ArticlePDF Available

Impact of Climate Change on Cocoa Yield in Ghana Using Vector Autoregressive Model

Authors:

Figures

Content may be subject to copyright.
32
G
GJ
JT
T
Vol. 1, No. 2, March, 2017
Impact of Climate Change on Cocoa Yield in Ghana Using Vector
Autoregressive Model*
1E. N. Wiah, and 2S.Twumasi-Ankrah
1University of Mines and Technology, P.O. Box 237, Tarkwa, Ghana
2Kwame Nkrumah University of Science and Technology, Ghana
Wiah, E. N., and Twumasi-Ankrah, S. (2017), “Impact of Climate Change on Cocoa Yield in Ghana Using
Vector Autoregressive Model”, Ghana Journal of Technology, Vol. 1, No. 2, pp. 32 - 39.
This study examined the impact of four main climate variables (maximum temperature, minimum temperature, precipitation
and number of rainy days) on cocoa yield in Ghana. The Vector Autoregression (VAR) model with the use of Granger
Causality test, impulse response functions and variance decomposition for the data was used to examine the dynamic impact of
climate change on cocoa yields. It was observed that the maximum temperature was the climatic variable which had the highest
number of significant cross correlation (lags) on yield followed by the minimum temperature. The VAR(2) model was selected
as the adequate model among competing models to describe the association between yield of cocoa and climatic factors. The R2
value indicated that 48.5% of the total variation in the yield of cocoa can be explained by the climate variables considered in
this study. Granger causality test indicated that the direction of causality is from maximum temperature (maxt), minimum
temperature (mint) and precipitation (PRE) to yield since their corresponding F-statistic are significant. However, there was no
causation from the number of rainy days to yield, since the F-statistic is statistically insignificant. Maximum temperature (by
significant lag), number of rainy days have negative effects on yield, whereas minimum temperature and precipitation affect
yield positively. It is therefore recommended that, the government should develop adaptation strategies that would fit into
climatic condition and the agriculture sector should develop new seedling of cocoa that will have resistance to higher
temperature, low precipitation in order to maintain high production of cocoa beans yield in Ghana.
Keywords:Vector Autoregression, Precipitation, Impulse Response, Granger Causality Test
1 Introduction
Climate is an important factor of agricultural
productivity. Climate change is caused by the
release of ‘greenhouse’ gases into the atmosphere.
These gases accumulate in the atmosphere, which
result in global warming. The related factors which
cause changes in global climate such as
temperature, precipitation and soil moisture, block
the transmission of heat level.
The agriculture sector is mostly affected by the
changing of climate (Cumhur and Malcolm, 2008).
Cocoa is produced in countries in a belt between
10ºN and 10ºS of the Equator, where the climate is
appropriate for growing cocoa trees. The natural
habitat of the cocoa tree is in the lower storey of the
evergreen rainforest, and climatic factors,
particularly temperature and rainfall, are important
in encouraging optimum growth. Cocoa plants
respond well to relatively high temperatures, with a
maximum annual average of 30 - 32ºC and a
minimum average of 18 - 21ºC. Variations in the
yield of cocoa trees from year to year are affected
more by rainfall than by any other climatic factor
(Anon. 2016). Therefore, rainfall should be
plentiful and well distributed through the year.
Cocoa is one of the most important tree crops of the
humid tropical regions. The average world’s annual
production of cocoa, from 2012 to 2015, is 3129
thousand tonnes, of which 72% comes from Africa
(ICCO, 2015). In the 2014/2015 cocoa season,
Ghana was ranked second in terms of the
contribution to the world cocoa production (i.e.,
17.6%). In Ghana, the cocoa sector is the second
largest source of export earnings accounting for
approximately 30% of Ghana’s total export
earnings (ICCO, 2015). The annual cocoa
production in Ghana declined from 897 thousand
tonnes in 2013/14 to 740 thousand tonnes in
2014/15 cocoa season representing a decrease of
17.5% (ICC0, 2015). Many factors have been
assigned to the declining of cocoa production in
Ghana. For instance Jaeger (1999) attributed the fall
in cocoa output in Ghana to poor management of
the Cocoa Marketing Board.
However, one of the factors that has received a lot
of concerns from both researchers and policy
