




































In ternationa l
Scholars
Journa ls

 

African Journal of Agricultural Marketing ISSN 2375-1061 Vol. 5 (5), pp. 001-009, May, 2017. Available online at 
www.internationalscholarsjournals.org © International Scholars Journals 

 

Author(s)                                           Author(s) retain the copyright of this article. 
 

 

Full Length Research Paper 

 

Efficiency of the plantain marketing system in 
Ghana: A co-integration analysis 

 
Akwasi Mensah - Bonsu, Akua Agyeiwaa-Afrane and John K. M. Kuwornu* 

 
Department of Agricultural Economics and Agribusiness, P. O. Box LG 68, University of Ghana, Legon, Accra, Ghana. 

 
Accepted 28 September, 2016 

 
The study assesses the efficiency of the plantain marketing system in Ghana using monthly wholesale prices 
in GHS/10 kg covering the period 2004 to 2009. The integration among the plantain markets was tested with 
the Johansen multivariate co-integration analysis and error correction model. The markets chosen for the 
study are Accra market as a central consumption market; Kumasi market, Sunyani market and Koforidua 
market as assembling markets; Goaso market, Begoro market and Obogo market as producing markets. 
These markets were chosen based on the volume of production and trade in the areas. The findings of the 
market integration analysis indicate that arbitrage in the plantain marketing system is working since there are 
both long and short run relationship between Accra market (central consumption market) and the three 
assembling and three producing markets. However, the speed with which prices are transmitted between 
Accra market (consumption market) and the other markets (plantain production and assembling markets) is 
relatively weak at 27.7%, compared to perfect adjustment of 100% threshold. This implies that there is the 
need for further integration especially in the short run. Improvement in market information systems and 
expansion especially into producing areas, as well as accurate, timely, and availability of information on 
plantain prices may be useful in the efficient distribution of plantain from surplus to deficit markets. 
 
Key words: Plantain marketing system, Johansen multivariate co-integration, error correction model, Ghana. 

 
 
INTRODUCTION 

 
Plantain is a basic food product contributing to the food 
security of millions of people in the developing world, and 
constitutes a source of employment and income for the 
rural population (Nkendah and Nzouessin, 2006). 
Plantain is grown in 52 countries with world production of 
33 million metric tonnes (FAO, 2006). Eight African 
countries are named among the top ten world producers 
of plantain, with Ghana producing 2,930,000 metric 
tonnes and ranked as the third world producer 
(FAOSTAT, 2007). It contributes about 770 Kcal/Kg and 
its supply of micronutrients such as iron and zinc is well 
established (Babatunde et al., 2007). 

Plantain is the third most important food crop after yam 
and cassava in terms of volume of production and 
contributes 13.1% to the Agricultural Gross Domestic  
 
 
 
*Corresponding author. E-mail: jkuwornu@ug.edu.gh, 

jkuwornu@gmail.com. 

 
 
 
 

 
Product in Ghana (FAO, 2006). About 8,348,865 ha of 
land area is used to cultivate plantain in Ghana, 
producing an annual average of 2.0 million tonnes of 
fruits, of which more than 95% is sold on the local market 
and the rest exported (SRID-MoFA, 2006). Six of the ten 
administrative regions of Ghana, namely Ashanti, 
Eastern, Brong-Ahafo, Western, Central and Volta 
regions, are designated as plantain producing regions 
(Banful, 1998). Recently, it has become an important 
export commodity in the international market (ISSER, 
2007) and Ghana obtained the highest international price 
with an average of US$ 1.53 per kg for the period 1995 to 
1998 (Martinez and Saavedra, 2001).  

If agricultural growth is to be realized, developing 
countries have to ensure effective and efficient marketing 
and distribution systems. Economic integration results in 
more efficient use of resources increase in trade, 
productivity and overall production (Ismet et al., 1998). 
Efficiency of markets depends, among other things, on 
the number of traders, the level of competition among 

file:///C:\Users\user\Documents\REPUBLICATION\AGRICULTURAL%20SCIENCES\AppData\Local\Temp\www.internationalscholarsjournals.org


 
 
 

 

them and on the amount and costs of information at their 
disposal (Federico, 2007). Food crop marketing in Ghana 
is dominated by the private sector which handles more 
than 95 percent of the products marketed (MoFA, 1987). 
This includes itinerant traders, market based traders and 
food contractors. An efficient farm marketing system is an 
important means for raising the income levels of farmers 
and for promoting the economic development of a country 
(Tamimi, 1999). The study uses integration of plantain 
markets to measure the efficiency of the plantain market 
system in Ghana. This approach is based on the concept 
originally developed by Bressler and King (1970) that an 
efficient commodity market will establish prices that are 
interrelated spatially by transaction and transfer costs and 
inter-temporally by storage costs. If a market is 
integrated, there will be a relatively low spatial and inter-
temporal variation in prices implying that commodity 
market prices will be functionally related (Onyuma et al., 
2006). 

