already evident, with predictions indicating that
Africa is warmer compared to the global average.
Temperature and rainfall are two key determinants
of agricultural production and food security (Abera
et. al., 2018). Climate change is worsening
agricultural production in Africa due to erratic
weather patterns and extreme weather events that
decrease the average yields (AGRA, 2014).
The agricultural production technical efficiency
(TE) study found that agricultural land, arable land,
rural population, average precipitation, land under
cereal production and economically active
population working in the agricultural sector,
access to credit and agricultural research influence
TE (Nsiah and Fayissa, 2019). An estimated 2.7
million hectares can be irrigated, but only 11% was
equipped for irrigation in 2001 (FAO, 2018). A
survey from 28 among 47 countries in SSA
indicated that 75% of the labour force worked in
household enterprises and the agricultural sector
(FAO, 2018) with households that are food secure
being TE and productive (Oyetunde-Usman and
Olagunju, 2017). The future maize yields in
Ethiopia are either increasing or decreasing based
on the region (Abera et. al., 2018) while Ngango
and Kim (2019) noted that coffee production TE in
Rwanda depends on technological adoption.
An input-oriented Data Envelopment Analysis
(DEA) employed to examine the TE of maize
production in northern Ghana noted that efficiency
can further be boosted through formal and informal
educational platforms to educate the farmers on
improved cultivation practices. The DEA employed
various variables, fertilizer consumption,
household size, household labor, maize plot size,
age of respondent, among other variables (Abdulai
et. al., 2018). The mean difference between food
secure and insecure households TE is and
was found to be statistically significant among the
agriculture households in Nigeria (Oyetunde-
Usman and Olagunju, 2017). A study on the
African agriculture and food production TE found
that it has decreased significantly over time
(Ogundari, 2014). The employment of the right
combination of productive resources to achieve
food sustainability is important for African
countries (Abdulai et. al., 2018). The technically
efficient food producers are more food secure to
non-technically significant producers. African
countries need to continue making agriculture a
critical component as it’s the principal source of
food, livelihood and a channel to reduce poverty
and attain food security (Ogundari, 2014).
A panel data analysis model was used to estimate
the impact of global warming-induced climate
change on agricultural production in Japan. The
results indicated that rising precipitation and
temperature and decreasing solar radiation reduced
rice production in Japan. A dynamic panel analysis
on rice, vegetable and potato showed a decline in
production. An increase of a degree in mean annual
temperature reduces rice production by 5.8% and
3.9%, and potato production by 5% and 8.6% in the
short and long term, respectively (Tokunaga et al.,
2015). In Burkina Faso, an increase in temperature
reduced the production of millet, maize and
sorghum while an increase in rainfall and
precipitation increased the production of the cereals
(Nana, 2019).
The goal of this study was to assess and identify
the climatic factors that influence technical
efficiency of agricultural production in SSA
through the DEA model and panel data analysis.
Climatic risks are changing the agricultural
production landscape in SSA with a reduction in
crop yields to cater for the increasing population.
An analysis of the relationship between the
environmental conditions and production system is
J. Agric. Environ. Sci. Vol. 6 No. 1 (2021) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Journal of the College of Agriculture & Environmental Sciences, Bahir Dar University 11
important to understand the influence of climatic
risks on agricultural production efficiency. The
intent is to understand to what extent climatic
conditions are influencing the agricultural
sustainability and food systems in SSA. We found
no similar study that has considered agriculture
value-added and agricultural land size in estimating
TE among SSA countries and use of forest area as
a predictor variable in panel data analysis. Forest
cover is key in absorbing greenhouse gases while
agricultural value-added is an indicator of the
interplay between the inputs and outputs in the
agricultural production systems.
