A
m
s
i
r
I
c
©
J
K
1
o
m
i
T
t
p
o
o
u
n
o
T
G
t
V
2
(
h
1
HOSTED BY Available  online  at  www.sciencedirect.com
ScienceDirect
Review of Development Finance 5 (2015) 53–63
Is West African Monetary Zone (WAMZ) a common currency area?
Simon K. Harvey a,∗, Matthew J. Cushing b
a Department of Finance, University of Ghana Business School, University of Ghana, Legon, Ghana
b Department of Economics, College of Business Administration, University of Nebraska, Lincoln, United States
bstract
In this paper, we test whether the West African Monetary Zone (WAMZ) is a common currency area by using a structural vector autoregressive
odel to study the variance decomposition, impulse responses of key economic variables and linear dependence of the underlying structural
hocks of the countries in the zone. The variance decomposition shows that the zone as a whole does not have common sources of shock, which
s expected because of the diverse economic structures of these countries. The correlation of the structural shocks also shows that these countries
espond asymmetrically to common supply, demand and monetary shocks and will therefore respond differently to a common monetary policy.
t is therefore not in the interest of the individual countries to go into a monetary union now or in the near future unless the economies of these
ountries converge further. 2015 Africagrowth Institute. Production and hosting by Elsevier B.V. All rights reserved.
EL classification: F33
c
i
m
t
A
W
l
c
o
r
I
p
W
A
A
i
i
a
t
eywords: Monetary zone; Common currency area; West Africa
.  Introduction
The quest for monetary union within Economic Community
f West African States (ECOWAS) began with the establish-
ent of the regional body in May 1975. This quest reflects
n the objectives, as stated in article 2 section 2h of the 1975
reaty of Lagos, a treaty that establishes the community, that
he community shall ensure “harmonization, required for the
roper functioning of the community, of the monetary policies
f the member states.” This is restated in article 3 section 2e
f the July 1991 treaty as “the establishment of an economic
nion through the adoption of common policies in the eco-
omic, financial, social and cultural sectors, and the creation
f a monetary union.” The 15 member states that ratified the
reaty of Lagos are Benin, Burkina Faso, Côte d’Ivoire, The
ambia, Ghana, Guinea, Guinea-Bissau, Liberia, Mali, Mauri-
ania, Niger, Nigeria, Senegal, Sierra Leone, and Togo. Cape
erde joined the community in 1976 and Mauritania left in
000, leaving the current membership still at 15 states. The∗ pCorresponding author. Tel.: +233 507427839.
E-mail addresses: sharvey@ug.edu.gh (S.K. Harvey), mcushing1@unl.edu
M.J. Cushing). Z
Peer review under responsibility of Africagrowth Institute. t
ttp://dx.doi.org/10.1016/j.rdf.2015.05.001
879-9337/© 2015 Africagrowth Institute. Production and hosting by Elsevier B.V. Aommunity is made up of English, French and Portuguese speak-
ng countries.
At the time of establishment of ECOWAS, there was one
onetary zone in West Africa West Africa Economic and Mone-
ary Union (WAEMU) which is composed of Francophone West
frican countries. The CFA, which is the single currency in the
est Africa Economic and Monetary Union (WAEMU), circu-
ates among the member countries. Anglophone West African
ountries, however, have their independent currencies. The idea
f introducing a single currency for ECOWAS as a whole was
e-enforced in the July 1991 Treaty ratified by all member states.
t has been proposed to implement the monetary integration
rocess in two stages by forming a second monetary zone, the
est African Monetary Zone (WAMZ) for the Anglophone West
frica, which will later merge with the existing zone, the West
frica Economic and Monetary Union (WAEMU). Since the
ntroduction of the proposed single currency is in two stages,
.e. forming a monetary union among the non-CFA countries
nd later merge with the CFA countries, we think that analyzing
he convergence of non-CFA countries alone will draw a better
icture of what is needed now by ECOWAS.In this paper, we test whether the West African Monetary
one (WAMZ) is a optimum currency area by using a vec-
or autoregressive model to study the variance decomposition,
ll rights reserved.
5  of De
i r
l e
r d
o d
r p
e b
t
t l
z t
o
t s
a r
t t
t s
u s
m c
o g
T a
r a
e t
a o
t
p t
a c
t B
c l
c a
a
a s
a F
n t
u t
m
I n
r s
a i
fl l
t m
w f
n c
t
c u
t c
m a
a l
q
( a
i t
m t
c
( 2
t W
K
b
d
e
a
4 S.K. Harvey, M.J. Cushing / Review
mpulse responses of key economic variables and by analyzing
inear dependence of and feedback between the structural shocks
ecovered from a structural vector autoregressive (SVAR) model
f key economic variables in the region. Countries with symmet-
ic shocks are expected to have linearly dependent shocks and
xhibit some level of feedback between these shocks. Also, if
he sources of shock to the region are common, then the struc-
ure of the variance decomposition should be similar across the
one.
Apart from contributing to the academic literature on mone-
ary integration in West Africa, the approach used in this paper
dds value to the previous studies in West Africa by measuring
he level of integration achieved by the participating countries in
erms of their response to common shocks. The methodologies
sed in the previous studies do not allow for the direct measure-
ent of supply, demand and monetary shocks to the economies
f the individual countries and their response to common shocks.
his will also inform policy on the adoption of the single cur-
ency, the eco, in the zone and also to have an idea of how the
conomies of the zone converge ex-ante or will converge ex-post
fter the introduction of the eco.