makers is climate change. The influences of climate
change on agriculture have been well researched in
the developing countries (Farook and Kannan,
2015). The results of these studies revealed that the
crop yield is more susceptible to climate change in
developing countries. The effect of climate change
on agriculture will differ with respect to countries.
In Ghana, the results from few studies have
emphasized on the alarming effect of climate
change on cocoa production.
Anim-Kwapong and Frimpong (2008) estimated the
impact of climate changes on the supply of dry
*Manuscript received August 23, 2016
Revised version accepted March 10, 2017
33
G
GJ
JT
T
Vol. 1, No. 2, March, 2017
cocoa beans. Their work sought to determine the
effect of changes in total annual rainfall, total
rainfall in the two driest months and sunshine
duration. They used multiple regression analysis to
show that over 60% of variation in dry cocoa beans
could be explained by the combination of the
preceding total annual rainfall, total rainfall in the
two driest months and the total sunshine duration.
Again, Ali (1969) reported both positive and
negative correlations between rainfalls in certain
months with the mean of yield crop in Ghana.
According to Brew (1991), a year with high rainfall
is followed by a year with larger crop output,
though the correlations not applicable in all years in
Ghana.
This is an indication of limited empirical
investigations of the impact of climate change on
cocoa yield in Ghana. However, it was observed
that the stationary properties of data collected over
long period of time have not been determined in
those studies. Thus, a study, on the effect of climate
change on cocoa yield, which will tackle the
stationary problem of time series data is desirable.
This study seeks to examine the impact of climate
change on cocoa yield in Ghana using vector
autoregressive (VAR) model.
2 Resources and Methods Used
2.1 Data Source
The average annual maximum temperature, average
annual minimum temperature, precipitation and the
number of rainy days for the period 1961-2013
were obtained from the Climate Research Unit
(CRU) while the yield of cocoa for the same period
was obtained from the Food and Agriculture
Organization website.
2.2 Stationary Test
The stationarity in this study was tested using unit
root test with Augmented Dickey-Fuller (ADF)
method. The ADF test has the following
mathematical equation (Joshua, 2007).
( )
0 1 1
21
p
t t i t i t
i
Y Y Yb q f e
- - +
=
D = + + D +
å
In which is time series value at the
t
time
minus time series value in 1 previous measurement
period (the time), θis constant-valued
(Joshua, 2007) which is used to
determine whether or not the unit roots exist with
hypothesis (the data contain unit roots)
and (the data do not contain unit roots).
Meanwhile, is trend coefficient on the time
series data of which the value is equal to
(Joshua, 2007).
2.3 Cross Correlation
In modeling and describing the relationship
between two time series
t
y
and
t
x
, the series
t
y
may be related to past lags of the
-series.
The sample cross correlation function (CCF) is
helpful for identifying lags of the
-variable that
might be useful predictors of
t
y
.
In R, the sample CCF is defined as the set of
sample correlations between
t h
x
and
t
y
for
0h
,
±1, ±2, ±3, and so on. A negative value for his a
correlation between the
-variable at a time
before tand the
y
- variable at time t(Anon. 2016).
2.4 Lag Length and Model Selection
Prior to forming the VAR model was completely,
the accuracy level should be evaluated by
calculating its lag value, which is generally
indicated by its p-value. In the conducted study, in
order to assess the feasibility level of the climatic
change on cocoa beans yield, it required applying
Aikake’s Information Criterion (AIC) and others
calculation for some kindependent variables where
the AIC value is generally defined using the
following mathematical equation (Schumway et.
al., 2011).
( )
22
log 2
k
n k
AIC n
s+
= +
where with .
In which
Yi
is observed value at the itime; kis the
number of parameters in the model; is mean; and n
is the number of observation times.
Yr
is mean; and
nis the number of observation times. In this case, it
can be stated that in case the AIC calculation value
is smaller, the taken lag value is the better lag value
as well as can be used as forecasting basis
(Gujarati, 2006; Schumway et. al., 2011).
2.5 The VAR Model
In general, a time series model of a random variable
Yi
is dependent on its past values and the present
and past values of its disturbance error term
 