The plantain business is faced with a lot of marketing 
problems which determine whether production can be 
expanded or not (Adetunji and Adesiyan, 2008). The 
relative attention given to plantain is focused on the 
technical and productive viability of plantain in Ghana, 
while little is done on its marketing (Ekboir et al., 2002; 
Owusu-Benoah et al., 2007). Agricultural production 
problems can be overcome through introducing new 
technology and efficient marketing systems (Adetunji and 
Adesiyan, 2008). According to Codjoe (2007) the area 
under production of plantain has been increasing over the 
years and Dankyi et al. (2007), also revealed that farmers 
are aware of the various production technologies, 
however low farm-gate pricing with traders determining 
prices has been a major hindrance to production and 
marketing. It is however obvious that increased pro-
duction without corresponding well-developed and 
efficient marketing system may amount to wastage of 
resources (Adetunji and Adesiyan, 2008). Hence, investi-
gating the spatial integration of producing and consuming 
plantain markets is useful in ascertaining the efficiency of 
plantain marketing in Ghana. The investigation provides a 
probable means of enhancing the efficiency of the 
marketing system for plantain.  

The objective of the study is to examine the efficiency 
of the plantain marketing system in Ghana by measuring 
price transmission and market integration for plantain 
(False horn) in some selected regional and district 
markets in Ghana. There is little disagreement on the 
benefits of a well-integrated market system. In general, 
producer marketing decisions are based on market price 
information, and poorly integrated markets may convey 
inaccurate price information, leading to inefficient product 
movements (Goodwin and Schroeder, 1991). Linkages to 
marketing centres have been found to contribute 
significantly to rural household’s escape from poverty 
(Krishna, 2004; Krishna et al., 2004). Market integration 
ensures that a regional balance occurs among food 

 
 
 
 

 

deficit, surplus and non-cash crop producing regions 
(Delgado, 1986). Moreover, how long an initially localized 
scarcity can be expected to persist depends entirely on 
knowledge of how well the region is connected by 
arbitrage to other regions (Ravallion, 1986). Improving 
marketing efficiency is a way to increase social welfare by 
generating income for the local producers and chain 
actors and by promoting its sustained use. The paper 
contributes towards identification of the degree and level 
of spatial market integration of the plantain market in 
Ghana, adds to the literature on the efficiency of the 
plantain marketing system, and provides guidance for 
policy aimed at enhancing economic activity in the 
plantain marketing system. 
 
 
METHODOLOGY 
 
Theoretical framework 
 
This research adopts market integration as a measure of marketing 
efficiency. Barret and Li (2002) defined market integration as 
tradability or contestability between markets. Market integration can 
be interpreted as the extent to which price shocks are transmitted 
between spatially separate markets (Goodwin and Piggott, 2001). 
This study assesses market integration and price transmission for 
plantain across some selected markets in Ghana using the 
Johansen multivariate co-integration procedure and the error 
correction model.  

The goal of market integration analysis is to determine marketing 
efficiency which is basically the extent and speed of price 
transmission between spatially separated markets (Goleti et al., 
1995). It is built on the premise that if a pair of markets is 
integrated, a price change in one of them will be reflected in a price 
change in the other. The demand for and price of a given unit of 
plantain in a market would have a dominant effect on the plantain 
trade and by extension, price formulation in other trading markets. 
This would be an indicator for marketing efficiency since price 
differences between the given markets would reflect only 
transportation costs including normal profit (Delgado, 1986). The 
more integrated a market is the more efficient it is. 
 

 
Estimation of plantain market integration: Johansen co-
integration test procedure 
 
Unit root test 
 
When investigating for market integration, the study first examined 
each price series for evidence of non-stationarity in order to confirm 
that co-integration approach is the appropriate tool (Fossati et al., 
2007). The number of lags in the Augmented Dickey Fuller (ADF) 
equation is chosen to ensure that serial correlation is absent using 
the Akaike’s information criterion. The ADF equation estimated by 
OLS is based on a model with a constant as follows: 
 

 = + −1 + −1−1 +  (1) 
Where D  is the differencing operator; Pt   is the price variable of 

interest and ε T  is a white noise process; the unit root test is: 

 
Null Hypothesis: H0: δ = 0 (Pt is non-stationary or has a unit root)  
Alternate Hypothesis: H1: δ ≠ 0 (Pt is stationary or has no unit root). 



 
 

 
If the null hypothesis which states that the price series is non-
stationary is not rejected, then in literature it has been suggested to 
difference the variable of interest to stabilize its mean. If a variable 
is found to have a unit root, the difference of the variable is included 
in the model. The procedure requires that the non-stationary 
variable be differenced sequentially until it attains stationarity. Since 
the price series were not stationary in levels (integrated of order 
zero), they were differenced once to attain stationarity. 

 

Testing for lag length 
 
A test for a suitable lag length to be included in the co-integration 
test was performed, because the results of co-integration tests can 
be quite sensitive to this (Hafer and Sheehan, 1991: Hai et al., 
2004). The number of lags is selected by applying three different 
multivariate lag selection criteria: the Akaike information criterion 
(AIC), the Hannan-Quin information criterion (HQIC), and the 
Schwarz’s Bayesian information criterion (SBIC). A vector auto-
regression (VAR) on the differenced series was conducted and lag 
length of the model with the least AIC, HQIC and SBIC values 
chosen as the appropriate lag length to be included in the co-
integration test. The test started with a lag length of 12 and then 
shortened till the least values of the AIC, HQIC and SBIC were 
obtained. 

 

Johansen multivariate co-integration test 
 
Following Ekpe (2005), the test of long-run market integration was 
done with the Johansen multivariate co-integration analysis. 
Johansen maximum likelihood estimation uses a rank test to define 
the number of co-integrating vectors r that can be found in a data. 
The rank is determined by estimating the p-dimensional VAR (k) 
model given in Equation (2): 
 

  
+ 

  
 

 = ŋ +     (2)   −1  
 

 =    
 

 
Where t = 1, 2, ... refers to the months from January 2004 to 
December, 2009; Pt is a n × 1 vector of the logarithmic prices at 

time t (Pt = P1t, P2t,…,Pnt); Ai are n × n matrices of parameters; ŋ is 

an n × 1 vector of intercept terms; εt is a n×1 vector of error terms, k 

is the lag length and εt is the vector of error terms assumed to be 
NID(0,).  

