J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 1
Reviving of Nigerian Rubber Industry through Export Potential Enhancement
Sadiq, M. S.
*1, 2
, Singh, I. P.
2
and Ahmad, M. M.
3
1
Department of Agricultural Economic & Extensions, FUD, Dutse, Nigeria
2
Department of Agricultural Economics, SKRAU, Bikaner, India
3
Department of Agricultural Economics, BUK, Kano, Nigeria
*Corresponding author: sadiqsanusi30@gmail.com
Received: March 20, 2020 Accepted: February 9, 2020
Abstract: The need to shore-up the continuous widening gap of the nation’s revenue owing to dwindling oil price
has forced the Nigerian government to start looking towards non-oil products especially agricultural cash crops for
foreign exchange earnings. Since agriculture was the mainstay of the Nigerian economy before the oil boom, the
need to look into this sub-sector to keep the nation’s economy afloat becomes a sine-qua-non. It is in view of this
thrust that the present research intends to chart a course towards examining the potential of reviving the rubber
sub-sector, being an important cash crop with high economic demand in the world. The research used dated data
that spanned from 1961 to 2017 and it covered production, area, yield and producer’s price (rubber). The data were
sourced from the FAO database and analyzed using both descriptive and inferential statistics. Empirical evidence
showed that the incremental change in the country’s rubber production in Nigeria was majorly driven by area
expansion which could be used for other purposes. Therefore, the future of the sub-sector is not promising to owe
the fact that the slight gentle rise in the forecasted production trend will be driven by a gentle incremental rise in the
annual production area while the level of productivity decreased year to year. The decrease in the forecasted annual
yield levels is as a result of non-productive income and not technology. This is because the farmers are at the mercy
of the Licensed Buyers (LBs) who exploit the producers through adopting collusive effects rather than to allow the
market forces to determine the prevailing market price. The Licensed Buyers also served as the major link to the
exporting markets. Therefore, the study recommends the establishment of farmer`s co-operative organizations so as
to venture into export marketing and to increase their bargaining power. Moreover, both governmental and non-
governmental organizations should facilitate viable export market linkages to these farmers’ co-operatives.
Furthermore, viable governmental policies should be framed to make the rubber market competitive for farmers,
middlemen, local and international industrial consumers in the value chain.
Keywords: Acreage response, Nigeria, Production forecast, Production instability, Rubber
This work is licensed under a Creative Commons Attribution 4.0 International License
1. Introduction
Today little is heard about foreign exchange earnings
from the three major cash crops viz. palm oil,
groundnuts and rubber which Nigeria was a leading
exporter (SENCE Agric, 2020). Even though rubber
plantation is one of the resources that thrive well in
Nigeria, policy inconsistency of successive
governments had neglected the sub-sector which used
to be the fourth-largest source of foreign exchange
earnings after crude oil, palm oil, and groundnut
(Umar et al., 2011; Ogbebor, 2013; Mesike and
Esekhade, 2014; Agbota, 2017). The mono-economy-
sole dependence on crude oil has rendered this
industry unattractive. In addition, the research
institutes mandated to develop improved seed
varieties have also gone comatose and even most of
the existing plantations are old as they were
established in the 1960s (Agbota, 2017). Besides, the
dearth of modern equipment necessary for value
addition of international standard has been attributed
to be responsible for the decline in rubber production
as argued by the stakeholders (SENCE Agric, 2020).
Due to the failure to replenish the old plantation and
establish new ones, Agbota (2017) reported that
experts have opined that the capacity of the country‟s
rubber industry has plummeted from well above
130,000 MT per annum to between the ranges of
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 2
55,000 and 60,000 MT. Thus, the country`s export
capacity has nosedived from about 100,000 MT per
annum to between 60,000 and 80,000 MT,
representing approximately a reduction by 40% to
20%, respectively (Ambrose, 2018). Furthermore,
Ambrose (2018) reported that the exit of the global
giants Michelin and Dunlop companies from Nigeria
was the highpoint of the decline in the rubber
industry. The rubber industry at its peak stage created
over 54 companies while currently; only less than 20
companies are working in the sector (Ambrose,
2018).
Rubber as a cash crop is at the verge of extinction
due to the neglect by the government despite being
one of the most important inputs in virtually every
industry. This caused the country colossal annual
revenue losses of approximately $6 billion in the
international market (Agbota, 2017). Rubber is
regarded as a money-spinner in some countries due to
its high demand for various purposes ranging from
engineering, aviation, education, sports, health etc.
The rate at which the automobile industry is mopping
up natural rubber around the world explains how
highly invaluable it is inhuman activities.
Dependency on oil may not sustain the economy of a
country where the price will continue to dampen
(Mesike and Esekhade, 2014; Pro-share, 2017). In
view of these, agriculture needs to be given desirable
attention to be able to sustain the economy. In this
regard, some countries are shifting from
hydrocarbon- eco-unfriendly to green energy-eco-
friendly approach (Agbota, 2017). Therefore, it is
very imperative for the country to embark on large-
scale cultivation of rubber to sustain the economy
when the oil fails.
It is on this thrust that the present research was
conceptualized with the aim of charting a pathway
for the revival of the sub-sector with a view of
exploring the economic potential of rubber
commodity for national development. Revitalization
of the neglected sector has the potential to drive the
Federal Government‟s on-going economic
diversification, job creation for the various
stakeholders along the value chain and to spur
industrialization. The specific objectives were to
determine the trend and growth patterns of rubber
production; determine the extent and sources of
instability in rubber production; determine the
sources of change in the average production;
determine the factors influencing farmers‟ acreage
response; and, to forecast the future production trend
of rubber in the studied area.