t
34
G
GJ
JT
T
Vol. 1, No. 2, March, 2017
(Griffiths et. al., 1993). An unrestricted form of
VAR also known as the reduced form of VAR can
be expressed as:
 
1 1 2 2 3
t t t m m t t
Y AY A Y A Y
 
 
where , is an matrix and
is the error term of the vector.
In terms of the use of VAR model, it required
applying stationary condition criteria, which is
defined as a condition where its mean and variance
are constant and the covariance is not time-
dependent (Okky et. al., 2012). VAR model are
assumed stationary, meaning that they exhibit
constant mean
 
t
E Y
over time and covariance
matrix between vector and depend on only
0, 1, 2,k
and not on . Thus, after determining
the appropriate (p)lag length by an information
criteria, a
 
VAR p
is fitted.
2.6 Cranger Causality Test
An important tool in VAR model is to perform the
Granger causality testing, which examines the
direction of causality among the variables (Granger,
1969). It is a technique for determining whether one
time series is useful in forecasting another. If a
variable say, Xis found to be helpful for predicting
another variable say, Y, then Xis said to Granger-
cause Y. To test the null hypothesis that Xdoes not
Granger-cause Y, the test statistic is given by
 
( ) / 4
/ ( )
R UR
UR
RSS RSS m
FRSS n k
where RSSR, restricted residual sum of squares,
RSSUR, unrestricted residual sum of squares, m,
number of lagged X terms, and k, number of
parameters estimated in the unrestricted regression.
The test statistic follows the F-distribution with m
and
 
n k
degrees of freedom.
2.7 Impulse Response Functions and
Variance Decompositions
Now, after an adequate VAR model is obtained, the
parameters are not interpreted as in the case of the
regression model. The two important tools used to
interpret the parameters of VAR model are the
Impulse Response Functions and Variance
Decompositions.
The Impulse response functions provided by VAR
models are used to know where the impact of
change in one variable can be found through all the
other variables. They exhibit the current and lagged
effects over time of changes in error terms
 
1 2
, ,...,
t t kt
 
on the endogenous variables
 
1 , 2 ,...,
t t kt
y y y
. When the VAR process of order p
is stable, the error term
 
1t
has immediate effects
and
 
1 2
, ,...,
t t kt
 
all have lagged effects on y1t
If any covariance stationary VAR (p) process has a
Wald representation of the form
 
1 1 2 2 ... 5
t t t t
Y 
  
 
where
s
are
 
n n
moving average matrices,
the impulse responses,
s
ij
,
 
,th
i j
element of
s
are defined by
 
, ,
, ,
, , 1,2, ..., 6
i t s i t s
ij
j t j t s
y y i j n
 
 
 
 
It is only possible if
 
var t
 
is a diagonal
matrix in which are uncorrelated.
The variance decomposition analysis is typically
performed by VAR models, which supplements
impulse response function analysis (Granger, 1969).
It shows how much the variance of the forecast
errors of each variable can be explained by
exogenous shocks to the other variables in the
VAR.
3 Results and Discussions
The plots in Fig. 1 indicate that yield of cocoa,
maximum temperature, minimum temperature are
depicting and upward trend. However, precipitation
and the number of rainy days are not showing any
obvious trend. Therefore, trend analysis was
conducted on each of the series by regressing each
series on year (time). It confirmed that yield of
cocoa, maximum temperature and minimum
temperature have positive trend whereas
precipitation and number of rainy days did not
show any form of trend.
The ADF test of stationarity resulted that the
original series of yield of cocoa, maximum
temperature, minimum temperature and the number
of rainy days were not stationary [i.e. I(1)] while
precipitation was stationary [i.e. I(0)]. Detailed
results are presented in Table 1.
Aj
Yt
t
35
G
GJ
JT
T
Vol. 1, No. 2, March, 2017
Fig. 1 Annual Variation in the Yield of Cocoa, Maximum Temperature, Minimum Temperature,
Precipitation and no. of Rainy days from 1961 to 2013
Table 1 Test of Unit Root Test Hypothesis with Intercept and Trend Using ADF Test
Variables
ADF Test
Conclusions
Test Statistic
P-value
Max. Temperature
Level
-2.43
0.39
Not stationary at level
First Difference
-4.39
0.01
Stationary at first difference
Min. Temperature
Level
-2.37
0.42
Not stationary at level
First Difference
-4.27
0.01
Stationary at first difference
Precipitation
Level
-4.03
0.01
Stationary at level
No. Rainy Days
Level
-2.97
0.18
Not stationary at level
First Difference
-5.46
0.01
Stationary at first difference
Yield
Level
-1.96
0.58
Not stationary at level
First Difference
-5.77
0.01
Stationary at first difference
3.1 Cross Correlation
In the analysis of relationship between two time
series, the dependent series may be related to past
lags of the independent series. Thus, the number of
significant cross correlations (significant lags) in
each cross correlogram was counted to identify
climatic variables which were highly influenced on
cocoa yield during the period of 40 years. For
example maximum temperature was the climatic
variable which showed highest number of
significant cross correlation followed by minimum
temperature. The corresponding results are
presented in Table 2.
Table 2 Number of Significant Cross Correlation
Climatic
Variable
Number of
significant lags
Max. Temperature
6
Min. Temperature
5
Precipitation
3
No. Rainy Days
3
3.2 Lag Length and Model Selection
Firstly, we use the information criteria in Table 3 to
select the order of the potential “best” model for our
data. The lag length with the minimum information
criterion is selected as the “best” model. According
to Table 3, the three information criteria are
selecting different lag length as the “best” order for
the model. For example, AIC selects model with lag
length 7 as the “best” order, SIC selects order 1 and
HQ selects order 2. Since the original series of
precipitation is stationary, the other series were
made stationary after the first difference, a VAR
model was fitted with the three lag length.
Therefore the three competing models [i.e. VAR(1),
VAR(2) and VAR(7)] were fitted to our data.
Table 3 Lag Order Selection Criteria
Lag
LogL
AIC
SC
HQ
0
-602.4
26.9
27.2
27.0
1
-551.2
25.8
27.0*
26.2
2
-516.5
25.4
27.6
26.2*
3
-507.6
26.1
29.3
27.3
4
-488.1
26.3
30.5
27.9
5
-463.3
26.3
31.5
28.3
6
-437.2
26.3
32.5
28.6
7
-352.7
23.6*
30.9
26.3
* indicates lag order selected by the criterion
36
G
GJ
JT
T
Vol. 1, No. 2, March, 2017
However, it is needed to test the goodness of fit of
these models, because minimum lag length may be
due to less number of parameters and not due to
better fit or both. Models VAR(1) and VAR(7)
violated the white noise tests. The model diagnostic
tests indicated that VAR(2) residuals are
uncorrelated, normally distributed. Therefore
VAR(2) model is selected as an adequate model to
describe the association between yield of cocoa and
climatic factors.
Thus, for cocoa yield, the following VAR(2) model
is employed by considering cocoa yield and climate
variables
 