The question of what constitutes an optimum currency area is
ioneered by Mundell (1961) who defines an optimum currency
rea as a domain within which exchange rates are fixed. Within
his domain, a single currency can be introduced under a single
entral bank with the power to issue and redeem currency and
onduct monetary policy. The issue of an appropriate domain is
ddressed by Mundell (1961) by suggesting that the domain is
 region that is defined such that there is internal factor mobility
nd external factor immobility and “if factors are mobile across
ational boundaries, then a flexible exchange system becomes
nnecessary and may even be positively harmful”.
The work of Mundell (1961) inspired a series of papers.
n particular, McKinnon (1963) describes the optimum cur-
ency area as an area within which there is a single currency
nd within which the same monetary and fiscal policies and
exible external exchange rates can be used to address the objec-
ives of employment, international payments and price stability
hich are sometimes in conflict. McKinnon emphasizes the
eed for price stability within the region and the openness of
he economies that should be considered optimum for a single
urrency. McKinnon (1963) also added the importance of fac-
or mobility across industries to Mundell’s argument for factor
obility across countries in determining an optimum currency
rea.
The issue of factor mobility is further examined by Kenen
1969). He asserts that “when regions are defined by their activ-
ties, not geographically or politically, perfect interregional labor
obility requires perfect occupational mobility and this can only
ome about when labor is homogeneous” (Kenen, 1969). Kenen
1969) also advances product diversification and fiscal integra-
ion of a region as major criteria for an optimum currency area.
enen (1969) argues that diversity in a region’s product mix may
e a more relevant criterion than labor mobility and that well
iversified is economy is more likely to have a well-diversified l
xport sector, which can mitigate external shocks by positive U
nd negative shocks canceling out without resulting to exchange ovelopment Finance 5 (2015) 53–63
ate changes in response to the shock. Fiscal integration also
nsures that weaker economies within the region are supported
uring recovery from external shocks. Eichengreen (1991) also
efines an optimum currency area as “an economic unit com-
osed of regions affected symmetrically by disturbances and
etween which labor and other factors production flow freely.”
These characterizations of the optimal currency area in the
iterature usually lead to categorization of all the criteria into
hree. Firstly, the region should be subject to common sources
f shocks and symmetric response to shocks. This means that
hocks that are external to the region should induce the similar
esponses across the region, that is, the response of the states in
he region to external shocks must be similar to ensure that the
ame monetary and fiscal policies can address shock recovery
imilarly across the region. Since the introduction of a single
urrency in a region means that the countries that form the region
ive up their autonomy over monetary policy, their individual
bility to respond to external shocks by using monetary policy is
lso surrendered, therefore shock symmetry in the region ensures
hat common monetary policy is feasible for the region. “The loss
f monetary flexibility has cost and benefit. One hand, a country
hat gives up its currency loses a stabilization devise targeted
o domestic shocks, on the other hand, the country may gain
redibility and thereby reduce undesired inflation” (Alesina and
arro, 2002). Alesina et al. (2002) also argue that the costs of
osing monetary autonomy are lower when shocks are symmetric
cross that region.
Secondly, factor mobility within the region ensures that
hocks to the region dissipates quickly and similarly across.
actors must be easily movable from surplus members states
o deficit member states in the region in times where shocks to
he region have asymmetric effects. This ensures full employ-
ent and price stability in the region. Lastly, fiscal integration is
eeded in the region to redistribute resources among the member
tate. This is a system where fiscal policies of the different states
n the region are coordinated by a common federal institution
ike the IRS and congress of the United States. By this arrange-
ent, collection and disbursement of certain taxes are done by
ederal institution and in time economic downturn, weak states
an easily be bailed out through these arrangements.
As summarized by Bayoumi (1994), “the choice of a currency
nion depends upon the size of the underlying disturbances, the
orrelation between these disturbances, the costs of transactions
cross currencies, factor mobility across regions, and the interre-
ationships between demand for different goods.” So the obvious
uestion to ask is whether ECOWAS is an optimum currency
rea, that is, does the region satisfy the criteria for the introduc-
ion of a common currency? This is the question this study sets
o investigate.
.  Evolution  of  the  West  African  Monetary  Union  and
est African  Monetary  ZoneAccording to Soyibo (1998) before ECOWAS was estab-
ished in 1975, there were two monetary unions in West Africa.
nder British colonial rule, Anglophone West Africa made up
f Gambia, Ghana, British Cameroon, Nigeria and Sierra Leone
of De
u p
a (
G
b
r
t
N
B
i
F
a (
F
s
n (
e
(
d
r
z
c
p
1
(
p
f (
t
c
i (
c
e T
o m
c w
g T
b a
m a
i d
E a
c t
s p
A o
t i
a c
p w
W a
t a
( i
2 i
r
m 3
M
t
g h
a
d
p
S.K. Harvey, M.J. Cushing / Review 
sed a common currency, the British West African pound man-
ged by the West African Currency Board. However, when
hana gained independence in 1957 and establish her central
ank, the Bank of Ghana, she began issuing her own national cur-
ency the cedi in 1958. Nigeria also issued her national currency,
he naira, in 1958 with establishment of the Central Bank of
igeria to replace the British West African pound. By 1968, the
ritish West African pound collapsed when the other members
ssued their own currencies.