On the basis of the Granger representation theorem, when the 
price series are integrated of order one, I(1), the VAR model can be 
re-parameterized into vector error correction model (VECM): 
 

 K      
 

P 
η∑ΠI 

P 
Π 

P 
−1 

(3) 
 

T T −1 T εT  

I 1 

  
  

 

 
If case (a) holds, then ordinary least squares estimate of the levels 
of models 2 and 3 are directly related by models 4 and 5. If case (b) 
holds, then model 3 is appropriate and Equation 2 is over-
parameterized and its estimates are inefficient. Between these two 
extremes are the vector error correction models that come under 
case (iii). In estimating the number of co-integrating equations 

present in the system, the rank of ∏ in Equation 3 was determined 

using the trace statistic (Johansen, 1991; 1988) results of the 

Johansen co-integration test available in EVIEWS 5. If ∏ = 0, then 

no co-integration relationship exists; if the rank of ∏ is equal to r < 

n, then there are r co-integrating relationships and s = (n - r) 
common trends. The extent of an integrated market requires that s  
= 1, that is perfect integration. If there is more than one common 
trend, then some prices could be generated by the first common 
trend and some by s combination of the first and other trends. 
According to Gonzalez-Rivera and Helfand (2001), such markets 
could not be said to be integrated because the long run movements 
in prices would be governed by different components. 
 
The error correction model (ECM) 
 
The error correction model is used to test the short-run integration 
restrictions and the speed of integration. Johansen defines two matrices 
α and β, such that ∏ = αβ', where both α and β are (n × r) matrices. This 

implies that; ∏−1 =′−1 . The error 
 
correction term which gives the short-run disequilibrium is X, hence 

=−1 . The matrix α is the weights with which each co-  
integrating vector enters the n equations of the vector error 
correction model (VECM), and β is the matrix of co-integrating 
relations. α is also the matrix of the speed of adjustment 
coefficients. 

Substituting for ∏−1 =in Equation 3 implies: 
 

∆ =   
+ + (6)  

ŋ 
+  ∏ ∆ 

 

 −1   
 

=1 
 
Where X is an (n-s)×1 vector of error correcting term given by the 
estimated residual from equation (2) and α is the speed of price 
adjustment to equilibrium level. A dummy variable was introduced 
into equation (6) for year 2007 where there was a shock in the 
plantain marketing system. Hence: 
 

∆ =   
+ + + (7)  

ŋ 
+  ∏ ∆ 

 

 −1  
 

 

  
 

=1 
 
The long-run co-integration of the plantain price series was 
determined by analyzing the normalised co-integrating coefficients 
(β). In estimating the co-integrating coefficients (β), the Johansen 
co-integration test as implemented in EVIEWS was used. 

 

Where ∏ and ∏i are defined by: 
 Wald test for market integration 

 

  
 

   The type and degree of market integration is determined by the 
 

 

(4) statistical significance of the estimated parameters based on the 
 

  
 

∏ = − =+1   results of the set of hypothesis using the F-statistic of the Wald 
 

  
(5) tests restrictions. The following restrictions were tested on the OLS 

 

∏ =   − 1 
estimation of the regression Equation (2): 

 

   
 

=1 
  

 

  
Long-Run Market Integration 

 

   
 

Where ∏i and ∏ are (n×n) matrices; ∏ has a reduced rank r = n – 
Ho: Plantain market prices are integrated in the long-run, that is  

s;  and  εt is  a  (n×1)  error  vector.  According  to  Silvapulle and  

 
 

Jayasuriya (1994), there are three possible cases for the rank of ∏. 
 

  = 1 . 
 

    
 

   =1 
 

a. Rank ∏ is equal to n, that is r = n (∏ is unrestricted).  HA: Plantain market prices are not integrated in the long-run, that is, 
 

b. Rank ∏ is equal to zero.   
 

   

c. Rank ∏ lies between zero and n, that is, 0 < r > n. 
   ≠ 1 .  

 =1 
 



 
 
 

 
Table 1. Results of unit root test in levels and first differences.  

 
 Market Levels (intercept only) First differences (intercept only) Order of 

 price ADF P-values ADF P-value integration 

 Accra -2.1140 0.0846 -6.6830 0.0000 1 

 Goaso -2.8890 0.0910 -9.4910 0.0000 1 

 Begoro -3.5510 0.6800 -9.6220 0.0000 1 

 Obogu -1.6240 0.5239 -14.0130 0.0000 1 

 Kumasi -0.8280 0.8620 -12.2000 0.0000 1 

 Sunyani -2.5250 0.4250 -6.5870 0.0000 1 

 Koforidua -2.7330 0.4160 -7.2410 0.0000 1 
 

MacKinnon critical value for the prices in levels at 5% is -2.9130. MacKinnon critical value for the prices in first differences at 5% is 
- 2.9140. 

 
 

 
b) Short –Run Market Integration 
 
Ho: A price change in a market is immediately transmitted to the 
other market, that is, Ai0 = 1.  
HA: A price change in a market is not immediately transmitted to the 
other market, that is, Ai0 ≠ 1. 
 