0 1 1 2 2
3 1 3 2 4 1
5 2 6 1 7 2
8 1 9 2
yield yield yield
maxt max t mint
mint pre p re
raday rada y 7
t t t
t t t
t t t
t t it
 
  
 
 
     
     
     
     
where yield is the cocoa yield (in kg per hectare),
maxt is the average maximum temperature (ºC),
mint is the average minimum temperature (ºC), pre
is the precipitation, raday is the number of rainy
days,
 
1t
is the error term and tis the time (i.e.,
year).
From Table 4, we are interested in column 2 (i.e.
YIELD), the effects of the climate variables on
cocoa yield is obtained and that the overall model is
statistically significant. The R2value indicates that
48.5% of the total variation in the yield of cocoa is
explained by climate variables. The significant
value of the F-statistic (2.44, p-value < 0.04) makes
all the lagged terms collectively statistically
significant.
Table 4 Estimated VAR(2) of Cocoa Yield
Vector Autoregression Estimates
Standard errors in ( ) and t-statistics in [ ]
YIELD
MINT
PRE
RADAY
MAXT
YIELD(-1)
-0.149081
(0.15967)
[-0.93370]
-4.36E-05
(7.5E-05)
[-0.58039]
0.000418
(0.00440)
[ 0.09508]
0.000189
(0.00039)
[ 0.49160]
-0.000106
(0.00013)
[-0.79079]
YIELD(-2)
0.068463
(0.15098)
[ 0.45345]
5.75E-05
(7.1E-05)
[ 0.80897]
0.003538
(0.00416)
[ 0.85120]
-0.000697
(0.00036)
[-1.91166]
-0.000140
(0.00013)
[-1.10181]
MINT(-1)
204.8956
(302.056)
[ 0.67834]
-0.116281
(0.14217)
[-0.81793]
3.847262
(8.31554)
[ 0.46266]
0.697323
(0.72895)
[ 0.95661]
-0.037044
(0.25381)
[-0.14595]
MINT(-2)
544.6557
(216.232)
[ 2.51884]
-0.260089
(0.10177)
[-2.55561]
1.199980
(5.95284)
[ 0.20158]
-0.363138
(0.52183)
[-0.69589]
-0.314469
(0.18170)
[-1.73072]
PRE(-1)
15.34327
(6.86953)
[ 2.23353]
-0.000535
(0.00323)
[-0.16534]
-0.268863
(0.18912)
[-1.42168]
-0.045850
(0.01658)
[-2.76571]
0.001056
(0.00577)
[ 0.18295]
PRE(-2)
-8.879805
(7.26487)
[-1.22229]
-0.001924
(0.00342)
[-0.56268]
0.273825
(0.20000)
[ 1.36912]
-0.030217
(0.01753)
[-1.72352]
-0.006100
(0.00610)
[-0.99923]
RADAY(-1)
-83.76321
(77.0594)
[-1.08700]
0.000987
(0.03627)
[ 0.02720]
4.078518
(2.12143)
[ 1.92253]
-0.301176
(0.18597)
[-1.61951]
-0.013786
(0.06475)
[-0.21291]
RADAY(-2)
-44.68136
(54.6922)
[-0.81696]
0.016132
(0.02574)
[ 0.62668]
0.088389
(1.50567)
[ 0.05870]
-0.084714
(0.13199)
[-0.64183]
0.081867
(0.04596)
[ 1.78136]
MAXT(-1)
-206.4889
(172.660)
[-1.19593]
0.579761
(0.08126)
[ 7.13429]
-8.676424
(4.75330)
[-1.82535]
-0.627283
(0.41668)
[-1.50543]
-0.729061
(0.14508)
[-5.02507]
MAXT(-2)
50.89773
(257.324)
[ 0.19780]
0.088581
(0.12111)
[ 0.73140]
-5.687434
(7.08407)
[-0.80285]
-0.374713
(0.62100)
[-0.60341]
-0.489644
(0.21623)
[-2.26449]
C
-581.5549
(752.431)
[-0.77290]
0.251023
(0.35414)
[ 0.70883]
96.20566
(20.7143)
[ 4.64441]
7.363788
(1.81584)
[ 4.05531]
0.544866
(0.63226)
[ 0.86177]
R-squared
0.485238
0.701739
0.184814
0.555497
0.467666
F-statistic
2.443922
9.175781
0.884182
4.873840
3.426231
Log likelihood
-365.4800
17.58878
-185.8557
-64.14189
-11.39191
Akaike AIC
15.05920
-0.263551
7.874229
3.005676
0.895676
Schwarz SC
15.47985
0.157094
8.294874
3.426321
1.316322
37
G
GJ
JT
T
Vol. 1, No. 2, March, 2017
3.3 Granger Causality
Granger causality test results from Table 5 suggest
that the direction of causality is from maximum
temperature (maxt), minimum temperature (mint)
and precipitation (PRE) to yield since their
corresponding F-statistic are significant. However,
there is no causation from the number of rainy days
to yield, since the F-statistic is statistically
insignificant.
Table 5 Results of Granger Causality Test
Null Hypothesis:
F-Statistic
P-value
Conclusion
MINT does not Granger Cause YIELD
4.12
0.00
Reject Ho
MAXT does not Granger Cause YIELD
3.04
0.02
Reject Ho
PRE does not Granger Cause YIELD
0.87
0.07
Reject* Ho
RADAY does not Granger Cause YIELD
0.47
0.75
Do not reject Ho
*Reject at 10% level of significance
The interpretation of the impulse response graph is given in Table 6.
Table 6 Interpretations of Impulse Response function
Periods
(Years)
Response of
yield to shocks
in yield
Response of
yield to shocks
in Max.
Temperature
Response of
yield to shocks
in Min.
Temperature
Response of
yield to shocks
in precipitation
Response of yield
to shocks in
number of rainy
days
1st to 2nd
Decreasing but
positive
Decreasing but
negative
Increasing but
positive
Increasing but
positive
Decreasing but
negative
2nd to 3rd
Increasing but
negative
Increasing but
negative
Decreasing but