The francophone West Africa, made up of Benin, Burkina
aso, Côte d’Ivoire, Mali, Niger, Senegal, and Togo, also had
nd still has a common currency: the CFA franc inherited from
rance, the colonial rulers of these countries. The CFA franc
urvived the post independence collapse of monetary harmo-
ization, unlike the case for Anglophone West Africa, and
stablished the West African Economic and Monetary Union
WAEMU) in 1994 with a single central bank BanqueCentrale
es Etats de l’Afrique de l’Ouest (BCEAO) and a common cur-
ency (CFA) which was fully convertible within the French franc
one. The WAEMU countries have common monetary and fis-
al policies. Lending to government, for example, is fixed at 20
ercent of the estimated revenue of the previous year (Soyibo,
998).
The West African Clearing House (WACH), a multilateral
ayment system, was set up in 1975, immediately after the
ounding of ECOWAS to provide settlement services among
he central banks and to facilitate the monetary integration pro-
ess in the whole of West Africa. This has been transformed
nto West African Monetary Agency (WAMA) in 1996. A more
omprehensive program called the ECOWAS Monetary Coop-
ration Programme (EMCP) was launched in 1987 with its main
bjective of creating a single monetary zone and introducing a
ommon currency. The initial idea had been to introduce a sin-
le currency for all the member states of ECOWAS at a time,
ut this idea was later changed to the formation of a second
onetary zone with a single currency, called the eco, when
n April 2000 Accra Declaration four Anglophone members of
COWAS Gambia, Ghana, Nigeria, Sierra Leone and one fran-
ophone member Guinea launched an initiative to establish the
econd monetary zone in West Africa. December 2000 Bamako
ccord established the West African Monetary Zone (WAMZ),
he West African Monetary Institute (WAMI) and Stabilization
nd Cooperation Fund (SCF) alongside eight-member franco-
hone West African Economic and Monetary Union (WAEMU).
AMI was established by this accord to undertake all necessary
asks leading to the setting up of the West African Central Bank
WACB) and the introduction of a common currency (WAMI,
002). The five countries had pledged to adopt a common cur-
ency by January 2003 and to work toward merging their planned
onetary union with the WAEMU by January 2004 (Asante and
asson, 2001). These ambitious targets could not be met by
hese countries because of the failure to meet the set conver-
ence criteria. Liberia later joined the WAMZ in February 2010
s the sixth member with Cape Verde an observer. o
In November 2002 the Forum of Finance Ministers of WAMZ s
ecided to facilitate the harmonization of fiscal and monetary o
olicies by introducing two sets of convergence criteria, four avelopment Finance 5 (2015) 53–63 55
rimary and six secondary, for members. According to WAMI
2002), these criteria are as follows:
The primary criteria
(i) Achieve and maintain price stability by recording single
digit end of period inflation rate by 2003 and 5 percent by
2004.
(ii) Ensure sustainable government fiscal position by reducing
the ratio of budget deficit (excluding grants) on commit-
ment basis to GDP to 4 percent or less throughout the period
2003–2005.
iii) Limit Central Bank financing of government budget deficit
as a percent of previous year’s tax revenue to 10 percent or
less throughout the period 2003–2005.
iv) Maintain sufficient level of gross official foreign exchange
reserves of at least 3 months of import cover throughout the
period 2003–2005.
The secondary criteria
(i) Prohibition of new domestic arrears and liquidation of
existing ones.
(ii) Tax revenue to GDP ratio equal to or greater that 20 percent.
iii) Wage bill to tax revenue ratio equal to or less than 35
percent.
iv) Public investment to tax revenue equal to or greater than
20 percent.
(v) Maintain real exchange rate stability.
vi) Positive real interest rate.
he primary criteria would ensure that the economies of the
ember states converge in the sense of having symmetric shocks
hile the secondary criteria would ensure fiscal convergence.
hroughout the period 2001–2009 only two countries, Gambia
nd Nigeria, satisfy all four primary criteria in 2007 and 2008
nd only Gambia satisfies all criteria in 2008 and 2009. These
evelopments brought a lot of doubts about the possibility of
 successful introduction of a common currency, the eco, in
he WAMZ. Debrun et al. (2005), for example, show that the
roposed monetary union is not incentive compatible for most
f the existing non-CFA members of ECOWAS unless there are
nstitutional changes. A new time for the introduction of the new
urrency, by which it is hoped all the economies in the region
ill meet the convergence criteria, is 2015. These unsuccessful
ttempts at introducing the currency in the previous set dates
lso bring into focus the sustainability of the eco when it is
ntroduced since there are still staggering efforts at introducing
t.
.  Empirical  literature
The empirical testing of the optimum currency area criteria
as taken several forms including testing the synchronization
f business cycles of the members of the region and measuring
hock asymmetry of the countries in the region, indices, among
thers. Among the many studies that use the business cycles
pproach are Frankel and Rose (1996, 1997) and Kouparitas
5  of De
( a
i b
c i
m t
f t
a (
( t
p
C h
u t
h a
d
o i
a w
a N
d w
S i
t fi
v
e p
( b
m
t u
e c
b b
n t
d p
t a
t t
t t
o o
c a
t
w e
a v
a s
o w
d a
p m
o d
t r
fi
a d
n A
o p
o
t c
c s
D t
t
r 4
t
u 4
a
i
t
6 S.K. Harvey, M.J. Cushing / Review
2001), and the general understanding is that the cost of los-
ng autonomy over monetary policy will be lower if business
ycles of the countries in the region synchronizes. Measure-
ent of shock asymmetry using VAR has been done mainly
or Europe, and the major studies in this area include Bayoumi
nd Eichengreen (1992), Kempa (2002) and Buigut and Valev
2005).