The null hypothesis is not rejected if the probability value of the F 
statistic test is less or equal to 0.10, but rejected if the probability 
value is greater than 0.10 (i.e. 10 percent critical level). 
 
 
Description of variables and sources of data 
 
Markets are said to be integrated if a process of arbitrage connects 
them. This will be reflected in the price series of commodities in 
spatially separated markets (Van Campenhout, 2005). Thus, the 
analysis of this study is entirely based on secondary price data 
collected from the Information Management System Units of seven 
District Agricultural Development Units in Ghana.  

Average monthly wholesale prices for 10 kg of plantain from 
January 2004 to December 2009 (72 observations) in four regional 
and three district markets in Ghana were used in the study. Prices 
of plantain (and some other agricultural commodities) are collected 
by the various District Agricultural Development Units throughout 
the country on weekly basis. Based on the weekly data, monthly 
price series are computed by taking averages. It is worth 
considering the appropriateness of using monthly series in testing 
spatial market integration. For instance, Hai et al. (2004) argues 
that average monthly data are inappropriate when analyzing rice 
market integration in the Mekong River Delta of Niger because 
these prices do not reflect the daily prices on which traders make 
their arbitrage decisions. However, weekly price series have several 
long periods in which prices are constant in almost every market. 
Such constancy of prices will invalidate a statistical analysis which 
is based on an assumption of independent and identically 
distributed innovations that follow a continuous distribution. 
Moreover, the use of weekly data is problematic due to the need to 
interpolate numerous missing values. For these reasons this study 
analyzes monthly data instead of the weekly data.  

In addition, the consumer price index (with 2002 as base year) 
was used to convert nominal price data into real values. This is 
justified because through this conversion, correlation by inflation 
can be excluded (Fafchamps and Gavian, 1995).  

The seven (7) markets were purposely selected based on the 
production and volume of plantain traded. They include one 
consuming market; Agbogbloshie (Greater Accra); three 
assembling markets; Kumasi (Ashanti Region), Koforidua (Eastern 

 
 
 

 
Region) and Sunyani (Brong –Ahafo Region); and three producing 
markets; Obogu (Asante Akyem- South in the Ashanti Region), 
Begoro (Fanteakwa District in the Eastern Region) and Goaso 
(Asunafo North District in the Brong –Ahafo Region). Based on the 
trade relations between the markets, spatial integration between 
these seven markets is assessed. 
 

 

RESULTS AND DISCUSSION 
 
Unit root test 
 

In order to ascertain whether the variables were 
stationary or not, the ADF unit root test was applied on 
the levels and first differences of the prices series. The 
results presented in Table 1 indicate that the real 
wholesale price series for the seven markets are 
integrated in order one, that is, I (1). The unit root test 
(with an intercept) shows that none of price series is 
stationary at the level at the 5 percent significance level. 
To make them stationary, their first differences were 
taken. When all the price series were differenced once, 
the results of the unit root test indicate that the null 
hypothesis of a unit root can be rejected at the 5% 
significance level.  

The lag selection-order criterion was used to select the 
appropriate lag length to be included in the cointegration 
model. Table 2 presents the results of the lag selection-
order for the various information criteria. On the basis of 
the Akaike’s information criterion (AIC), Hannan-Quin 
information criterion (HQIC) and Schwarz’s Bayesian 
information criterion (SBIC), one month of lag was 
selected for the model. 
 

 

Johansen multivariate results for co-integration 
 
To examine the hypothesis that there are r co-integrating 
vectors, the trace test was performed. Table 3 reports  
the results for the Johansen trace statistic (λtrace) test 
based on the smallest value of the AIC and SBIC values.  
Comparing the trace statistic  with the  corresponding 



  
 
 

 
Table 2. Lag selection-order criteria.  

 
Lag AIC HQIC SBIC 

0 3.6838 3.7731 3.9086 

1 1.0924* 1.8069* 2.8912* 

2 1.2026 2.5423 4.5754 
 

*AIC, HQIC and SBIC values are smallest at 1% significant level. 
 

 
Table 3. Results of Johansen co-integration tests.  

 

 Null hypothesis Alternative hypothesis λtrace value 5% critical value 

 r = 0 r > 0 157.9151* 124.24 

 r ≤ 1 r > 1 92.8623 94.15 

 r ≤ 2 r > 2 48.9786 68.52 

 r ≤ 3 r > 3 30.689 47.21 

 r ≤ 4 r > 4 15.4227 29.79  
*If value of λtrace exceeds the critical value, we reject the null hypothesis and accept the alternative of more co-

integrating vectors. 
 

 

critical values, it can be seen that the null hypothesis of 
no co-integrating relationship can be rejected at the 5% 
significance level for the plantain wholesale market real 
prices. In carrying out the co-integration test, the 
deterministic trend was used because of the 
characteristics of the line plots of the plantain prices in 
the various selected markets. The trace test statistics 
reported in Table 3 indicates that at least a stationarity 
relationship exists among the seven plantain wholesale 
markets. The paper discusses the price relationship 
between a central market (consuming, but non-producing 
market) and other markets (assembling and producing 
markets) for plantain.  

The co-integration results for wholesale real price in 
Accra (the central market) are presented in Table 4. The 
explanatory variables jointly explain the variation in the 
Accra wholesale price with the F -statistic significant at 

the 1% level. The coefficient of determination (R
2
) of 

approximately 83% means that the variables in the model 
explain 83% of the variations in current wholesale real 
price in Accra. The coefficients of the current and 
previous month wholesale real prices in Kumasi and 
Obogu markets are insignificant. The coefficients of the 
other markets wholesale real prices are significant.  