2. Materials and Methods
The two-step DEA model was applied to estimate
the agricultural production TE among the selected
countries in SSA. The first step is to estimate the
efficiency scores on agricultural production. The
second step performs the panel data analysis to
estimate which climate variables have an influence
on the agricultural production TE scores. The study
covers a period of 25 years from to the year
with countries sampled from SSA. The DEA
input variable is agricultural land and the output
variable is the agricultural value-added, with data
sourced from FSP (2020). The five climate change
variables are forest area from World Bank portal
(World Bank (2020)) while rainfall, temperature,
population and greenhouse gases are from climate
watch data (CWD, 2020). The study has two
input/output variables. Therefore at least 8
Decision-Making Units (DMU) were required as
indicated by Ntwiga (2020). The study population
has 28 SSA countries with an overall food security
score of between 34.3 and 67.3% (Economist
Intelligent Unit, 2020). A total of top 9 DMUs with
no missing data points were selected from the 28
countries to form the study sample. The countries
include Benin (BEN), Botswana (BWA), Burkina
Faso (BFA), Cameroon (CMR), Ethiopia (ETH),
Ghana (GHA), Kenya (KEN), Mali (MLI) and
Nigeria (NGA) (Economist Intelligent Unit, 2020).
The efficiency scores were analyzed using DEA.
Then, the results were assessed using descriptive
statistics and econometrics model. The influence of
climatic factors on TE of agricultural production
was estimated using regression panel analysis. The
efficiency scores summary statistics were grouped
into four periods; 1991-2000, 2001-2010, 2011-
2015 and 1991-2015. The purpose was to check if
any significant changes can be attributed to these
time segments compared to the overall period. In
the determinant of TE, four models were derived
where model M1 was the two-dimensional
variables panel analysis. Model M2 and M3
captured one-way effect controlling for the year
and country, respectively while M4 captured two-
way effects controlling for both year and country.
2.1. Variables definition and measurement
The DEA input and output variables resulted to the
TE scores as the output variable. In the panel
analysis, the TE scores were the response variable
and climate factors were the predictor variables.
The study variables and their descriptions for step
one DEA model and the step two-panel analysis are
presented in Table 1. In the DEA model, the input
and output variables estimate the TE of agricultural
production. In the panel data analysis, the
efficiency scores are the response variable, while
the five climate change variables are the predictor
variables. The goal is to assess the influence of the
climatic factors on TE of the agricultural
production in SSA.
Table 1: Data Envelopment Analysis and panel regression models variables
Variable Name
Description
Agricultural land (AL) – Input
variable
Percentage of total land that is arable, used for permanent crops, and used for
permanent pastures
Agriculture value-added (AVA)
- Output variable
Net output for the agriculture sector, forestry, hunting, cultivation of crops, fishing and
livestock production, after adding up all outputs and subtracting intermediate inputs
(Value added is outputs minus inputs)
Forest Area (FA)
Land under natural/planted trees (5 meters), whether productive or not
Greenhouse Gases (GHG)
Total including land-use change and forestry/agriculture
Population (POP)
People living in the country as defined by the national statistics office
Rainfall (RN)
Average annual rainfall observed in the country
Temperature (TP)
Average annual temperature observed in the country
J. Agric. Environ. Sci. Vol. 6 No. 1 (2021) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Journal of the College of Agriculture & Environmental Sciences, Bahir Dar University 12
2.2. Data envelopment analysis
The non-parametric DEA technique was applied to
estimate the efficiency scores of a DMU relative to
other DMU. The Charnes, Cooper and Rhodes
(CCR) model is the basic DEA technique with the
Constant Return to Scale (CRS), which assumes no
significant relationship between the scale of
operations and efficiency (Charnes, et al., 1978). A
modification of CRS by (Banker et al., 1984)
became the Banker, Charnes and Cooper (BCC)
model which accommodates the variable return to
scale (VRS). The TE entails overall TE estimated
by the CRS. In the DEA, an efficient frontier is
created that evaluates the efficiency of a DMU and
is designed to maximize the relative efficiency of
each DMU. The efficiency score is estimated as the
ratio of weighted outputs to weighted inputs for
each variable of every DMU in order to maximize
its efficiency score (Ntwiga, 2020; Abel and Bara,
2017). Weights were determined by solving the
following linear programming problem.