Bayoumi and Eichengreen (1997) construct an Optimum
urrency Area (OCA) index for the European countries and
se that to divide the countries in euro zone into three groups:
igh level of readiness, tendency to converge and little or no evi-
ence to converge. Bayoumi and Ostry (2010) use correlation of
utput growth and inflation across countries in the regions and
lso regress real output per capita on its first and second lags
nd interpret the residuals to mean the underlying real output
isturbances and conclude that there is little evidence that sub-
aharan African countries would benefit from currency union in
he near future.
Unlike the European Union, studies on the ECOWAS mon-
tary integration are scanty. Since the seminal work by Soyibo
1998), little empirical work has been done on the ECOWAS
onetary integration process. Debrun et al. (2003) examine
he rationale for establishing regional currency unions in west-
rn Africa and conclude that monetary unification might well
e beneficial for a number of the member states of the Eco-
omic Community of West African States (ECOWAS) despite
ramatic economic, political and historical differences between
he two regions in the community. This is because the costs of
hese countries losing their monetary autonomy are often more
han offset by the gains originating in the (partial) separation
f monetary and fiscal powers. They argue however, that large
ountries with relatively ambitious public expenditure objec-
ives, like Nigeria, would not be attractive partners because they
ould be expected to pressure the common central bank, cre-
ting excessive inflation in the entire union. Based on those
rguments, they conclude that the desirability and sustainability
f a currency union within ECOWAS critically depend on fiscal
iscipline among its members and on a strong fiscal surveillance
rocedure both in the transition phase and after the establishment
f the union. Masson and Patillo (2003) conclude that “mone-
ary union in West Africa can be effective agency of restraint on
scal policies only if the hands of the fiscal policy authorities
re also tied by a strong set of fiscal restraint criteria, applicable
ot just for accession to monetary union, but throughout the life
f the union”.
Ogunkola (2005) uses real exchange rate model to analyze
he viability of a single monetary zone in ECOWAS and con-
ludes that ECOWAS is closer to a monetary union than before.
ebrun et al. (2005) also conclude, based on the calibration of
heir model, that lack of fiscal convergence, not the low level of
egional trade or asymmetry of shocks, is the primary obstacle
o the creation of a well-functioning and acceptable monetary
nion in West Africa. These two studies consider both the CFA
nd non-CFA zones and conclude based on the two zones form-
ng a single monetary union.
On the failure of the introduction of the single currency for s
hree consecutive times, Ojo (2005) notes that the failures are mvelopment Finance 5 (2015) 53–63
ttributable to inadequate political commitment, political insta-
ility and inability to sufficiently carry along all the stakeholders
n the process of program implementation. There is the need for
he common market program to be implemented to complement
he monetary integration program (Obaseki, 2005). Sagbamah
2005) highlights the important lessons of the European Union
hat should be learned by ECOWAS and provide the needed
olitical will, social enlightenment campaign and mobilization,
omogeneous product and financial markets, basic infrastruc-
ural production and economic structures, before transiting into
 monetary union.
Debrun et al. (2005) are of the view that fiscal heterogeneity
ndeed appears critical in shaping regional currency blocs that
ould be mutually beneficial for all their members. In particular,
igeria’s membership in the configurations currently envisaged
ould not be in the interests of other ECOWAS countries unless
t were accompanied by effective containment on Nigeria’s
nancing needs. But Iyare et al. (2005) note that while fiscal con-
ergence among members is desirable, other mechanisms like
ayment systems and labor mobility issues should be established
eyond fiscal convergence, if such a union is to be successful.
Balogun (2007) estimates a dynamic panel data model
sing data available on West African Monetary Zone (WAMZ)
ountries and examines the monetary and macroeconomic sta-
ility perspective for entering into monetary union. By testing
he hypothesis that independent monetary and exchange rate
olicies have been relatively ineffective in influencing domestic
ctivities (especially GDP and inflation), and that when they do,
hey are counterproductive, he concludes that the members of
he WAMZ would be better off surrendering their independence
ver some policy instruments to the planned regional body under
ppropriate monetary union arrangements.
Balogun (2009) examines the determinants of inflation differ-
ntials in a panel of West African Monetary Zone (WAMZ) states
is-à-vis its set benchmark for macroeconomic convergence
ince 2000. Over the sample period, he finds that the un-
eighted average regional inflation rates were most often above
 single digit target and vary widely among the countries. The
ajor monetary policy instruments determinants of inflationary
ivergence are the pursuit of distorted interest rates, exchange
ates overvaluation and expansionary monetary policies.
It is clear from the empirical evidence that the ex ante con-
itionality for the introduction of the single currency in West
frican Monetary Union will be difficult to achieve. But it is
ossible, like the UEMOA countries, for the WAMZ to achieve
ptimality ex post. Achieving ex post optimum currency area
an be ensured if there are common sources of shock and shock
ymmetry across the region and that is what this study sets out
o measure.