From Table 4, the coefficient of previous month 
wholesale real price for Accra shows a positive sign, 
meaning that a higher wholesale real price in Accra in the 
previous month would reflect in a higher wholesale price 
in Accra in the current month. Also, the current plantain 
wholesale real prices in Sunyani and Goaso have positive 
and significant effects on the current wholesale real price 
for plantain in Accra. Thus, within the same month, higher 
(lower) wholesale real prices in the Sunyani and Goaso 
markets will result in higher (lower) real prices in the 
Accra market. On the other hand, lower 

 
 

 

previous month wholesale real prices in both Sunyani and 
Goaso would reflect in higher current month wholesale 
real prices in Accra.  

On the contrary, current wholesale real prices in 
Koforidua and Begoro have negative and significant 
effects on the current wholesale real prices in Accra. This 
means that higher (lower) current wholesale prices in 
Koforidua and Begoro would reflect in lower (higher) 
wholesale real prices for Accra market in the current 
month. The coefficients of previous month wholesale real 
prices in Koforidua and Begoro markets show positive 
signs, indicating that lower (higher) wholesale real prices 
in Koforidua and Begoro in the previous month would 
reflect in lower (higher) wholesale real prices in Accra 
during the current month. 

The significant coefficients in the co-integration model 
indicate that in the long-run most of the plantain markets 
are highly co-integrated with the wholesale real price in 
Accra (central market). In particular, the wholesale real 
prices for Koforidua and Begoro, assembling and 
producing markets, respectively, in the Eastern region 
and Sunyani and Goaso, assembling and producing 
markets, respectively, in the Brong-Ahafo region are 
spatially integrated with the wholesale real price in Accra 
(the central consuming market), while the wholesale real 
prices in Kumasi and Obogu assembling and producing 
markets respectively in the Ashanti region are not 
spatially integrated with the wholesale real price in Accra 
(the central consuming market).  

But the directional effects of the integration between 
Accra market and the markets in the Eastern region are 
different from (opposite to) the directional effects of the 
integration between Accra market and the markets in the 
Brong-Ahafo region. Thus, while the current wholesale 
real prices in Sunyani and Goaso have positive effects on 



     

 Table 4. Co-integration regression results.     
      

 Variable Coefficient Std. error t-Statistic P-value 

 Accra wholesale real price (-1) 0.5590*** 0.1027 5.4500 0.0000 

 Kumasi wholesale real price 0.0517 0.0966 0.5300 0.5930 

 Kumasi wholesale real price (-1) 0.0678 0.0711 0.9500 0.3440 

 Sunyani wholesale real price 0.2837** 0.1133 2.5000 0.0150 

 Sunyani wholesale real price (-1) -0.3006** 0.1164 -2.5800 0.0120 

 Koforidua wholesale real price -0.6112*** 0.1496 -4.0900 0.0000 

 Koforidua wholesale real price (-1) 0.6112*** 0.1551 3.9400 0.0000 

 Goaso wholesale real price 0.2852** 0.1348 2.1200 0.0340 

 Goaso wholesale real price (-1) -0.2852** 0.1292 -2.2100 0.0270 

 Begoro wholesale real price -0.2781* 0.1426 -1.9500 0.0510 

 Begoro wholesale real price (-1) 0.2781** 0.1376 2.0200 0.0430 

 Obogu wholesale real price -0.0470 0.0573 -0.8200 0.4120 

 Obogu wholesale real price (-1) 0.0470 0.0629 0.7500 0.4550 

 C 0.1902*** 0.0648 2.9300 0.0050 

 R-squared 0.83144 F-statistic 21.6200 

 Adjusted R-squared 0.79303 Prob (F-statistic) 0.0000 
 

Source: Authors’ computation based on data from MoFA, 2010. *** indicates significance at the 1 percent level, ** indicates significance 
at the 5 percent level and * indicates significance at the 10 percent level. Dependent Variable: Accra Wholesale Price. Included 
observations: 71 after adjustment. 

 

 

the current wholesale real price in Accra, the current 
wholesale real prices in Koforidua and Begoro have 
negative effects on the current wholesale real price in 
Accra. On the other hand, while the previous month 
wholesale real prices in Sunyani and Goaso have 
negative effects on the current wholesale real price in 
Accra, the previous month wholesale real prices in 
Koforidua and Begoro have positive effects on the current 
wholesale real price in Accra. The reasons for these 
mixed results are not very clear, but the negative effects 
do not meet prior expectations.  

The implication is that the positive response of the 
wholesale real price in Accra market to the wholesale real 
prices for plantain in the Sunyani and Goaso markets 
(Borng-Ahafo region) is prompt, while its positive 
response to the wholesale real prices in Koforidua and 
Begoro markets (in the Eastern region) lags by about a 
month. On the other hand, the negative response of the 
wholesale real price in Accra market to the wholesale real 
prices for plantain in the Sunyani and Goaso markets 
(Borng-Ahafo region) lags by about a month, while its 
negative response to the wholesale real prices in 
Koforidua and Begoro markets (in the Eastern region) is 
prompt. The differences in these effects could be due to 
differences in trade volume/levels and information flow 

 
 

 

among the different regions. 
 