[1]
[2]
Where,
is the output for the
country at
year
with weight
is the input for the
country at
year with
weight
s and m are the number of countries for the output
and input variables respectively;
k is the number of years
is the efficiency score to be maximized
The maximal efficiency score is equal to 1 and the
lower values indicate relative inefficiency of
analyzed objects (Ntwiga, 2020). We apply the
output-oriented DEA model to estimate the
efficiency scores of agricultural land used to
produce agriculture value addition.
2.3. Panel data analysis
The panel regression model response variable is the
TE scores with the predictor variables being GHG,
FA, POP, RN and TP as explained in Table 1. The
panel data comprises of nine countries from SSA,
with 25 annual data points for five predictor
variables and one response variable. The equation
for the panel model is indicated below.
[3]
Where,
is the TE scores of country and time
and
represent the
forest cover, greenhouse gases, population, rainfall
and temperature in country at time , respectively
3. Results and Discussion
The results were analyzed in two steps. Efficiency
scores of two-dimensional variables, individual and
time period, were computed using DEA. Then, the
panel data of computed efficiency score regressed
on the explanatory variables to find the
determinants of TE. The descriptive statistics
provide the summary statistics for the variables in
Table 1. The diagnostic tests for the panel data
comprising nine countries, for 25 years with five
predictor variables were performed and the data did
not exhibit multi-collinearity but heteroscedasticity
and autocorrelation were observed. The panel AR
package and function (Panel Regression with AR
(1) Prais-Winsten correction and panel-corrected
standard errors) in R statistical software were used
to correct for heteroscedasticity and
autocorrelation.
Table 2 presents the summary statistics for the nine
countries based on the mean, standard deviation
and percentage change of the two DEA variables
(AVA and AL) and the five-panel regression model
variables (GHG, FA, POP, RN and TP). Nigeria
had the highest GHG production on average in the
25 years followed by Cameroon and Ethiopia. The
ratio of AVA to AL was an indicator for the TE
with Benin having a ratio of one-to-one, Botswana
had a ratio of one-to-fourteen and Ethiopia's ratio
was one-to-less than one. These ratios indicated the
efficiency levels with Ethiopia being more TE
based on the selected variables among the nine
countries in the sample. The average temperature
did not vary much across the nine countries during
these 25 years but major variations were observed
in mean rainfall amounts. The highest average
rainfall was observed in Cameroon, followed by
Ethiopia, Nigeria and Benin. Nigeria had the
highest population, followed by Ethiopia, Kenya
and Ghana. Cameroon had the highest forest area,
then Benin and Ghana with Mali having the lowest
forest area.
J. Agric. Environ. Sci. Vol. 6 No. 1 (2021) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Journal of the College of Agriculture & Environmental Sciences, Bahir Dar University 13
The variables summary for the 25 years among the
nine countries indicates a decrease in forest area,
rain and agriculture value addition with an increase
in population, greenhouse gases, agricultural land
and temperature. This is a paradox as population
increase requires more food that led to agricultural
land increase that increases agricultural production
and in the process of reducing forest cover. The
increase in greenhouse gases and temperature and
reduction of rainfall complicates the agricultural
production due to the compound nature of climate
change. Nsiah and Fayissa (2019) observed that the
SSA population is expected to increase by 22% to
1.5 billion by 2050, to exceed agricultural
production and declining agricultural land. This
study noted that population increase far exceeds the
growth in agricultural land which will further lead
to food security vulnerabilities in the short and long
term. An increase in temperature by 1.37% is
similar to observations by FAO (2018) that Africa
is becoming warmer compared to the rest of the
globe. On average, among the nine countries
between 1991 and 2015, forest cover had reduced
by 20.85%, rainfall by 17.49% and agricultural
value-added by 23.06% while the population,
greenhouse gases, agricultural land, temperature
increased by 90.81, 46.94, 14.93, 1.37%,
respectively which agrees with the sentiments from
Tokunaga et al. (2015).