.  Methodology
.1.  The  SVAR  modelThe empirical analysis of shock asymmetry is based on the
tochastic rational expectations open economy macroeconomic
odel developed by Clarida and Gali (1994) and also used by
of De
K s
c
r y
b
m c
t
A
p
w
t T
r s
i i
p B
i ∑i
Γ
⎛
w
a ⎜⎝
i
g
I
y
t
y
w
w
T
y a
w i
t t
a t
s
( C
y l
s
p
y S
b
N d
( l
b t
b a
s
y r
a s 2t
p tary shocks v3t have no long run impact on either change in real
output levels or real exchange rates” (Clarida and Gali, 1994).
y
f
e
S.K. Harvey, M.J. Cushing / Review 
empa (2002) to analyze the convergence of the euro zone
ountries an optimum currency area. “The model exhibits the
esults of the standard Mundell-Fleming-Dornbusch model in
oth the short run when prices adjust sluggishly to demand,
oney and supply shocks and long run properties that charac-
erize macroeconomic equilibrium in the open economy once
rices adjust full to all shocks” (Clarida and Gali, 1994).
′
Let yt = (x1tx2tx3t) be a vector of endogenous variables
here x1t is a measure of growth of economic activity of a coun-
ry relative to the US, x2t is the change in bilateral real exchange
ates between each country’s currency and the US dollar and x3t
s the change in price level of each country relative to the US
rice level. The dynamic structural representation of the model
s
∑p
yt =  μ  + Btyt−j +  vt (1)
j=1
here Γ  is a 3 ×  3 matrix of contemporaneous coefficients
mong the endogenous variables, μ  is a vector of constants, Bj
s a 3 ×  3 matrix of structural coefficients, vt i∑s a vector( of or)tho-′
onal structural shocks to the system so that v =  E vtvt =
. The reduced∑ form of Eq. (1) is
= −1 −1 −1t  Γ μ  + Γ Btyt−j +  Γ vt (2)
his can be written as
∑p
t =  ν  + Θjyt−j +  et (3)
j=1
here Γ−1μ,Θ = Γ−1 j  j Bj , and etΓ−1vt . Eq. (3) can also be
ritten as
t =  ν  +  Θ(L)yt +  et (4)
here Θ(L) = 2 p L  + L + · · · + L and L is a lag operator. Given that
he system in Eq. (4) is stable, we can re-write (4) as a moving
verage representation, by Wold’s decomposition.
I −  Θ(L)) yt =  ν  +  et (5)
t(I  −  Θ(L))−1ν  + (I  −  Θ(L))−1et (6)
∑∞
t =  μ0 + Φjet−j (7)
j=0
ow suppose, as in Blachard and Quah (1989), Clarida and Gali
1994) and Kempa (2002) that the estimated MA representation,
ased on estimation of the reduced form equation in (4), is given
y
t =  et +  C1et−1 +  C2et−2 +  · ·  · (8)
nd the true MA representation of the actual data generating
rocess is
t =  A0vt +  A1vt−1 +  A2vt−2 +  · ·  · (9) T
rom Eq. (3) o
= −1t a Γ vt (10)velopment Finance 5 (2015) 53–63 57
ubstituting Eq. (10) to Eq. (8) gives
= Γ−1t  vt + C −1 1Γ vt−1 +  C2Γ−1vt−2 +  ·  · ·  (11)
omparing Eqs. (9) and (11) give us
= Γ−10  ,  et =  A0vt for j  =  0 Aj =  CjA0 for j  >  0
(12)
his shows the relationship between the vector of structural
hocks vt and the vector of reduced form residuals et, which
s equivalent to the C-model of Amisano and Giannini (1997).
y knowing A0, we can recover the structural shocks from the
nnovations. From Eq. (12) we can write
′ ′ ′ ′
E(etet) =  A0E(vtvt)A0 =  A0A0 (13)e
⎞ ⎛ ⎞⎛ ⎞
w11 w21 w31 ⎟ ⎜ a110 a210 a310 ⎟⎜ a110 a120 a130 ⎟
w12 w22 w32 ⎠ =⎝ a120 a220 a320 ⎠⎝ a210 a220 a230 ⎠
w13 w23 w33 a130 a230 a330 a310 a320 a330
w11 =  a2 2 2110 +  a210 +  a310
w22 = 2 2 2 a120 +  a220 +  a320
w 2 2 233 =  a130 +  a230 +  a330
w21 =  w12 +  a120a110 +  a220a210 +  a320a310
w31 =  w13 +  a130a110 +  a230a210 +  a330a310
w32 =  w23 +  a130a120 +  a230a220 +  a330a320 ∑
his is a system of 6 equations with 9 unknowns since e is
 symmetric matrix estimated from the VAR in Eq. (3), this
mplies that A0 is not identified. In order to be able to iden-
ify A0 and recover the structural shocks vt we need to impose
hree additional restrictions on the elements of A0. In this paper
ince the structural model derives from the structural model of
larida and Gali (1994), the alignment of the shocks also fol-
ows. The shock to economic growth is aligned as the supply
hock because supply shocks are known to be the main unex-
ected changes in output in developing countries especially.