2. Materials and Methods
Nigeria lies between latitudes 4ʹ to14ʹ N and
longitudes 2ʹ to 15ʹ E of the Greenwich meridian time
(CIA, 2011). It has a vast area of land with suitable
prevailing agro-climatic conditions for the production
of various agricultural purposes viz. livestock,
fisheries, crop production etc. The country has
abundant untapped potential human and
environmental resources.
The study used time series data which ranged for 56
years (1961-2017) and it covered production, area,
yield and producer‟s price (rubber). The data were
sourced from the FAO database and analyzed using
both descriptive and inferential statistics. For better
understanding, the study was categorized based on
three economic reforms in the country viz. pre-
Structural Adjustment Period (pre-SAP) (1961-1984),
SAP (1985-1999) and post-SAP (2000-2017). The
trend and growth pattern were determined using
descriptive statistics and growth model while the
extent and source of instability in the production
were determined using instability indexes viz.
coefficient of variation (CV), Cuddy-Della Valle
index, Coppock‟s index and Hazell‟s variance
decomposition model. Changes in average production
and farmers‟ acreage response were determined using
Instantaneous change and Hazell‟s average
decomposition; and Nerlove‟s Adjustment models,
while the Autoregressive Integrated Moving Average
(ARIMA) model was used to forecast the future
rubber production trend.
2.1. Empirical Model
2.1.1. Growth rate
The compound annual growth rate was calculated
using the exponential model as indicated below:
[1]
[2]
[3]
Where,
CAGR is compound growth rate
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 3
t = time period in a year
= area/yield/production
= intercepting
= estimated parameter coefficient
2.1.2. Instability index
Coefficient of variation (CV), Cuddy-Della Valle
Index and Coppock‟s index were used to measure the
variability in the production, area and yield of rubber.
Following Sandeep et al. (2016) and Boyal et al.
(2015), the CV was calculated using the formula
below.
[4]
Where,
The simple CV overestimates the level of instability
in time series data characterized by long-term trends,
whereas the Cuddy-Della Valle Index corrects the
coefficient of variation by instability index as it de-
trend the annual production and show the exact
direction of the instability (Cuddy-Della Valle,
1978). Thus, it is a better measure to capture the
instability of agricultural production and prices using
the formula indicated below.
CDII = CV*(1-R
2
)
0.5
[5]
Where,
CDII = Cuddy-Della instability index
CV = Coefficient of variation
R
2
= coefficient of multiple determination
Following Shimla (2014) as adopted by Umar et al.
(2019), the instability index was classified as low
instability (20%), moderate instability (21-40%) and
high instability (>40%).
Unlike a CV, Coppock‟s instability index gives a
close approximation of the average year-to-year
percentage variation adjusted for trend (Ahmed and
Joshi, 2013; Kumar et al., 2017; Umar et al., 2019)
and it measures the instability in relation to the trend
in prices (Kumar et al., 2017). According to Kumar
et al. (2017), a higher numerical value for the index
represents greater instability. Following Coppock
(1962), the algebraic economic formula used by
Ahmed and Joshi (2013), Sandeep et al. (2016),
Kumar et al. (2017) and Umar et al. (2019) is
indicated below.
[6]
[7]
Where,
,
CII = Coppock‟s instability index
M = mean difference between the log of X
t+1
&
X
t
Log V = logarithm variance of the series
2.1.3. Source of change in rubber production
Instantaneous change: Following Sandeep et al.
(2016) the instantaneous decomposition analysis
model used to measure the relative contribution of
area and yield to the total output changes indicated
below.
P_0 = A_0×Y_0 [8]
P_n = A_n × Y_n [9]
Where
P, A and Y represent production, area and yield,
respectively
The subscript 0 and n represent the base and the
n
th
years, respectively.
P_n-P_0 = ∆P [10]
A_n-A_0 = ∆A [11]
Y_n-Y_0 = ∆Y [12]
From equation [5] and [9] we can write
P_0+∆P = (A_0+ ∆A)(Y_0+∆Y) [13]
Therefore,
P = (Y_0 ∆A)/∆P × 100 + (A_0 ∆Y)/ ∆P × 100+
∆A∆Y/∆P×100 [14]
Production = Area effect + Yield effect + Interaction
effect [15]
Hazell’s decomposition model: In estimating the
change in average production and change in the
variance of production between regimes and the
overall period, Hazell‟s (1982) decomposition model
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 4
was used. Hazell‟s decomposed the sources of
change in the average of production and change in
production variance into four (Table 1) and ten
(Table 2) components as cited by Umar et al. (2017,
2019). Decomposition analysis of change in
production assesses the quantum of increase or
otherwise of production in year „n‟ over the base year
that results from a change in the area, productivity or
interaction.
Changes in average production: It is caused by
changes in the covariance between area and yield and
changes in the mean area and mean yield. The model
is shown below.
[16]
[17]
Table 1: Components of change in the average production
Sources of change
Symbols
Components of change
Change in mean area
Change in mean yield
Interaction effect
Changes in area-yield covariance
Table 2: Components of change in variance production
Sources of change
Symbols
Components of change
Change in mean area
Change in mean yield
Change in area variance
Change in yield variance
Interaction effect I (changes in mean
area and mean yield)
Changes in area-yield covariance
Interaction effect II (changes in mean
area and yield variance)
Interaction effect II (changes in mean
yield and area variance)
Interaction effect IV (changes in mean
area and mean yield and changes in
area-yield covariance)
Residual
Change in variance decomposition: The source of
instability is caused by ten factors and the model is
shown below.