Table 2: Summary statistics of variables in DEA and panel analysis
Variable
Statistic
BEN
BWA
BFA
CMR
ETH
GHA
KEN
MLI
NGA
Mean
FA
Mean
43.87
22.18
21.51
45.39
13.39
39.53
7.21
4.64
13.08
SD
3.70
1.61
1.51
3.43
0.89
0.92
0.49
0.48
3.31
%
-24.25
-21.17
-20.29
-21.91
-16.52
7.88
-4.22
-28.68
-58.44
-20.85
GHG
Mean
21.89
66.66
29.92
192.47
134.89
40.81
51.24
32.09
402.48
SD
1.84
29.48
4.68
21.07
25.95
13.70
41.02
6.00
30.21
%
26.59
117.6
59.73
52.13
70.96
46.74
-55.50
87.23
16.97
46.94
POP
Mean
7.65
1.81
13.00
16.88
73.42
20.81
34.74
12.43
134.75
SD
1.65
0.23
2.78
3.28
15.21
3.86
7.02
2.74
25.48
%
105.4
55.94
100.12
89.33
100.5
83.40
95.36
101.9
85.40
90.81
RN
Mean
87.68
31.56
66.28
131.06
67.30
95.97
55.96
26.55
95.29
SD
8.58
6.53
5.90
8.97
5.78
8.80
12.87
3.10
7.56
%
-25.50
-36.25
-8.21
-11.81
8.86
-22.93
-47.83
6.51
-20.22
-17.49
TP
Mean
27.87
22.27
28.68
24.91
23.13
27.60
25.20
28.87
27.25
SD
0.31
0.38
0.30
0.24
0.35
0.26
0.70
0.35
0.31
%
2.95
5.37
1.95
1.50
4.17
2.71
-9.95
1.89
1.71
1.37
AL
Mean
28.15
45.68
39.29
19.71
34.47
63.99
47.57
31.70
76.76
SD
4.31
0.12
3.82
0.52
5.51
4.86
0.65
2.44
2.91
%
64.47
0.39
26.7
6.56
-28.91
23.43
2.8
28.2
10.75
14.93
AVA
Mean
28.89
3.24
35.41
17.44
49.10
35.75
29.28
37.88
31.43
SD
4.18
0.93
2.72
3.29
7.21
7.76
2.79
2.89
7.17
%
-27.74
-49.37
8.07
-35.26
-36.05
-53.93
18.34
1.6
-33.2
-23.06
Note: Percent (%) was the change of the variable from the year 1991 to 2015
Forest area = FA, Greenhouse gases = GHG, Population = POP, Rain = RN, Temperature =TP, agricultural land
= AL, agricultural value-added = AVA
In table 3, Ethiopia had generally the highest
average efficiency scores followed by Mali and the
lowest efficiency scores were observed in
Botswana as indicated in Table 3. In the 25 years,
the efficiency scores in descending order of country
were Ethiopia (0.971), Mali (0.816), Benin (0.708),
Burkina Faso (0.617), Cameroon (0.598), Kenya
(0.420), Ghana (0.378), Nigeria (0.277) and
Botswana (0.047). The difference between
Ethiopian and Botswana TE was about 92.4%,
which is a wide margin for TE of agricultural
production of the two countries. TE of Kenya,
Ghana, Botswana and Nigeria were below 50%
during the segmented period. The major difference
was observed between the country with the lowest
and highest TE ranging between 4.7% and 97.1%.
Between 1991 and 2015, the overall change in TE
showed a decline. Highest negative change in TE
was observed in Ghana, followed by Benin and
J. Agric. Environ. Sci. Vol. 6 No. 1 (2021) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Journal of the College of Agriculture & Environmental Sciences, Bahir Dar University 14
Botswana while the highest positive change was
observed in Kenya and Ethiopia.
Generally, the technical efficiency of agricultural
production among the selected countries is
decreasing trend with an average of 3.6%. The
percentage change of efficiency between 1991 and
2015 ranged from -47.97% to 60.47%. Similar
sentiments were also noted by Ogundari (2014)
where TE in Africa has decreased drastically over
time.