positive
Decreasing but
positive
Increasing but
negative
3rd to 4th
Decreasing but
positive
Decreasing but
positive
Constant but
negative
Increasing but
negative
Decreasing but
positive
4th to 5th
Decreasing but
negative
Decreasing but
negative
Increasing but
positive
Decreasing but
negative
Increasing but
negative
5th to 6th
Increasing but
negative
Increasing but
negative
Response start
dying out
Increasing but
negative
Decreasing but
positive
6th to 7th
Decreasing but
positive
Increasing but
positive
Constant but
positive
Increasing but
negative
7th to 10th
Response start
dying out
Response start
dying out
Response start
dying out
Response start
dying out
-400
-200
0
200
400
600
1 2 3 4 5 6 7 8 9 10
Res pons e of YI ELD to YI ELD
-400
-200
0
200
400
600
1 2 3 4 5 6 7 8 9 10
Res pons e of YI ELD to Maximu m Tem perat ure
-400
-200
0
200
400
600
1 2 3 4 5 6 7 8 9 10
Res pons e o f Y IE LD to Mini mum Temp eratu re
-400
-200
0
200
400
600
1 2 3 4 5 6 7 8 9 10
Res pons e of YI ELD to PREC IPI TATION
-400
-200
0
200
400
600
1 2 3 4 5 6 7 8 9 10
Res pons e of Y IE LD t o N o. R ainy D ay s
Res ponse to No nfactorizedOne S.D. Innovations ± 2 S.E.
Fig. 2 Impulse Response Graph for Yield, Maximum Temperature, Minimum Temperature, Precipitation
and Number of Rainy Days
38
G
GJ
JT
T
Vol. 1, No. 2, March, 2017
Table 7 gives the variance decomposition values of
cocoa yield. During the changes of yield, its effect on
itself is 100% in first period and then gradually
declines to 71.3% in the 10th period. 0%~5.4%,
0%~8.7%, 0%~3.5% and 0%~11.4% fluctuations in
cocoa yield can be explained by the volatility of mint,
pre, raday and maxt respectively.
Table 7 Forecast Error Variance Decompositions
of Cocoa Yield
Period
S.E.
YIEL
D
MIN
T
PRE
RAD
AY
MAX
T
1
409.4
100.0
0.0
0.0
0.0
0.0
2
445.4
90.2
0.4
5.1
1.7
2.4
3
481.3
79.6
5.2
6.9
2.3
5.7
4
492.9
75.9
5.1
8.7
2.9
7.2
5
506.5
72.2
5.3
8.4
3.4
10.5
6
509.9
72.3
5.3
8.4
3.4
10.4
7
515.0
71.6
5.2
8.3
3.4
11.3
8
515.5
71.5
5.2
8.3
3.5
11.3
9
516.2
71.4
5.2
8.3
3.5
11.4
10
516.5
71.3
5.3
8.4
3.5
11.4
4 Conclusions
This paper uses the VAR model to examine the
impact of climate change on yield of cocoa in
Ghana. The climatic variables considered in this
study are average annual maximum temperature,
average annual minimum temperature, precipitation
and number of rainy days. The stationary test
revealed that the original series of yield of cocoa,
maximum temperature, minimum temperature and
the number of rainy days were not stationary but
precipitation was stationary. However, all non-
stationary series became stationary after the first
difference. Also, we observed that maximum
temperature was the climatic variable which
showed highest number of significant cross
correlation (lags) followed by minimum
temperature.
The VAR (2) model was selected as the adequate
model among competing models to describe the
association between yield of cocoa and climatic
factors. The R2value indicated that 48.5% of the
total variation in the yield of cocoa can be
explained by the climate variables considered in
this study. Granger causality test indicated that the
direction of causality is from maximum temperature
(maxt), minimum temperature (mint) and
precipitation (PRE) to yield since their
corresponding F-statistic are significant. However,
there was no causation from the number of rainy
days to yield, since the F-statistic is statistically
insignificant. Maximum temperature (by significant
lag), number of rainy days have negative effects on
yield, whereas minimum temperature and
precipitation affect yield positively.
It is therefore recommended that, the government
should develop adaptation strategies that would fit
into climatic condition and the agriculture sector
should develop new seedling of cocoa that will
have resistance to higher temperature, low
precipitation in order to maintain high production of
cocoa beans yield in Ghana.
References
Ali, F.M. (1969). “Effects of rainfall on yield of
cocoa in Ghana”. Experimental Agricultural,
Vol. 5, pp. 209-213.
Anim-Kwapong G. J. and Frimpong E. B.
(2008).”Vulnerability of agriculture to climate
change impact of climate change on cocoa
production”. Cocoa Research Institute of
Ghana, pp. 5-12.
Anon. (2016). International Cocoa Organisation
(ICCO). www.icco.org, Accessed: 22nd
February, 2016.
Brew, K.M. (1991), “Relationship between yield,