[18]
2.1.4. Nerlovian model
Following Sadiq et al. (2017), the basic model which
is called as Nerlovian price expectation model is
indicated below.
[19]
[20]
Where,
The Nerlovian model depicting farmer‟s behavior in
its simplest form is indicated below.
[21]
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 5
) [22]
As expected, variables are not observable, for
estimation purpose, a reduced form containing only
observable variables may be written after substituting
the value of
from equation (22) into equation (21),
and is indicated below.
[23]
Equation (21) is a behavioral equation that states the
desired acreage
depends upon the following
independent variables:
Where,
A
t
= current area under studied crop
P
t-1
= one year lagged price of studied crop
PR
t-1
= one year lagged price risk of studied crop
Y
t-1
= one year lagged yield of studied crop
YR
t-1
= one year lagged yield risk of studied crop
T
t
= time trend at time t
WI
t
= one year lagged weather index
A
t-1
= one year lagged area under studied crop
Price and yield risks were measured by the standard
deviation of the three preceding years. For the
weather index, the impact of weather on yield
variability was measured with a Stalling index
(Stalling, 1960). The yield was regressed on time to
obtain the expected yield. The actual to the predicted
yield ratio is defined as the weather variable. The
weather effects such as rainfall, temperature etc. may
be captured by this index in the acreage response
model (Ayalew, 2015).
The extent of adjustment to changes in the price
and/or non-price factors is measured in terms of the
“coefficient of adjustment”. The adjustment takes
place in accordance with the actual planted area in
the preceding year. If the coefficient of adjustment is
one, farmers fully adjust area under the crop in the
current year itself and there will be „no lags‟ in the
adjustment. But if the coefficient of adjustment is less
than one, the adjustment goes on and gives rise to
lags, which are distributed over time. The number of
years required for 95 percent of the effect of the price
to materialize is given below (Sadiq et al., 2017).
[24]
Where,
r = coefficient of adjustment (1-coefficient of a
lagged area)
n = number of year
In the present study, both short-run (SRE) and long-
run (LRE) elasticities of the area under the crop with
respect to price were estimated to examine and
compare the effect of price on the responsiveness of
area in the short-run as well as in the long-run. The
price elasticities were calculated using the formula
indicated below.
[25]
[26]
2.1.5. Autoregressive integrated moving average
(ARIMA)
Box and Jenkins (1976) posited that a non-seasonal
ARIMA model is denoted by ARIMA (p,d,q), which
is a combination of Auto-regressive (AR) and
Moving Average (MA) with an order of integration
or differencing (d). The p and q are the order of
autocorrelation and the moving average, respectively
(Gujarati et al., 2012).
The Auto-regressive of order p denoted as AR (p) is
indicated below:
[27]
Where
= constant
= the p-
th
autoregressive parameter
= the error term at time „t‟
The general Moving Average (MA) of order q or MA
(q) can be written as follow:
[28]
Where
=
constant
=
the q-
th
moving average parameter
=
-
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 6
ARIMA in general form is as follows:
[29]
Where
[30]
[31]
Here,
are values of past series with
lag 1,.., p, respectively.
Modeling using ARMA methodology consists of four
steps viz. model identification, model estimation,
diagnostic checking and forecasting.
2.1.6. Forecasting accuracy
For measuring the accuracy in fitted time series
model, mean absolute prediction error (MAPE),
relative mean square prediction error (RMSPE),
relative mean absolute prediction error (RMAPE)
(Paul, 2014), Theil‟s U statistic and R2 were
computed using the following formulae
[32]
[33]
[34]
[35]
[36]
Where
= coefficient of multiple determination
= Actual value
= Future value
T = time period
3. Results and Discussion
3.1. Trend and growth patterns of rubber
production
The trend pattern of rubber production during the
pre-SAP period was marked by a slight increase and
decrease throughout the studied period with yield
been the driving force for the rise as the change in the
area remains stagnant over a long period of time
(Figure 1 and 2). During the SAP period, the
production trend was marked by a gentle rise during
the early and late nineties with the incremental rise in
the area been the driving force as successive changes
in yield trend plummeted throughout the specified
study period (Figure 3). Furthermore, during the post-
SAP period, a slight rise was only visible during the
early twenties owing to slight incremental change in
successive annual yield level as the successive annual
area was stagnant. Thereafter, the production trend
from the mid-twenties till the expiration period was
marked by marginal rise due to small rise and fall in
both the successive annual yield and area which
interchange at different points in time (Figure 4).
Therefore, it can be inferred that rubber production in
the country was driven by incremental change in
yield during the pre-SAP period while successive
steep increase and a slight increase in the annual area
were the driving forces behind rubber production
during the SAP and post-SAP periods respectively.
The explosive effect of yield during the pre-SAP
period did not come as a surprise as the first national
plan was export-oriented with a focus on cash crops
as a source of raw material which was the driving
force of western industrialization. Also, for the post-
SAP period, it can be inferred that the sector was
almost redundant as the successive annual production
levels throughout the most part of the period
remained stagnant owing to poor yield even with the
slight rise in the area across the studied period.