Table 3: DEA Technical efficiency scores of Sub-Saharan countries in the last 25 years
Variable
Statistic
BEN
BWA
BFA
CMR
ETH
GHA
KEN
MLI
NGA
1991-2000
Mean
0.801
0.051
0.558
0.599
0.947
0.409
0.376
0.771
0.252
SD
0.138
0.009
0.057
0.122
0.113
0.042
0.037
0.073
0.030
Min
0.585
0.043
0.442
0.407
0.694
0.353
0.327
0.674
0.214
Max
1.000
0.068
0.634
0.844
1.000
0.482
0.443
0.905
0.297
2001-2010
Mean
0.664
0.043
0.671
0.604
1.000
0.402
0.424
0.790
0.333
SD
0.038
0.006
0.045
0.068
0.000
0.065
0.044
0.037
0.065
Min
0.587
0.033
0.595
0.509
1.000
0.330
0.363
0.743
0.248
Max
0.719
0.051
0.731
0.703
1.000
0.490
0.484
0.858
0.459
2011-2015
Mean
0.608
0.045
0.627
0.582
0.961
0.269
0.499
0.961
0.218
SD
0.035
0.003
0.017
0.033
0.058
0.023
0.035
0.049
0.009
Min
0.582
0.042
0.603
0.539
0.870
0.245
0.454
0.896
0.211
Max
0.666
0.049
0.639
0.626
1.000
0.306
0.552
1.000
0.233
1991-2015
Mean
0.708
0.047
0.617
0.598
0.971
0.378
0.420
0.816
0.277
SD
0.120
0.008
0.069
0.087
0.077
0.074
0.060
0.092
0.065
Min
0.582
0.033
0.442
0.407
0.694
0.245
0.327
0.674
0.211
Max
%
1.000
-38.76
0.068
-29.70
0.731
18.90
0.844
-15.31
1.000
25.40
0.490
-47.97
0.552
60.47
1.000
10.48
0.459
-15.92
Note: SD = Standard deviation, Min = minimum value, max = maximum value
Percent (%) is the change of the variable from the year 1991 to 2015
The results presented in Table 4 highlights only the
countries whose results are statistically significant.
Model 1 showed that a unit increase in temperature
and forest area increased the technical efficiency by
3% and 0.7%, respectively while a unit increase in
greenhouse gases decreased technical efficiency by
0.03%. The predictor variables explained 73.94%
variations in the efficiency. In model 2, a unit
increase in greenhouse gases decreased efficiency
by 0.054%, while a unit increase in temperature
and forest area increased efficiency by 3.1 and
0.84%, respectively. Controlling for the year
significantly increased the magnitude of
temperature and forest area and reduced the
magnitude of greenhouse gases in influencing
efficiency. When the year 1992 is compared to that
of 1991, technical efficiency was reduced by 4.5%.
The influencing efficiency of temperature change
was 3.01% in Model 1 while 3.11% in Model 2.
Similarly, the influencing efficiency of forest area
change was 0.663 % and 0.835% in Model 1 and 2,
respectively, in Table 4. The influencing efficiency
of greenhouse gases in model 1 was 0.034 while in
model 2 it was 0.054 as indicated in Table 4. The
predictive power of the model 1 (73.94%) had been
reduced to 66.22% in model 2 after controlling for
the year. The changes from year to year reduced the
predictive power of the model.
In model 3, the one-way effect of the country, there
is an increase in efficiency by 0.882% when forest
area increases by one unit. Compared to Benin,
technical efficiency of Botswana and Ghana has
been reduced by 43.56% and 32.8%, respectively.
On the other hand, technical efficiency of Ethiopia
and Mali has been increased by 44.28% and
49.96%, respectively with an overall R-squared of
89.68%. In model 4, the two-way effect of country
and year showed that a unit increase in forest area
significantly increased efficiency by 1.58%.
Controlling for the year showed significant changes
on the influence of predictor variables on technical
efficiency in agricultural production from the year
2002 to 2015.