rainfall and total sunshine hours”. Rep. Cocoa
Research Institute of Ghana. 1988/89, pp. 30-
32.
Cumhur A. and Malcolm S. C. (2008). “The Effects
of Global Climate Change on Agriculture”.
American – Eurasian Journal of Agriculture and
Environmental Scienc, Vol. 3. No. 5, pp. 672-
676.
Farook, A. J. and Kannan K. S. (2015). “Climate
Change Impact on Rice Yield in India Vector
Autoregression Approach”. Sri Lanka Journal of
Applied Statistics. Vol. 16, No. 3, pp. 161-178.
Granger, C. W. J. (1969), “Investigating Causal
Relations by Econmetric Models and Cross-
spectral Methods”, Econometrica, Vol. 37, No.
3, pp. 424-438.
Griffiths, W. E., Hill, R. C., and Judge, G. G.,
(1993), Learning and practicing econometrics,
Wiley, New York, pp. 110-123.
Gujarati, D. N. (2006), Essential of Econometrics,
New York, McGraw-Hill Co., pp. 87-121.
Jaeger, P. (1999). The market for cocoa and its
relevance to African production. www.trecrops
.org/events/octconf/cocorep.pdf Accessed: 22nd
February, 2016.
Joshua, (2007), Analysis of vector autoregression
(VAR) to interrelationship of GDP growth and
employment opportunity growth (A case study
of Indonesia in 1977-2006), University Press,
pp. 34-45.
ICCO (2015). ICCO Quarterly Bulletin of Cocoa
Statistics. Vol. XLI, No. 4, pp. 5-12.
Okky, D. S. (2012), “Composite stock price index
modeling (IHSG), exchange rate and world oil
price modeling using vector autoregression
39
G
GJ
JT
T
Vol. 1, No. 2, March, 2017
approach” Journal of Science and Arts of ITS,
vol. 1, No. 1, pp. 54-69.
Schumway, R. H., Stoffer, D. S. (2011), Time
series analysis and its application, New York:
Springer Science+Business Media, Inc., pp.
218-235.
Authors
Eric Neebo Wiah is a Lecturer at the
Department of Mathematics, University of
Mines and Technology (UMaT), Tarkwa,
Ghana. He holds a Bachelor of Science
degree and Master of Philosophy degree in
Mathematics from the KNUST, Kumasi,
Ghana. He obtained his Doctor of
Philosophy from UMaT. His current
research interest focuses on applying
discrete and continuous element modeling of climatic patterns
and extreme events.
Sampson Twumasi-Ankrah is a Lecturer at
the Department of Mathematics, Kwame
Nkrumah University of Science and
Technology, Ghana. He holds a Ph.D in
Statistics from the University of Peradeniya,
Sri Lanka. His current research interest
focuses on model selection, diagnostics
analysis, medical statistics, agricultural
statistics, pharmaceutical statistics, extreme events and
econometrics.
... Non-governmental organizations (NGOs) have made significant efforts in developing sustainable practices related to cocoa production and climate change across the West African sub region and other developing nations in the cocoa belt. However, the development of cocoa varieties with tolerance for higher temperature and low precipitation is needed [8], particularly in Ghana, where strategic climate change ameliorating strategies are essential to sustaining cocoa production [9]. Current and emerging climatic trends could render smallholder cocoa farmers vulnerable and pose a significant threat to livelihoods centered on cocoa production [10]. ...
... Numerous studies have examined the perceptions of farmers on such topics as the impact of climate change on cocoa yields [8,9,20,21], smallholder choice of cocoa production systems [22][23][24], the potential benefits of cocoa agroforestry [25,26] and advantages of REDD+ in cocoa production [13,27]. There is however, limited information on how farmers perceive the inclusion of climate change mitigation strategies into their land/farm management objectives. ...
... Respondents suggest that plummeting cocoa yield in the area is a manifestation of climatic changes. These concerns have been raised previously by [3,[8][9][10]. The correct observation of increasing temperature trends and length of dry season in this study also brings into perspective that farmers accurately perceive weather patterns in relation to crop production and tend to amend their farming practices accordingly [34]. ...
Article
Full-text available