hocks to real exchange rates are identified as demand shocks
ecause these countries are import-dependent, so excess import
emand drives exchange rates. Shocks to price changes are also
abeled as monetary shocks because money is assumed to be neu-
ral. Clarida and Gali (1994) and Kempa (2002) use Blachard
nd Quah (1989) decomposition to identify A0. This decompo-
ition states that “only supply shocks v1t influence changes in
eal output levels in the long run, while both supply and demand
hocks v influence real exchange rates in the long run. Mone-his statement imposes three restrictions on A0. A short run view
f these restrictions is
210 =  a310 =  a320 =  0
5  of De
g X
Q U
r n
∑ b
r
n
a
∑ d
e
d
i t
t c
v
4
e
s
o I
S
4 d
s s
i
e
d l
b G
b r
s G
a t
c S
F D
f
w t
a t
m
s
x e
s
n
x a
a
s
y e
i
r
T
y d
t
C
x
(exports +  imports)US
a
The exchange rate variables are bilateral real exchange rates
F of the countries’ currencies to the US dollar. The real exchange
F
8 S.K. Harvey, M.J. Cushing / Review
iven the arrangement of the variables in the VAR. Blachard and
uah (1989) restrictions are long run restriction; therefore the
estrictions imply that
∞ ∑∞ ∑∞
a21i = a31i = a32i =  0
i=0 i=0 i=0
These restrictions imply that the matrix
∞ ∑∞
Aj = CjA0
j=0 j=0
s a lower triangular matrix. Blachard and Quah (1989) show
hat these restrictions identify A0 and we can recover vt as
t = −1 A0 et
In order to ensure the stability of the VAR and be able to
xplore it’s properties I check the stationarity properties of the
eries using Augmented Dickey–Fuller test. The optimum lag
rder selection is based on Akaike Information Criterion.
.2.  Linear  dependence  of  and  feedback  between  the
tructural shocks
The linear dependence of two time series xt and yt can be
ecomposed into a sum of contemporaneous linear feedback
etween xt and yt, linear feedback from xt to yt and linear feed-
ack from xt to yt. Geweke (1982) shows that if the series are
tationary, nondeterministic, autoregressive and have moving
verage representation, then linear dependence of xt and yt (FX,Y)
an be decomposed as
X,Y = F̂X→Y + F̂Y→X + F̂XY (14)
here FX→Y, FX→Y, and FX→Y are calculated from the variances
nd covariance of the residuals in the following autoregressive
odels:
∑p
= E x − + u Σ̂ = Û ′t 1s t s  1t 1 1Û1 (15)
s=1
∑p ∑p
t = E ′2sxt−s + E2syt−s +  u1t Σ̂2 = Û2Û2 (16)
s=1 s=1
∑p
t = G1syt−s + ′ v1t T̂1 = V̂1V̂1 (17)
s=1
∑p ∑p
t = ′G2syt−s + H2sxt−s +  v2t T̂2 = V̂2V̂2 (18)
s=1 (s=1 )
Σ̂2 Ĉ
ˆ ′
 = Û2V̂2 Υ̂ = ′ (19)
Ĉ T̂2
nd ( | ) ( )= T | |Σ | ln 1 1, F =  ln ,X→Y
( |T2|
Y→X |T2| r
= |T2|.|Σ2|
) ( |Σ2|.| |)T1
X.Y  ln | ,  FX,Y =  ln (20)Υ x | |Υ  |velopment Finance 5 (2015) 53–63
 and Y  are linearly independent if and only if Σ1 = Σ2.
nder the null hypothesis of no linear feedback where
F̂X,Y ,  nF̂X→YnF̂Y→X, and nF̂X.Y have chi-squared distri-
ution with degrees of freedom kl(2p  + 1), klp, klp, and kl,
espectively, where k is the number of variables in xt, l is the
umber of variables in yt, p is the number of autoregressive lags
nd n  is the number of observations.
Geweke (1982) is used to measure and decompose linear
ependence between pairs of countries and compared with lin-
ar dependence of Germany, France, Spain, Italy and Greece to
etermine if the West African countries are ready for a mone-
ary union. We expect the structural shocks of economies that
onverge to be linearly dependent.
.3.  Data
The data for the estimation of the models are extracted from
nternational Financial Statistics (IFS) and Direction of Trade
tatistics (DOTS) by the International Monetary Fund (IMF) and
irectly from some central banks and statistical organizations of
ome of the countries. The data on consumer price index and
nflation for all the countries are extracted from October 2011
dition of IFS, except for Guinea and where these data are col-
ected from the website of BanqueCentrale de la République de
uinée (BCRG) (bcrg-guinee.org). Data on nominal exchange
ates of each of the countries are taken from the IFS except for
uinea where they are taken from the IFS for 1980 to 2005 and
he rest of the years from the website of Institut National de la
tatistique (stat-guinee.org) while the trade data are taken from
OTS. I use monthly data from February 1987 to April 2011
or all the series. The period of the data for the paper is chosen
o insure that the data are available for all the variables for all
he countries in the study.
In measuring the variables that go into the models, many
tudies use real GDP growth as a measure of real growth of the
conomic activity but in the context of developing countries
uch as the ECOWAS countries Bayoumi and Ostry (2010)
otes that “in Africa many of the shocks which affect economies
re temporary supply disturbances such as climatic shocks to
griculture or terms of trade disturbances”. This is due to the sub-
istence nature of agriculture, which is the dominant sector in the
conomies of many of these countries. In this study I use growth
n total trade of each these countries, that is, exports plus imports
elative to US trade to measure growth of economic activity.
he use of the trade data also makes it possible to use monthly
ata which increases frequency and range of the data. Therefore
he real growth for country i in the region is measured as
( )
= (exports +  imports)1i  dln i .ate for country i is measured as
2i =  d(ei −  pi +  pus)
S.K. Harvey, M.J. Cushing / Review of Development Finance 5 (2015) 53–63 59
Table 1
Variance decomposition of the variables in the model.