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 7
Figure 1: Production trend of Rubber (1961-2017)
Figure 2: Pre-SAP production trend of Rubber (1961-1984)
Figure 3: SAP production trend of Rubber (1985-1999)
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
0
50000
100000
150000
200000
250000
300000
350000
400000
450000
Yield (hg/ha)
Production (ton) & Area (ha)
Production Area Yied
Linear (Production) Linear (Area) Linear (Yied)
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
0
20000
40000
60000
80000
100000
120000
140000
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
Yield (hg/ha)
Production (ton)/Area (ha)
Production Area Yied
Linear (Production) Linear (Area) Linear (Yied)
0
2000
4000
6000
8000
10000
0
100000
200000
300000
400000
500000
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
Yield (hg)
Production (ton)/Area (ha)
Production Area Yied
Linear (Production) Linear (Area) Linear (Yied)
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 8
Figure 4: Post-SAP production trend of Rubber (2000-2017)
A cursory review of the results showed the average
annual production of rubber to have increased by
almost two-fold between pre-SAP and SAP periods
and thereafter a small increase between the SAP and
post-SAP period (Table 3). The explosive increase in
area by almost three-fold was the major cause of the
galloping rise in rubber production between pre-SAP
and SAP regimes. Generally, the annual average area
was on the increase across the regime shifts while the
annual average yield was on the decrease across the
regime shifts.
Furthermore, a perusal of the growth pattern results
showed the annual growth rate of rubber production
to be on the trough during the pre-SAP period despite
that yield recorded a positive annual growth rate.
However, the negative annual growth rate which
makes the production to trough owes to negative
annual growth rate observed for the area. During the
SAP period, the annual production growth rate
increased by 4.6% with an instantaneous 10.6% rise
in the growth rate of area been the only driving force
as the yield was found to be on the trough due to the
negative growth rate. The trending behavior of yield
may be attributed to poor investment in technology
and infrastructure, which intern affects the
productivity of the crop. For the post-SAP period, the
rubber production witnessed a small incremental
change in the growth rate by 1.9% with the small
incremental rise in the annual growth rates of both
area and yield been the driving forces. Though, yield
effect (1.3%) was more pronounced in incremental
growth which marked rubber production during the
post-SAP era. However, for the overall period, the
incremental rise in the annual growth rate of
production was due to a positive annual growth rate
in the area as yield observed a negative growth rate.
Therefore, it can be concluded that rubber production
during the SAP period witnessed an impressive
growth due to an increase in the area rather than
technology, which is not a healthy growth (Table 3).
Table 3: Growth pattern of rubber production
Variables
Pre-SAP
SAP
Post-SAP
Overall
Area (ha)
95500 (-3.1)***
223566.7 (10.6)***
350347.9 (0.7)***
209680.1(3.2)***
Yield (hg)
6534.708 (2.0)***
5484.2 (-5.9)***
3997.667 (1.3)***
5457.088 (-1.1)***
Production (ton)
60560.38 (-1.1)***
110297.7 (4.6)**
140328.2 (1.9)***
98838.98 (2.1)***
Source: Authors‟ computation, 2019
Note: Values in parenthesis are CAGR
***, **, * = means significant at 1, 5 and 10%, respectively
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
0
50000
100000
150000
200000
250000
300000
350000
400000
Yield (hg)
Production (ton)/Area (ha)
Production Area Yied
Linear (Production) Linear (Area) Linear (Yied)
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 9
3.2. Instability in rubber production and source
of production risk
A perusal of the Table showed production instability
to be low and moderate during the pre-SAP and post-
SAP, and SAP regimes respectively as evident by the
CV indexes which were less than 20% for the former
and higher than 20% but less than 40% for the latter
(Table 4). While for the overall period, rubber
production was marked by high instability as evident
by the CV index which is greater than 40%.
However, the major reasons for low and moderate
fluctuation in rubber production during the pre-SAP
and SAP periods owed to low and moderate
instability in the yield levels respectively. For high
instability which marked the production for the
overall period, high fluctuation in the area was
observed to be the major cause. Therefore, it can be
inferred that despite the poor production performance
of rubber during the pre-SAP period, the extent of the
production instability was low, thus exogenous factor
may be the likely cause of negative production
growth rate recorded during the stipulated study
period.
Furthermore, a review of the exact direction of the
production instability showed a similar trend with
what was obtained when CV was applied except that
the production instability for the overall period turn-
out to be moderate as indicated by the CDII index
which was between 20 to 39%. Therefore, it can be
inferred that rubber production across the policy
regime periods in the country was within the comfort
zone from the exact directional point of view (Table
4).
The results of the CII index showed high fluctuation
in rubber production across the policy regime periods
as evident by the CII indexes which were higher than
40%. Evidence showed all the production parameters
viz. area and yield to be marked by explosive
fluctuation across all the policy periods under
investigation. Therefore, from the point of production
instability in relation to price trend, it can be inferred
that rubber production in the country was not in the
comfort zone across the policy regime periods (Table
4).
Furthermore, a critical investigation of the source of
instability across regime shifts showed „change in
area variance‟ to be the major source of fluctuation in
rubber production between pre-SAP and SAP
periods; while between SAP and post-SAP regimes,
„change in the average yield‟ was the major cause of
production instability. However, for the overall
period i.e. across the policy regime shifts, „interaction
between change in average area and yield‟ featured as
the major cause of instability in rubber production in
the country (Table 5). Therefore, it can be inferred
that the major cause of rubber production instability
centered majorly on risk.