This study investigates the knowledge and perception of smallholder cocoa farmers on the potential impacts of climate change on cocoa production in Ghana. It addresses opinions on the inclusion of climate change mitigation strategies (such as Reducing Emissions from Deforestation and Forest Degradation—REDD+) into cocoa production, and potential obstacles and roles of stakeholders in ensuring community acceptance of such strategies in a unique multiple land use area—the Krokosua Hills Forest Reserve. Data from the Ghana Meteorological Agency and through survey of 205 cocoa farmers were assessed with Mann-Kendall, Kruskal Wallis and Mann-Whitney tests. Farmers’ perceptions of changes in climate were notably diverse and did not always match historic weather data, but accurately described increases in temperature and drought which are linked to cocoa productivity. Farmers appreciate the importance of tree maintenance for ecosystem services but were skeptical of financially rewarding climate change strategies which favor tree protection. Cultural practices associated with cocoa production encourage carbon release and may pose a threat to the objectives of REDD+. Farmers’ experience on the land, interactions with other farmers, government extension agents and cocoa buyers all influence cocoa agroforestry practices in the area, and communication through existing entities (particularly extension agents) presents a pathway to community acceptance of climate change mitigation strategies. The study recommends reforms in REDD+ strategies to adopt flexible and participatory frameworks to facilitate adoption and acceptability due to pronounced heterogeneity in community perceptions and knowledge of climate change and related issues.
... It is, therefore, indisputable that, climate changes and some socioeconomic factors interrelatedly constitute immensely to the yield of cocoa in such a manner that tackling them separately will not overcome the problem. Over the past decades, many researchers like [2,4,9,14,15] have devoted considerable resources to modelling cocoa production in Ghana. In literature, there are several econometric models that have been developed for the Ghanaian cocoa sector since the 1960's [1]. ...
Article
Full-text available
This study examined the combined effect of socio-economic and climatic variables on Ghana’s cocoa production. First, only multiple regression analysis was applied to all the variables affecting cocoa production which caused multicollinearity problems. In order to eliminate multicollinearity problems and perform a reliable regression, factor analysis was employed after its appropriateness on the data has been tested. Thus the problems were removed by using factor scores, summated scales and surrogate variables for the regression analysis. Also, the most significant determinants affecting cocoa production in Ghana was identified through which various parsimonious models were developed using regression models with ARIMA errors technique. The model parameters had least multicollinearity values that best describe and predict Ghana’s cocoa production. The parsimonious models were then compared in terms of prediction accuracy. The factor scores model (with an interaction term) was concluded to give better interpretation and good estimates of Ghana’s cocoa production as compared to the remaining models.
... This implies that, since Ghana's economy depends mostly on local cocoa production, it means the livelihood of the majority of the population will be greatly affected by climate change. A recent study concluded that changes in precipitation and temperature levels over the years (i.e., constitutes natural disaster and climate variability component in vulnerability assessments) has affected the quality of the cocoa yield in Ghana (Wiah and Twumasi-Ankrah 2017). Peprah (2015) was interested in the sustainability of livelihoods of cocoa farmers in Asunafo North district using both field-based and secondary data. ...
Article
Full-text available
Climate variability poses serious livelihood problems in most developing countries, especially in farming communities. This study assessed the vulnerability to climate change in two cocoa growing districts in Ghana. A total of 400 households from both districts were surveyed. Data were collected on socio-demographics, livelihoods, social networks, health, food, water, and natural disasters and climate variability. The composite index, differential, and integrated approach were used to aggregate the data, and differential vulnerabilities of the two districts were compared. The contributing factors (exposure, sensitivity, and adaptive capacity) were integrated to estimate