Period Gambia Ghana Guinea Nigeria Sierra Leone
Monetary Demand Supply Monetary Demand Supply Monetary Demand Supply Monetary Demand Supply Monetary Demand Supply
Variance decomposition of real growth
1 0.01 0.34 99.65 0.09 1.94 97.97 0.41 0.28 99.31 0.1 0.19 99.71 0.04 9.23 90.73
3 0.31 0.49 99.2 0.14 2.4 97.47 1.49 0.75 97.76 0.24 0.73 99.04 0.12 12.03 87.84
6 0.31 0.49 99.2 0.48 2.59 96.93 1.5 0.86 97.64 0.24 0.73 99.04 0.3 12.59 87.11
12 0.31 0.49 99.2 0.49 2.61 96.9 1.5 0.86 97.64 0.24 0.73 99.04 0.31 12.59 87.1
24 0.31 0.49 99.2 0.49 2.61 96.9 1.5 0.86 97.64 0.24 0.73 99.04 0.31 12.59 87.1
36 0.31 0.49 99.2 0.49 2.61 96.9 1.5 0.86 97.64 0.24 0.73 99.04 0.31 12.59 87.1
Variance decomposition of real exchange rate changes
1 0.425 99.211 0.364 1.908 96.042 2.05 0.485 97.951 1.564 0.375 99.605 0.02 7.85 85.392 6.758
3 1.689 96.976 1.335 1.35 94 4.651 0.916 97.234 1.85 0.549 98.47 0.98 11.685 81.941 6.374
6 1.716 96.95 1.334 1.439 92.696 5.865 0.931 96.349 2.72 0.553 98.402 1.044 12.505 81.093 6.402
12 1.716 96.95 1.334 1.61 92.454 5.936 0.931 96.348 2.721 0.553 98.402 1.044 12.535 81.061 6.404
24 1.716 96.95 1.334 1.61 92.452 5.938 0.931 96.348 2.721 0.553 98.402 1.044 12.536 81.06 6.404
36 1.716 96.95 1.334 1.61 92.452 5.938 0.931 96.348 2.721 0.553 98.402 1.044 12.536 81.06 6.404
Variance decomposition of price changes
1 91.971 7.785 0.244 90.307 9.187 0.506 33.573 64.728 1.699 95.6 4.226 0.174 84.02 10.064 5.917
3 89.896 8.921 1.183 92.524 7.036 0.44 33.806 64.265 1.929 94.517 4.698 0.785 83.165 9.923 6.912
6 89.696 9.12 1.185 91.519 7.652 0.829 33.512 63.687 2.801 94.503 4.711 0.786 82.253 10.806 6.941
12 89.696 9.12 1.185 91.196 7.931 0.873 33.511 63.686 2.803 94.503 4.711 0.786 82.194 10.833 6.973
2 63
3 63
w d
i c
t e
a
w v
p
x t
t
h
U d
s r
e
5 l
1
p
m a
A r
a o
C p
2 a
G o
f
5 s
a
c
s c
c
s 5
t
g
c
2
4 89.696 9.12 1.185 91.195 7.931 0.874 33.511 
6 89.696 9.12 1.185 91.195 7.931 0.874 33.511 
here ei is the log of bilateral nominal exchange rate for country
, pi is the log of CPI of country i  and pUS is the log of CPI of
he USA.
The price variable is measured according to Kempa (2002)
here ( )
= pi3i  d  ln .
pUS
These variables are measured relative to the USA because the
S dollar is seen as an anchored currency of these countries as
hown in Alesina et al. (2002).
.  Empirical  results
The measurement of the variables that go into the models
akes them naturally to be stationary at their levels. However,
DF tests are used to formally check the stationarity properties
nd found that they do not have unit roots. Akaike Information
riterion is used to select the optimum lag for the models and
, 5, 3, 2, and 3 lags are found to be optimum for the models of
ambia, Ghana, Guinea, Nigeria and Sierra Leone, respectively.
.1.  Variance  decomposition
Variance decomposition is important in identifying the
ources of variability in the variables in the models for each
ountry in the region. This helps in determining whether the
ources of shock to variables in the models are common across
he region. Table 1 shows the variance decomposition of real
rowth of economic activity, real exchange rates and price level
hanges. The variance decomposition is presented for 1, 3, 6, 12, c
4 and 36 lags to enable us compare the structure of the variance f.686 2.803 94.503 4.711 0.786 82.194 10.833 6.972
.686 2.803 94.503 4.711 0.786 82.194 10.833 6.972
ecomposition after the system stabilizes across. Panels a, b and
 of Table 1 show the variance decomposition of real growth, real
xchange rate change and price level changes, respectively for
ll the countries. Panel a shows that the supply shocks dominate
ariability in output growth throughout, explaining at least 99
ercent for Gambia, Ghana and Nigeria. While the structure of
he variance decomposition looks similar for these countries in
he region, the magnitude looks different for Sierra Leone with a
igher percentage of the variability in real growth explained by
emand shocks. Panel b of shows that for all the countries in the
egion, demand shocks are dominant and persistent over time in
xplaining exchange rate variability. Demand shocks explain at
east 97 percent of real exchange rate across the region for the
-month ahead forecast variance and this stabilizes after third
eriod except for Sierra Leone. In panel c, the pattern of the vari-
nce decomposition for price level changes is different across the
egion. At 36 lags, after the system stabilizes, about 90 percent
f the variance for Gambia is explained by monetary shocks, 91
ercent for Ghana, 34 percent for Guinea 95 percent for Nigeria
nd 82 percent for Sierra Leone. Clearly, the forecast variance
f prices is explained by different shocks across the zone except
or Ghana and Nigeria which are close for all the variables. This
uggests that the sources of external shock to real output growth
nd real exchange rates in the region are common to the four
ountries but the sources of shock to price level changes are not
ommon to any.