Table 4: Extent of instability in rubber production
Regimes
Variables
CV
CDII
CII
Pre-SAP
Area
0.236
8.509101
46.25296
Yield
0.171
10.51337
43.83258
Production
0.144
12.31179
42.53909
SAP
Area
0.398
21.28463
64.3402
Yield
0.29
9.791527
48.79965
Production
0.29
23.07278
51.71077
Post-SAP
Area
0.038
1.354208
38.21286
Yield
0.088
6.172559
40.42428
Production
0.112
5.968665
41.53075
Overall
Area
0.569
32.88407
70.8594
Yield
0.282
21.66083
48.6214
Production
0.40198
21.90699
56.27885
Source: Authors‟ computation, 2019
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 10
Table 5: Sources of instability in rubber production
Source of variance
Pre-SAP to SAP
SAP to Post-SAP
Overall
Change in mean yield
-177.30
2137.30
206.26
Change in mean area
11.56
-47.45
9.72
Change in yield variance
80.54
659.44
137.05
Change in area variance
1547.00
1098.46
1405.53
Interaction between changes in mean yield and mean
area
30.49
-247.35
-45.75
Change in area yield covariance
-775.96
-1671.72
-749.69
Interaction between changes in mean area and yield
variance
360.86
959.98
104.48
Interaction between changes in mean yield and area
variance
-457.41
-514.79
-1332.79
Interaction between changes in mean area and yield
and change in area-yield covariance
-687.37
-229.01
1042.51
Change in residual
167.57
-2044.86
-677.33
Total change in variance of production
100
100
100
Source: Authors‟ computation, 2019
Table 6: Sources of change in rubber production (Intra-wise %)
Source of change
Pre-SAP
SAP
Post-SAP
Overall
Area effect
0
305.4656
31.84607
134.7209
Yield effect
0
-168.753
69.06265
-16.5962
Interaction effect
0
-36.6874
-0.8735
-18.1263
Total change
100
100
100
100
Source: Authors‟ computation, 2019
Table 7: Sources of change in rubber production (Inter-regime wise %)
Source of change
Pre-SAP to SAP
SAP to Post-SAP
Area effect
160.17
257.71
Yield effect
-19.20
-123.18
Interaction effect
-25.75
-69.85
Co-variance effect
-15.22
35.33
Total change
100
100
Source: Authors‟ computation, 2019
3.3. Source of change in rubber production
The results of the instantaneous source of an increase
in the average annual production level of rubber
across the policy regimes independently showed area
effect to be the major source of incremental change in
the average annual output level during the SAP and
the overall periods; while technology effect was
observed to be the production incremental driving
force during the post-SAP period. However, the
average annual rubber output level was dormant
during the pre-SAP period as evident by the
parameter estimates which were zero (Table 6).
Furthermore, for the inter-regime wise, evidence
showed an area effect to be the source of the
incremental rise in the output level of rubber between
the pre-SAP and SAP, and SAP and post-SAP. This
indicates that technology was the driving force
behind the increase in rubber production in the
country when the sub-sector was undergoing a
paradigm shift (Table 7).
3.4. Farmers’ acreage response
The results of the Autoregressive distributed lag
model showed the linear functional form to be the
best fit for the specified equation as the classical
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 11
linear regression model satisfied the economic,
statistical and econometric criteria, thus selected as
the best fit (Table 8). The auto-regression test at lag
four (4) showed the residuals to be independent as
indicated by the Langrage Multiplier (LM) test which
was not different from zero at 10% degree of
freedom, thus no serial correlation. The test of
heteroscedasticity using the robust variant test
showed the variance of the residual to be constant as
indicated by the LM test statistic which is not
different from zero at 10% degree of freedom, thus
implying that the sum of the square of the residual is
homoscedasticity. This clearly showed that the least-
squares are efficient and reliable for prediction. There
is no evidence of a correlation between the variance
of the residuals as indicated by the LM test statistic
which is not different from zero at 10% degree of
freedom, thus implying that the residual has no Arch
effect. Furthermore, the diagnostic test results
showed that the parameter estimates were stable and
the specification of the functional form is adequate as
indicated by the CUSUM (Figure 5) and RESET test
statistics respectively, which were not different from
zero at 10% degree of freedom, thus implying no
change in the parameters and the model specification
is adequate. However, the residual was found not to
be normally skewed as indicated by the Chi
2
test
statistic which is different from zero at 10% degree of
freedom.
Though, literature has shown that non-normality in
the distribution of residual is not considered a serious
challenge as data in their natural forms are likely not
to be normally distributed (Sadiq et al., 2017).
Therefore, with the above ample evidence, it can be
inferred that the least-squares are efficient and
reliable to predict farmers‟ acreage response with
high precision and certainty.
Figure 5: Parameter stability test
The value of the coefficient of multiple determination
being 0.984 implies that 98.4% of the current acreage
cultivated is been determined by the explanatory
variables captured in the model while the disturbing
economic reality accounts for the left-over
percentage. The high value of the R
2
will create a
suspicious of spurious correlation but with the
absence of a serial correlation as indicated by the LM
test, it can be affirmed that the variables did not move
together with the time trend. Thus, the model is
devoid of spurious correlation and is not a nonsense
regression. Thus, the model is reliable for long-run
prediction.
Furthermore, the variables found to have an impact
on the current rubber acreage cultivated are weather
vagaries, producer price, time factor and the lagged
area under rubber cultivation as evident by their
respective parameter estimates which were different
from zero at 10% degree of freedom. The negative
significant of the Weather Index (WI) implies that the
current acreage under rubber production decrease
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 12
owing to weather vagaries which are caused by
climate change. The decline in the productivity of
rubber has dire consequences on the turn-over of the
enterprise as it has the tendency of forcing the
farmers to diversify to other cash crops, thus
affecting the hectare cultivated for rubber in the
country. The consequence of climate change due to
indiscriminate human exploitation of the
environmental resources caused dry-spell and flood
with dare consequences on the production of rubber
in the country.