livelihood vulnerability index (LVI)-IPCC. Results show that Asunafo North (ASN) had LVI–IPCC score of − 0.0236 compared to Bia West (BIW) of 0.0073. The results suggest that BIW may be more vulnerable regarding socio-demographics, social networks, health, food, water, natural disasters, and climate variability while ASN may be vulnerable to only livelihood strategies. The study also found that BIW was highly vulnerable to average receive–give and borrowed–lend money ratio. This assessment highlights how climate variability is affecting the livelihood of the cocoa-producing districts in Ghana. The study will be beneficial to the government of Ghana and non-governmental organizations in developing programs and projects to reduce vulnerabilities and enhance adaptive capacities in both ASN and BIW districts. Diversifying sources of income and livelihood could be the alternative to ameliorate vulnerability in both districts.
Article
Full-text available
Climate change plays an important role in agricultural production. Agricultural productivity is highly affected by a number of factors including precipitation, temperature. This paper examines the relationship between the yield of two major rice crops (e.g., Kharif and Rabi) and three main climate variables (e.g., maximum temperature, minimum temperature and rainfall). The dynamic relationships among the variables considered for observing the impact of climate change on rice yields are examined based on Vector Autoregression (VAR) model with the use of Granger causality test, impulse response functions and variance decomposition for the data. Maximum and Minimum temperature have significant effects, meanwhile rainfall has negative impact on Kharif rice yield. Adverse effects on Rabi rice yield are observed by maximum temperature and rainfall, whereas minimum temperature affect yield positively. Appropriate adaptive techniques are recommended to overcome this emerging hazard of climate change on rice production. Sri Lankan Journal of Applied Statistics 2015;16(3): 161-17
Article
Full-text available
2 Abstract: Climate is the primary important factor for agricultural production. Concerning the potential effects of climatic change on agriculture has motivated important change of research during the last decade. The research topics concentrate possible physical effects of climatic change on agriculture, such as changes in crop and livestock yields as well as the economic consequences of these potential yield changes. This study reviews the effects of climate change on agriculture. The main interests are findings concerning the role of huma n adaptations in responding to climate change, possible regional impacts to agricultural systems and potential changes in patterns of food production and prices.
Article
Effects of rainfall on yield of cocoa from a large-scale experiment has been investigated at 10 and 12 sites, over seven years, in the Eastern and Ashanti regions of Ghana. There was a positive association between yield and rainfall at certain times of year, but a negative association at other times. Variation due to the regression of yield on monthly rainfall varied up to 23 per cent, while a maximum variation of 31 per cent accounted for the regression of yield on the total rainfall of February to April in the Eastern region.
Vulnerability of agriculture to climate change -impact of climate change on cocoa production
  • G J Anim-Kwapong
  • E B Frimpong
Anim-Kwapong G. J. and Frimpong E. B. (2008)."Vulnerability of agriculture to climate change -impact of climate change on cocoa production". Cocoa Research Institute of Ghana, pp. 5-12.
International Cocoa Organisation (ICCO)
  • Anon
Anon. (2016). International Cocoa Organisation (ICCO). www.icco.org, Accessed: 22 nd February, 2016.
Relationship between yield, rainfall and total sunshine hours". Rep. Cocoa Research Institute of Ghana
  • K M Brew
Brew, K.M. (1991), "Relationship between yield, rainfall and total sunshine hours". Rep. Cocoa Research Institute of Ghana. 1988/89, pp. 30-32.
Learning and practicing econometrics
  • W E Griffiths
  • R C Hill
  • G G Judge
Griffiths, W. E., Hill, R. C., and Judge, G. G., (1993), Learning and practicing econometrics, Wiley, New York, pp. 110-123.
Essential of Econometrics
  • D N Gujarati
Gujarati, D. N. (2006), Essential of Econometrics, New York, McGraw-Hill Co., pp. 87-121.