 

Error correlation model 
 

The vector error correction model (VECM) was 
constructed in order to analyze the short-run dynamics of 
the effects of plantain prices in other selected markets on 
plantain prices in Accra, having established that a long-
run relationship exists between the variables. The results 
of the error correction model capture short run 
relationships. The result of the estimated error correction 
model which expresses the first difference of the Accra 
wholesale price of plantain as a function of the first 
difference of the explanatory variables is presented in 
Table 5. The empirical results suggest good explanatory 

power of the model as indicated by the R
2
 statistics as 

well as the F-statistics for the overall regression. The F-
statistic of 21.73 indicates the joint significance of current 
and previous plantain prices in Kumasi, Sunyani, 
Koforidua, Goaso, Begoro and Obogu and previous 
prices in Accra in explaining about 68.82% of the 
variation in the plantain wholesale prices in Accra in the 
short period.  

The  adjustment coefficient  of the lagged  of the first 



      

 Table 5. Error correction model results.       
        

 Variable Coefficient Std. error t-Statistic Prob.  

 1st diff Accra wholesale real price (-1) 0.6543 0.0930 7.0585 0.0000   

 1st diff Kumasi wholesale real price 0.3117 0.4023 0.7748 0.1529   
 1st diff Kumasi wholesale real price (-1) -0.1707 0.2191 -0.7792 0.4371   

 1st diff Sunyani wholesale real price 0.3802 2.9429 0.1292 0.0031   
 1st diff Sunyani wholesale real price (-1) -0.2564 0.1356 -1.9050 0.1005   

 1st diff Koforidua wholesale real price 0.2094 0.1183 1.7704 0.0809   
 1st diff Koforidua wholesale real price (-1) -0.2749 0.1327 -2.0720 0.0502   

 1st Diff Goaso wholesale real price 0.1400 0.1642 0.8525 0.4004   
 1st diff Goaso wholesale real price (-1) 0.4051 0.3700 2.6323 0.0052   

 1st diff Begoro wholesale real price 0.2708 0.1120 2.4174 0.0209   
 1st diff Begoro wholesale real price (-1) -0.2781 0.1202 -2.3129 0.0194   

 1st diff Obogu wholesale real price -0.0621 0.3698 -0.1679 0.8731   
 1st diff Obogu wholesale real price (-1) -0.4019 0.3785 -1.0617 0.1535   

 Dummy (2007) -0.0101 0.0592 -0.1706 0.8652   

 X (Residual term) -0.2773 0.1391 -1.9930 0.0473   
 Constant 0.0140 0.0329 0.4244 0.3913   

 R-squared 0.6882 F-statistic 21.7277   

 Adjusted R-squared 0.6403 Prob(F-statistic) 0.0001   
 

X (residual term) denotes the error correcting term. Dependent variable:  First differenced Accra wholesale real price.  
Included observations: 70 after adjustments. 

 
 

 

differenced of the wholesale real price in Accra market 
has a positive and significant effect on price stability, 
suggesting that any previous disequilibrium in the long-
run wholesale real price in Accra market would be 
corrected in the short-run. The short-run effects of both 
the current and lagged first differenced of the wholesale 
real prices in Kumasi and Obogu markets are again 
insignificant. Also, the short-run effect of the current first 
differenced of the wholesale real price in Goaso market is 
also insignificant. However, the current first differenced of 
the wholesale real prices in Sunyani, Koforidua, and 
Begoro markets have positive and significant effects on 
stabilizing the wholesale real price in Accra, while the 
lagged of the first differenced of their wholesale real 
prices have negative and significant effects. Also, the 
lagged of the first differenced of the wholesale real price 
in Goaso has positive and significant effect on stabilizing 
the wholesale real price in Accra. These significant 
adjustment coefficients suggest that any distortion in the 
long-run wholesale real price in Accra would be corrected 
in the short-run, by the required either upward or 
downward movement towards the long-run equilibrium 
level. The dummy variable for price shock in the plantain 

 
 
 

 

marketing system in year 2007 also has the expected 
negative sign but an insignificant effect on the stability of 
the wholesale real price in Accra.  

The coefficient of the error correction term, X, which 
signifies the speed at which plantain wholesale real price 
in Accra adjusts to their long run equilibrium level, is 
negative and statistically significant at the 5% level. The 
significant coefficient of the error correction term confirms 
the existence of a long-run equilibrium relationship of 
wholesale real price for plantain in Accra with the 
wholesale real prices in the other markets included in the 
analysis. The coefficient of the error correction term of 
0.2773% implies that, the feedback into the short -run 
dynamic process from the previous period is 27.73% and 
the negative sign suggests that the adjustment is from a 
higher price shock (price rise) to the long-run price level. 
This means that the adjustment from the short-run to 
long-run equilibrium is about 27.73% which is relatively 
weak compared to perfect adjustment of 100% threshold. 
It suggests that the wholesale real price in Accra adjust 
partially to its long-run level after a price rise (shock). 
Nkendah and Nzouessin (2006) found a 21 to 27% 
speeds of adjustment of prices in pairs of plantain 



 
 
 

 
Table 6. Wald test results for market integration.  
 
Null hypothesis Wald test statistics P value   

F-statistics   
Long-run market integration  

 

0.00 0.9614 
 

 
Short-run market integration   

Ai0 = 1 0.11 0.7426  
 

 

markets in Cameroon and concluded that a weak 
integration exist in the markets and hence urban 
consumer price increase because there is bad 
information circulation between the various markets. 