The negatively significant producer price result
showed how dampening of the price due to glut in the
supply which intern affected the business turn-over
led to a decrease in the current acreage cultivated for
rubber. This did not come as a surprise as the farmers
have no direct link with the exporting market and
have to rely on the licensed buyers (LBs) who
engaged in the exportation of the commodity. The
LBs take advantage of lack of market tie-up between
the producers and exporters to exploit the farmers
with respect to the price to their advantage. This
exploitation tendency of the middlemen i.e. LBs is
what is killing the cash crop sub-sector in Africa as
most of the farmers leased-in the plantations, thus
living them with very marginal turn-over which can
hardly sustain their family expenditure more or less
the going concern of the business (Adelodun, 2017).
In addition, the negative relationship has a link with
the non-availability of remunerative substitute crops
to cultivate, given that substitute crops require extra
capital for cultivation, well known that capital is the
major constraint affecting the farmers, they tend to
glue themselves to the cultivation of this crop. Thus,
they produce the crop irrespective of the prevailing
price in the market.
The SRE and LRE showed acreage responsiveness of
the rubber farmers to price changes in the preceding
crop period and the elasticities were 0.05 and 0.18%
respectively. Given that the LRE reflects the acreage
responsiveness of rubber crop to a price change if
given sufficient time for adjustment, thus the impact
of price policy on rubber production in the long-run
is small owing to the low LRE value. Furthermore, it
will take approximately 2.29 years for the price effect
to materialize. Since the required time for price effect
to adjust is small, it can be inferred that the price
policy instrument in bringing the desired change in
the supply of rubber production will be effective.
The positive significant of the time trend implies that
the different policy regimes witnessed by the
economy impacted positively on rubber production in
the country thus encouraged an increase in the
current acreage cultivated under the rubber. The
establishment of rubber research institutes to develop
technologies aimed at increasing rubber production,
the establishment of agencies with the mandate of
export promotion both at state and national level,
bilateral and multilateral agreement in trade
organization both at regional, continental and
international (WTO) levels, foreign investment in the
sub-sector, protection policies viz. embargo, tariff,
export promotion etc. are all visible policies aimed at
reviving the sub-sector in the country. Furthermore,
empirical evidence showed that the rate of adjustment
of the area under rubber was low as evident by the
calculated adjustment coefficient value of 0.27. Thus,
the adjustment in accordance with the actual planted
area in the preceding year will go on and give rise to
lags that are distributed over time. The positive
significant of the efficiency parameter indicated that
technology impacted positively on the current
acreage under rubber production in the country.
3.5. Production Forecast of Rubber in Nigeria
The unit root test results showed that the residuals of
variables viz. area, yield and production at the level
were pure white noise but after differencing (first
difference) the residuals became Gaussian white
noise as evident by their respective Augmented
Dickey-Fuller (ADF) test statistics which were not
different from zero and different from zero at 5%
error gap respectively (Table 9). This implies that the
variables after differencing became stationary which
is a pre-requisite for the efficiency of time series
data. Thereafter, for forecasting, the variables were
subjected to various ARIMA stages to determine the
most suitable ARIMA model with the lowest Akaike
information criterion (AIC). The empirical evidence
showed that for the area, yield and production;
ARIMA (1,1,0), ARIMA (0,1,1) and ARIMA (0,1,1)
respectively, were chosen as the best fit for the
forecast because they had the lowest AICs among all
the tried ARIMA forms (Table 9). In addition, the
residuals of the chosen ARIMAs were devoid of
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 13
autocorrelation, Arch effect and are normally
distributed as indicated by their respective test
statistics which were not different from zero at 10%
error gap. However, the residual of the selected
ARIMA for the production variable was found not to
be normally skewed as it is different from zero at
10% degree of freedom. Though, literature has shown
that non-normality in the distribution of the error
term is not a serious problem given that data in their
natural form in most cases are not normally
distributed.
Table 8. Farmers’ acreage response
Variables
Parameters
t-stat
Mean
SRE
LRE
Intercept
34928.9 (34928.9)
2.150**
-
-
-
P
t-1
-0.2468 (0.1355)
1.821*
41865.26
-0.04754
-0.17607
PR
t-1
0.066 (0.3195)
0.206
NS
5667.686
0.001721
0.006374
Y
t-1
-1.85 (4.39326)
0.421
NS
5387.075
-0.04585
-0.16983
YR
t-1
1.662 (9.364)
0.177
NS
419.4214
0.003207
0.011879
T
t
2449.38 (577.17)
4.244***
26
0.293013
1.085233
WI
t
-57703.4 (22012.0)
2.621**
1.003672
-0.26647
-0.98693
A
t-1
0.73 (0.0953)
7.655***
213240.8
0.716226
2.65269
R
2
0.9843
F-stat
385.95{1.17e-36}***
Autocorrelation
1.961{0.119}
NS
Arch effect
0.0466{0.828}
NS
Heteroscedasticity
7.932{0.338}
NS
Normality
21.11{2.6e-5}***
CUSUM test
1.502{0.140}
NS
RESET test
5.2778{0.8208}
NS
Source: Authors‟ computation, 2019
***, **, * = significant at 1%, 5%, 10% probabilities, respectively; NS = Non-significant
Values in ( ), [ ] and { } are standard error, t-statistic and probability level, respectively
Table 9. ARIMA model
ARIMA
Production (AIC)
Area (AIC)
Yield (AIC)
ARIMA (1,1,1)
1220.82
1258.15
894.406
ARIMA (1,1,0)
1218.84
1257.05
894.299
ARIMA (0,1,1)
1218.82
1257.69
892.839
Autocorrelation
2.959(0.3960)
NS
2.1319(0.545)
NS
2.922(0.495)
NS
Arch effect
2.488(0.477)
NS
8.0259(0.1548)
NS
14.786(0.1400)
NS
Normality
16.454(0.00026)***
28.888(5.33e-07)***
2.427(0.297)
NS
ADF
Level
-0.889(0.7846)
NS
-0.237(0.9313)
NS
-0.9223(0.7818)
NS
1
st
Diff.