J. Agric. Environ. Sci. Vol. 6 No. 1 (2021) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Journal of the College of Agriculture & Environmental Sciences, Bahir Dar University 15
AGRA (2014) observed that an increase in
greenhouse gases in the last 100 years, which has
worsened agricultural production in Africa due to
erratic weather patterns. The study found an
increase in the greenhouse gases and temperature
and they reduced and increased TE respectively in
the last 25 years. Muller et al. (2011) noted that
temperature and rainfall changes are the two major
determinants of agricultural production. On the
other hand, rainfall did not influence the technical
efficiency significantly although the rainfall
amount declined in the last 25 years. The technical
efficiency of agricultural production in selected
countries was decreasing in the last 25 years with
an average of 3.6%. Population increase far
exceeds the growth in agricultural land which will
further lead to food insecurity in the short and long
term.
J. Agric. Environ. Sci. Vol. 6 No. 1 (2021) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Journal of the College of Agriculture & Environmental Sciences, Bahir Dar University 16
Table 4: Panel data analysis results of five selected variables
Variables
Model 1
Model 2
Model 3
Model 4
Greenhouse gases
Rain
Temperature
Forest Area
Population
Factor (Year) 1992
Factor (Year) 1993
Factor (Year) 2002
Factor (Year) 2003
Factor (Year) 2004
Factor (Year) 2005
Factor (Year) 2010
Factor (Year) 2013
Factor (Year) 2014
Factor (Year) 2015
Factor (country) BWA
Factor (country) ETH
Factor (country) GHA
Factor (country) MLI
Constant term
R-squared
Wald statistic
Total Obs.
Significance codes: '*' 0.05, '**' 0.01, '***' 0.001
4. Conclusions and Recommendation
A significant decrease in technical efficiency of
agricultural production has been observed in
selected SSA countries with an average downward
trend of 23.06% for the last 25 years. Temperature
and forest cover had a significant and positive
influence on efficiency and greenhouse gases had a
significant and negative influence on efficiency.
Rainfall and population changes did not
J. Agric. Environ. Sci. Vol. 6 No. 1 (2021) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Journal of the College of Agriculture & Environmental Sciences, Bahir Dar University 17
significantly influence technical efficiency. In the
last 25 years, technical efficiency declined while
greenhouse gases, temperature and agricultural land
increased due to population pressure and climate
change. The increase in population and agricultural
land reduced forest coverage with climatic changes
influencing rainfall amount. The agricultural value
addition decreased during this period, an indication
that farmers are becoming less efficient in adding
value to agricultural production even in the face of
climatic risks and population increase. The paradox
observed was, the increase in population increased
greenhouse gases and agricultural land and reduced
forest cover that in turn reduced climatic mitigation
with an increase in temperature and reduction in
rainfall.
Therefore, there is need for concerted efforts to
increase agricultural value addition and adopt more
efficient agricultural practices. This will reduce
deforestation, have sustainable agricultural food
production for the increasing population and deal
with the adverse effects of climate change in SSA
countries.
Conflict of Interest
The author declares no conflict of interest.
References
Abdulai, S., Nkegbe, P.K., and Donkoh, S.A.
(2018). Assessing the technical efficiency of
maize production in northern Ghana; The data
envelopment analysis approach. Cogent Food
and Agriculture, 4:1,
DOI: 10.1080/23311932.2018.1512390
Abel, S., and Bara, A. (2017). Decomposition of
the technical efficiency: Pure technical and
scale efficiency of the financial system.
Economic Research Southern Africa ERSA
Working paper 683
Abera, K., Crespo, O., Seid, J., and Mequanent, F.
(2018). Simulating the impact of climate
change on maize production in Ethiopia, East
Africa. Environmental Systems Research, 7:4,
DOI: 10.1186/s40068-018-0107-z
Alliance for a Green Revolution in Africa (AGRA).