 

 

Wald test for market integration 
 
In order to find the nature of long-run and short-run 
market integration, the Wald test restriction of the F-
statistic was applied to determine market integration in 
the plantain markets. Table 6 presents the results of the 
Wald test. The F statistical values of 0.00 and 0.11 with 
probability values of 0.9614 and 0.7426, respectively, 
show that they are not significant even at 10%. The long-
run and short-run null hypotheses that plantain market 
real prices are integrated and a price change in a market 
is immediately transmitted to other markets, respectively, 
therefore cannot be rejected. The results mean that there 
exist both long-run and short run market integrations 
between Accra and the other selected markets (both 
producing and assembling markets). Thus, changes in 
plantain real price in any producing or assembling market 
would cause plantain real price in Accra to adjust 
immediately and the estimated speed of adjustment is 
about 27.73% (Table 5). 

 

 

Conclusions 
 

This study explored market integration for plantain 

monthly prices in Ghana, for the time period 2004 to 
2009, using the Johansen multivariate co-integration 
approach and the error correction model. The results of 
the market integration analysis obtained by the Johansen 
multivariate co-integration approach indicate at least one 
co-integrating equation. While the wholesale real prices 
for markets in the Eastern region (Koforidua and Begoro 
markets) and Brong-Ahafo region (Sunyani and Goaso 
markets) are spatially integrated with the wholesale real 
price in Accra (the central consumption market), the 
wholesale real prices for markets in the Ashanti region 
(Kumasi and Obogu markets) are not. But the directional 
effects (signs of the coefficients) of the integration 

 
 
 
 

 

between Accra market and the markets in the Eastern 
region are different from (opposite to) the directional 
effects of the integration between Accra market and the 
markets in the Brong-Ahafo region. The differences in 
these effects could be due to differences in trade 
volumes/levels and information flow among the different 
regions.  

Again the results of the market integration analysis 
obtained by employing the error correction model (ECM) 
show that price signal is transmitted in the short-run 
between the current wholesale real price in Accra market 
and the other markets. The Wald test results for market 
integration suggest that the long-run and short-run null 
hypotheses that plantain market real prices are integrated 
and a price change in a market is immediately transmitted 
to other markets, respectively, cannot be rejected. The 
estimated coefficient for the error correction term 
suggests that plantain real price signals are transmitted 
between Accra market (the consumption market) and 
other selected markets in the short-run with a negative 
speed of adjustment of 27.73%. These results show that 
there is relatively weak integration/adjustment of the 
plantain production and assembly markets to the 
consumption market compared to perfect adjustment of 
100% threshold. It suggests that the wholesale real price 
in Accra adjust partially to its long-run level after a 
(higher) price shock. This would mean urban consumer 
real price for plantain increases and returns only partially 
to its expected long-run real price level.  

The speed with which price signals are transmitted in 
the short-run shows that there is the need for further 
market integration between the plantain markets 
especially in the short-run. Thus, high development of 
market information delivery will further help to enhance 
market integration. Expansion of market information 
systems especially into producing areas can be 
considered. More information on availability and prices of 
plantain may partially be useful in monitoring the status of 
food security situations in the country. If market 
participants have accurate and timely information on 
plantain conditions, plantain markets may be able to 
respond more quickly to market shocks, and market 
channel members can efficiently and effectively distribute 
plantain from surplus to deficit markets. 
 
 
REFERENCES 
 
Adetunji MO, Adesiyan IO (2008). Economic Analysis of Plantain 

Marketing in Akinyele Local Government Area in Oyo State, Nigeria. 
Int. J. Agric. Econ. Rural Dev., 1(1): 15-21.  

Babatunde RO, Omotesho OA, Sholotan OS (2007). Socio-economic 
Characteristics and Food Security Status of Farming Households in 
Kwara State, North-Central Nigeria. Pakistan J. Nutr., 6(1): 49-58.  

Banful B (1998). Production of Plantain, an Economic Prospect for Food 
Security in Ghana, In C. Picq, E. Fouré, and E. A. Frison (eds.), 
Bananas and Food Security/Les Productions Bananières: Un Enjeu 
Économique Majeur Pou La Sécurité Alimentaire, Proceedings of an 
International Symposium Held in Douala, Cameroun, 10 – 14 
November 1998, International Network for the Improvement of 
Banana and Plantain, Montpellier, France. 



 
 
 

 
Barret CB, Li JR (2002). Distinguishing between Equilibrium and 

Integration in Spatial Price Analysis. Am. J. Agric. Econ., 84: 292-
307. 

Bressler RG, King RA (1970). Markets, Prices and International Trade. 
International Research and Rural Studies, New York: John Wiley and 
Sons.  

Codjoe SNA (2007). Supply and Utilisation of Food Crops in Ghana, 
1960-2010. Afr. J. Food Agric. Nutr. Dev., 7(2): 1-15.  

Dankyi AA, Dzomeku BM, Anno-Nyako FO, Adu-Appiah A, Antwi G 
(2007). Plantain Production Practices in the Ashanti, Brong-Ahafo 
and Eastern Regions of Ghana. Asian J. Agric. Res., 1(1): 1-9. 

Delgado CA (1986). Variance Components Approach to Food Grain 
Market Integration in Northern Nigeria. Amer. J. Agric. Econ., 68(4): 
970-979. 