-6.542(6.08e-7)***
-5.674(1.19e-05)***
-7.340(3.86e-011)***
Source: Authors‟ computation, 2019
Note: *** = significant at 1% probability; NS = Non-significant
Values in ( ), [ ] and { } are standard error, t-statistic and probability level, respectively
In determining the predictive power of the chosen
ARIMAs, one-step-ahead forecast of the variables
along with their corresponding standard errors using
the naïve approach for the periods 2013 to 2017 were
computed (Table 10). This was done to validate how
closely the sample periods could track the path of
actual observation.
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 14
Table 10. One-step ahead forecast of rubber production
Period
Production
Area
Yield
Actual
Forecast
Actual
Forecast
Actual
Forecast
2013
150110
150020.4
359859
359978.6
4171
4103.45
2014
152298
151932.6
362658
363721.1
4199
4145.14
2015
154571
154149.7
365622
366586.7
4228
4177.18
2016
156900
156428.6
368676
369590.4
4256
4207.08
2017
159264
158762.9
371775
372666.1
4284
4235.64
Source: Authors‟ computation, 2019
Table 11. Validation of models
Variable
R
2
RMSE
RMSPE
MAPE
RMAPE (%)
Theil‟s U
Production
0.997725
396.0568
1.003552
351.834
0.225509
0.193138
Area
0.997904
859.1978
2.013157
-766.67
-0.20894
0.322523
Yield
0.990446
45.2005
0.481984
40.392
0.952592
1.789165
Source: Authors‟ computation, 2019
In measuring the reliability of the selected ARIMAs
for the forecast, the mean absolute prediction error
(MAPE), root mean square error (RMSE), Theil‟s
inequality coefficient (U) and the relative mean
absolute prediction error (RMAPE) were determined.
Empirical evidence showed the RMAPE and U
coefficients to be less than 5% and 1 respectively,
indicating the predictive error associated with the
estimated equations in tracking the actual data (ex-
post prediction) to be very low and insignificant, thus
could be used for ex-ante projection with high
projection validity, efficiency and consistency (Table
11).
Table 12 and Figure 6-8 showed the estimated one-
step-ahead out of sample forecasts of rubber
production (ton), area (hectare) and yield (hg)
spanning through 2018 to 2029. It was observed that
the production will be marked by a gentle rise
throughout the forecasted period owing to a gentle
rise in the area. It is saddened to observe that the
yield will be marked by a gentle fall throughout the
forecasted period as evidenced by the plummeting
forecasted yield trend. Therefore, the study calls for
urgent intervention by both government and non-
governmental agencies in linking the producers with
the importing market in order to have a better price
for their products. In addition, the farmers should
form viable co-operative organizations that will be
directly involved in marketing, especially exportation
in order to have strong bargaining power for their
products, thus fetching them remunerative prices that
will make their income productive.
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 15
Table 12: Production, area and yield forecasts of rubber (2018-2029)
Year
Production
Area
Forecast
LCL
UCL
Forecast
LCL
UCL
2018
161130.03
137205.30
185054.76
375775.92
342132.33
409419.52
2019
162943.19
127278.74
198607.63
379993.87
326383.81
433603.92
2020
164756.34
120354.62
209158.07
384264.04
315107.77
453420.30
2021
166569.50
114887.07
218251.93
388546.77
306496.08
470597.45
2022
168382.65
110325.49
226439.81
392832.52
299601.56
486063.49
2023
170195.81
106397.73
233993.88
397119.01
293900.24
500337.78
2024
172008.96
102945.55
241072.37
401405.67
289081.16
513730.18
2025
173822.12
99867.30
247776.93
405692.37
284946.29
526438.44
2026
175635.27
97093.09
254177.46
409979.08
281361.55
538596.60
2027
177448.43
94572.40
260324.45
414265.79
278231.50
550300.09
2028
179261.58
92267.36
266255.81
418552.51
275485.42
561619.60
2029
181074.74
90148.64
272000.84
422839.22
273069.21
572609.23
Year
Yield
Forecast
LCL
UCL
2018
4263.80
2962.16
5565.45
2019
4257.87
2665.23
5850.50
2020
4251.93
2413.80
6090.06
2021
4246.00
2191.50
6300.49
2022
4240.06
1989.92
6490.21
2023
4234.13
1804.03
6664.23
2024
4228.19
1630.58
6825.81
2025
4222.26
1467.30
6977.22
2026
4216.32
1312.53
7120.12
2027
4210.39
1165.02
7255.76
2028
4204.46
1023.81
7385.10
2029
4198.52
888.13
7508.91
Source: Authors‟ computation, 2019
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 16
Figure 6: Production forecast of Rubber in Nigeria
Figure 7: Area forecast of Rubber in Nigeria
80000
100000
120000
140000
160000
180000
200000
220000
240000
260000
280000
2014 2016 2018 2020 2022 2024 2026 2028
Production (Ton)
Year
Production
forecast
95 percent interval
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Publication of College of Agriculture and Environmental Sciences, Bahir Dar University 17
Figure 8: Yield forecast of Rubber in Nigeria
4. Conclusion and Recommendations
From the findings, it can be inferred that the
incremental changes in rubber production between
the policy regime shifts in the country have been
driven by area effect which does not signify healthy
growth in the rubber sub-sector of the country.