(2014). Africa agriculture status report:
Climate change and smallholder agriculture in
sub-Saharan Africa, Nairobi, Kenya
Banker, Charnes and Cooper (1984). Some model
for estimating technical and scale
inefficiencies in DEA. Management Science,
30: 1078-1092
Charnes, Cooper and Rhodes (1978). Measuring
the efficiency of DMUs. European Journal of
Operational Research, 2:115-139
Climate Watch Data (CWD). (2020).
https://www.climatewatchdata.org/ [Accessed
July 2020]
Economist Intelligence Unit. (2020). Global food
security index.
https://foodsecurityindex.eiu.com/Index
[Accessed July 2020]
Eniko, V., Imre, F., and Jozsef, F. (2018). Impacts
of climate on technical efficiency in the
Hungarian arable sector. Studies in
Agricultural Economics, 12: 41-46,
https://doi.org/10.7896/j.1729
FAO (2018). African regional overview of food
security and nutrition: Addressing the threat
from climate variability and extremes for food
security and nutrition. FAO, Accra, Ghana
FSP (2020). Data dashboard. Food Security Portal.
http://www.foodsecurityportal.org/ [Accessed
July 2020]
Gregory, P.J., Ingram, J.S.I., and Brklacich, M.
(2005). Climate change and food security.
Philosophical Transactions of the Royal
Society, 360: 2139-2148,
https://doi.org/10.1098/rstb.2005.1745
Islam, M.S., and Wong, A.T. (2017). Climate
change and food in/security: A critical nexus.
Environments, 4,
38, https://doi.org/10.3390/environments40200
38
Muller, C., Cramer, W., Hare, W.L., and Lotze-
Campen, H. (2011). Climate change risks for
African agriculture. Earth System Analysis,
Potsdam Institute for Climate Impact Research,
108(11): 4313-5, DOI:
10.1073/pnas.1015078108
Nana, T.J. (2019). Impact of climate change on
cereal production in Burkina Faso. Journal of
Agriculture and Environmental Sciences, 8(1):
14-24. DOI: 10.15640/jaes.v8n1a2
Ngango, J. and Kim, S.G. (2019). Assessment of
technical efficiency and its potential
determinants among small scale coffee farmers
in Rwanda. MDPI Agriculture, 9 (161): 1-12
Nsiah, C., and Fayissa, B. (2019). Trends in
agricultural production efficiency and their
implications on food security in Sub-Saharan
J. Agric. Environ. Sci. Vol. 6 No. 1 (2021) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Journal of the College of Agriculture & Environmental Sciences, Bahir Dar University 18
African countries. African Development
Review, 31 (1): 28-42,
https://doi.org/10.1111/1467-8268.12361
Ntwiga, D.B. (2020). Technical efficiency in the
Kenyan banking sector: Influence of Fintech
and banks collaborations. Journal of Finance
and Economics, 8(1): 13-20. DOI:
10.12691/JFE-8-1-3
Ogundari, K. (2014). The paradigm of agricultural
efficiency and its implication on food security
in Africa: What does meta-analysis reveal?
World Development, 64: 690-702,
DOI: 10.1016/j.worlddev.2014.07.005
Oyetunde-Usman, Z., and Olagunju, K.O. (2017).
Determinants of food security and technical
efficiency among agricultural households in
Nigeria. Economics, 7, 103,
https://doi.org/10.3390/economies7040103
Popp, J., Olah, J., Kiss, A., and Lakner, Z. (2019).
Food security perspectives in Sub-Saharan
Africa. Amfiteatru Economic, 21(51):361-376,
DOI: 10.24818/EA/2019/51/361
Ray, D.K., West, P.C., Clark, M., Gerberl, J.S.,
Prishchepov, A.V., and Chatterjee, S. (2019).
Climate change has likely already affected
global food production. PLoS ONE 14(5):
e0217148. https://
doi.org/10.1371/journal.pone.0217148
Tokunaga, S., Okiyama, M., and Ikegawa, M.
(2015). Dynamic panel data analysis of the
impacts of climate change on agricultural
production in Japan. JARQ, 49(2):149-157,
https://doi.org/10.6090/jarq.49.149
World Bank. (2020). World Bank Data.
https://data.worldbank.org/ [Accessed July
2020]