Ekboir J, Boa K, Dankyi AA (2002). Impacts of No-Till Technologies in 
Ghana. Economic Program Paper, 02-01, International Maize and 
Wheat Improvement (CIMMYT), Mexico, D.F. 
apps.cimmyt.org/Research/Economics/map/research_results/progra 
m_papers/pdf/EPP02_01.pdf 

Ekpe SK (2005). Evaluation of the Performance of the Egg Marketing 
System in Ghana (Unpublished Master’s Thesis). Department of 
Agricultural Economics and Agribusiness, University of Ghana, 
Legon-Accra, Ghana.  

Fafchamps M, Gavian S (1995). The Spatial Integration of Livestock 
Markets in Niger. J. Afr. Econ., 5(3): 366-405. 

FAO (2006). Food and Agricultural Indicators, Statistical Analysis 
Service (ESSA), Food and Agriculture Organization, Rome.  

FAOSTAT (2007). Production Year Book, Food and Agriculture 
Organization Statistics, Food and Agriculture Organization of the 
United Nations, Rome. 

Federico G (2007). Market Integration and Market Efficiency: The Case 
of 19th Century Italy.  Explor. Econ. Hist., 2(44):  293-316.  

Fossati S, Lorenzo F, Rodriguez CM (2007). Regional and International 
market Integration of a Small Open Economy. J. Appl. Econ., 0: 77-
98, Universidad del CEMA, 

Golett F, Raisuddin A, Naser F (1995). Structural Determinants of 
Market Integration: The Case of Rice markets in Bangladesh. Dev. 
Econ., 33(2): 185-202. 

Gonzalez-Revera G, Helfand SM (2001). The Extent, Pattern, and 
Degree of Market Integration: A Multivariate Approach for the 
Brazilian Rice Market. Am. J. Agric. Econ., 83(3): 576-592. 

Goodwin BK, Piggott N (2001). Spatial Market Integration in the 
Presence of Threshold Effects. Am. J. Agri. Econ., 83(2): 302-317. 

Goodwin BK, Schroeder TC (1991). Cointegration Tests and Spatial 
Price Linkages in Regional Cattle Markets. Am. J. Agric. Econ., 73: 
452-464. 

Hafer RW, Sheehan RG (1991). Policy Inference Using VAR Models. 
Econ. Enquiry, 29(1): 44-52. Oxford University Press. 

Hai LTD, Lutz C, Praagman C (2004). Rice Market Integration in the 
Mekong River Delta: the Successful Liberalisation of the Domestic 
Food Market in Vietnam. Research Report 04B10, University of 
Groningen, Research Institute, Systems, Organisation and 
Management (SOM).  

Ismet M, Barkley AP, Llewelyn RV (1998). Government Intervention and 
Market Integration in Indonesian Rice Markets. Agric. Econ., 19: 283-
295. 

ISSER (2007). The State of the Ghanaian Economy in 2006. Institute of 
Statistical Social and Economic Research, University of Ghana, 
Legon, Ghana. 

  
  

 
 

 
Johansen S (1988). Statistical Analysis of  Co-integration Vectors.  J. 

Econ. Dyn. Control, 12: 231-234.  
Johansen S (1991). Estimation and Hypothesis Testing of Co-

integrating Vectors in Gaussian Autoregressive Models. 
Econometrica, 59: 1551-80. 

Krishna A (2004). Escaping Poverty and Becoming Poor: Who Gains,  
Who Loses, and Why? World Dev., 32: 121-136. 

Krishna A, Janson PK, Radeny M, Nindo W (2004). Escaping Poverty 
and Becoming Poor in 20 Kenyan Villages. J. Hum. Dev., 5: 211-226.  

Martinez JLR, Saavedra AR (2001). Socioeconomic Aspects of Plantain 
Cultivation in Colombia. InfoMusa, 10(1): 4-9.  

MoFA (1987). Strengthening Agricultural Market Service. National 
Agricultural Marketing Development Plan, Ministry of Food and 
Agriculture, Ghana. Food and Agriculture Organization Project. 
TCP/GHA/4504; Accra, Ghana.  

Nkendah R, Nzouessin CB (2006). Economic Analysis of the Spatial 
Integration of Plantain Markets in Cameroon. Research paper, 
African Economic Research Consortium, July 2006.  

Onyuma SO, Icark E, Owuor G (2006). Testing Market Integration for 
Fresh Pineapples in Kenya. Paper Presented at the International 
Association of Agricultural Economics Conference, Gold Coast, 
Australia, August 12-18.  

Owusu-Bennoah E, Anno-Nyako FO, Egyir IS, Banful B (2007). 
Methodological Framework: Analyzing the Agricultural Science 
Technology and Innovation (ASTI) Systems in ACP Countries: The 
Ghana Case Study on Plantains. Wageningen, The Netherlands: 
Technical Centre for Agricultural and Rural Cooperation (CTA).  

Ravallion M (1986). Testing Market Integration. Am. J. Agric. Econ., 68: 
102-109.  

Silvapulle P, Jayasuriya S (1994). Testing for Philippines Rice Market 
Integration: A Multiple Co-integration Approach. J. Agric. Econ., 45: 
369-380. 

SRID-MoFA (2006). Statistics, Research and Information Directorate, 
Ministry of Food and Agriculture, Accra, Ghana. 

Tamimi T (1999). Investigating Production Opportunities, Marketing 
Efficiency and Option of Trade for Fruits and Vegetables in Palestine. 
Stuttgart, Germany: Verlag Ulrich E. Grauer. 

Van-Campenhout B (2005). Modelling Trade in Food Market 
Integration: Method and an Application to Tanzanian Maize Markets. 
Institute of Development Policy and Management (IDPM), University 
of Antwerp-Belgium. 