Furthermore, area risk due to competing demand for
limited available land seriously affected the
production of rubber in Nigeria. The future of the
rubber sub-sector is not impressive as the gentle rise
in the forecasted production trend will be driven by
area increase as the future annual yield levels
plummeted. Price dampening due to lack of market-
tie of the producers with the importing markets which
makes the farmers be at the mercy of the LBs will
affect farmers‟ income productivity, thus affecting
their livelihoods and the enterprise business going
concerned. Therefore, for healthy market
competition, the farmers should constitute themselves
into viable co-operative associations in order to be
competitive enough to venture into direct marketing
i.e. exportation of their products and the government
and non-government agencies should assist in linking
these farmers‟ co-operatives directly with the
importing markets. By doing so, the farmers will
have bargaining power for their products, thus
making them earn productive income owing to
remunerative prices from their products.
Conflict of Interest
The authors declare no conflict of interest.
References
Adelodun, A. (2017). Cocoa production in Nigeria: A
literature review. Analysis, Food and
Agribusiness. Retrieved from
http://cpparesearch.org/nu-en-pl/cocoa-
production-in-nigeria/
Agbota, S. (2017). Nigeria‟s rubber plantation going
extinct. The Sun Newspaper.
J. Agric. Environ. Sci. Vol. 5 No. 2 (2020) ISSN: 2616-3721 (Online); 2616-3713 (Print)
Journal of the College of Agriculture & Environmental Sciences, Bahir Dar University 18
https://www.sunnewsonline.com/nigerias-
rubber-plantation-going-extinct/
Ahmed, S.I., and Joshi, M.B. (2013). Analysis of
instability and growth rate of cotton in three
districts of Marathwada. International Journal of
Statistika and Mathematika, 6(3):121-124
Ambrose, N. (2018). How to rejuvenate rubber
Industry. The Nation Newspaper.
https://thenationonlineng.net/how-to-rejuvenate-
rubber-industry/
Ayalew, B. (2015). Supply response of maize in
Ethiopia: Cointegration and Vector Error
Correction Approach. Trends in Agricultural
Economics, 8 (1):13-20
Boyal, V.K., Pant, D.C., and Mehra, J. (2015).
Growth, instability and acreage response
function in production of cumin in Rajasthan.
The Bioscan, 10(1):359-362
Central Intelligence Agency (CIA) (2011). The
World Fact book. Central Intelligence Agency.
Coppock, J.D. (1962). International Economic
Instability. McGraw-Hill, New York, pp. 523-
525.
Cuddy, J.D.A., and Valle, P.A.D. (1978). Measuring
the instability of time series data. Oxford Bulletin
and Economic Statistics, 40:53-78.
Gujarati, D., Porter, D., and Gunasekar, S. (2012).
Basic Econometrics. McGraw Hill, New Delhi.
Hazell, P.B.R. (1982). Instability in Indian food grain
production. Research Report 30, Washington,
D.C., USA: International Food Policy Research
Institute.
Kumar, N.S., Joseph, B., and Muhammed, J.P.K.
(2017). Growth and Instability in Area,
Production, and Productivity of Cassava
(Manihot esculenta) in Kerala. International
Journal of Advance Research, Ideas and
Innovations in Technology, 4(1):446-448
Mesike, C.S., and Esekhade, T.U. (2014). Rainfall
variability and rubber production in Nigeria.
African Journal of Environmental Science and
Technology, 8(1): 54-57
Ogbebor, O.O. (2013). The Sustainability of
Agriculture in Nigeria Using Rubber as a Case
Study. Electronic Theses and Dissertations.
Paul, R.K. (2014). Forecasting wholesale price of
pigeon pea using long memory time-series
models. Agricultural Economics Research
Review, 27(2): 167-176.
Pro-share (2020). Nigerian Agriculture. Accessed
from https://www.proshareng.com/news/Agri
culture/Reviving-the-Cocoa-Industry-in-
Nigeria/37675
Sadiq, M.S., Singh, I.P., and Karunakaran, N. (2017).
Supply response of cereal crop farmers to price
and non-price factors in Rajasthan state of
Nigeria. Journal of Agricultural Economics and
Rural Development, 3(2):203-210
Sandeep, M.V., Thakare, S.S., and Ulemale, D.H.
(2016). Decomposition analysis and acreage
response of pigeon-pea in western Vidarbha.
Indian Journal of Agricultural Research, 50(5):
461-465
SENCE Agric (2020). Agriculture at glance.
Retrieved from https://www.agriculturenigeria.
com/ production
Sihmar, R. (2014). Growth and instability in
agricultural production in Haryana: A District
level analysis. International Journal of
Scientific and Research Publications, 4:1-12.
Stalling, J.L. (1960). Weather indexes. Journal of
Farm Economics, 42: 180-186.
Umar, H.Y., Giroh, D.Y., Agbonkpolor, N.B., and
Mesike, C.S. (2011). An overview of world
natural rubber production and consumption: An
implication for economic empowerment and
poverty alleviation in Nigeria. Journal of Human
Ecology, 33(1):53-59.
Umar, S.M., Suhasini, K., Sadiq, M.S., and Aminu,
A. (2017). Growth and Instability in Yam
Production in Nigeria: An Inter Zone and State
Level Analysis. Dutse Journal of Agriculture
and Food Security, 4(1):10-24.
Umar, S.M., Suhasini, K., Jainuddin, S.M., and
Makama, S.A. (2019). Sources of growth and
instability in cassava production in Nigeria:
evidence from Hazell‟s Decomposition Model.
SKUAST Journal of Research, 21(1):86-95.