.2.  Impulse  response  functionsImpulse response functions of all the variables for all the
ountries in the region, which are not shown in the paper, show
or each country the response of real growth to a 1-standard
6  of De
d s
r i
r t
d (
t l
r G
t w
s e
L t
t a
o o
s d
s t
t c
t s
t Z
s
p
5 p
s a
c
T
R
L
G
G
G
N
S
F
G
G
I
S
C
G
G
G
N
S
F
G
G
I
S
L
G
G
G
N
S
F
G
G
I
S
0 S.K. Harvey, M.J. Cushing / Review
eviation of supply, demand and monetary shocks. Also, the
esponse of real exchange rate changes and price level changes,
espectively to the same shocks is drawn. The graphs display the
ynamics of how the variables respond to the shocks. On impact,
he economies of all the countries in the region shrink at different
ates. The time it takes for the shock to dissipate varies among
he countries. While it takes Gambia about 6 periods for the
hock to dissipate, it takes Ghana about 5 periods, while Sierra
eone’s shock lingers through to the 10th period. The response
o supply and monetary shock is not similar either. The response
f real exchange rates changes and price level changes to all
hocks across the region differs greatly in structure and inten-
ity. The differences in the rate at which the shocks dissipate
hrough impulse response functions of the countries show that
hese countries have asymmetric shocks which is further inves-
igated in the next section using a correlation of the structural
hocks.
.3.  Linear  dependence  of  and  feedback  between  the
tructural shocks
c
A measure of the level of convergence r between the
ountries, in the sense of an optimum currency area, is shock S
able 2
elationship between supply shocks.
Gambia Ghana Guinea Nigeria Sierra
inear dependence
ambia
hana 0.025
uinea 0.049 0.044
igeria 0.018 0.057 0.007
ierra Leone 0.029 0.002 0.024 0.018
rance 0.065 0.087 0.046 0.051 0.051
ermany 0.040 0.026 0.013 0.035 0.060
reece 0.041 0.030 0.033 0.075 0.051
taly 0.071 0.077 0.058 0.047 0.045
pain 0.057 0.038 0.045 0.077 0.039
ontemporaneous linear feedback
ambia
hana 0.004
uinea 0.008 0.014
igeria 0.004 0.044 0.002
ierra Leone 0.010 0.000 0.000 0.000
rance 0.010 0.008 0.006 0.004 0.010
ermany 0.036 0.015 0.004 0.010 0.026
reece 0.010 0.003 0.011 0.052 0.009
taly 0.019 0.005 0.014 0.009 0.002
pain 0.001 0.005 0.002 0.007 0.001
inear feedback
ambia 0 0.00 0.02 0.01 0.02 
hana 0.02 0 0.02 0.01 0.00 
uinea 0.02 0.01 0 0.00 0.01 
igeria 0.00 0.00 0.00 0 0.01 
ierra Leone 0.00 0.00 0.01 0.01 0 
rance 0.00 0.05 0.01 0.01 0.00 
ermany 0.00 0.01 0.01 0.00 0.01 
reece 0.00 0.00 0.01 0.01 0.01 
taly 0.01 0.03 0.01 0.00 0.01 
pain 0.00 0.00 0.00 0.02 0.00 velopment Finance 5 (2015) 53–63
ymmetry between underlying structural shocks of the countries
n the region. The size and correlation of the underlying dis-
urbances are important for the choice of a currency union
Bayoumi, 1994). If two economies converge, we expect under-
ying disturbances to be linearly dependent, in the sense of
eweke (1982), because their response to the external shocks
ill be similar. Shocks are symmetric if and only if they are lin-
arly dependent in this sense. The existence of feedback between
he shocks of these countries in a region suggests that they have
 mechanism to correct any imbalances that will arise as a result
f external shocks that are specific to any of them. Also, linear
ependence of shocks ensures that common policies transmit to
hese countries similarly. In this section we discuss the empiri-
al results of measuring linear dependence and feedback of the
tructural shocks of the countries in the West African Monetary
one (WAMZ) using Geweke (1982).
In order to put the results of the WAMZ countries in proper
erspective, these same measures are computed for five Euro-
ean countries in the euro zone: France, Germany, Greece, Italy
nd Spain. These countries are chosen to include all the different
haracters within the euro zone presently. France and Germany
emain strong after the introduction of the euro while Italy and
pain are troubled and Greece is at the brink. The idea is to
 Leone France Germany Greece Italy Spain
 0.358
 0.218 0.291
 0.500 0.276 0.284
 0.254 0.138 0.220 0.213
 0.263
 0.064 0.255
 0.474 0.182 0.115
 0.200 0.054 0.073 0.163
0.05 0.00 0.03 0.04 0.05
0.03 0.00 0.03 0.04 0.03
0.03 0.00 0.01 0.04 0.04
0.03 0.02 0.01 0.03 0.05
0.04 0.03 0.03 0.03 0.04
0 0.01 0.01 0.02 0.05
0.08 0 0.02 0.08 0.08
0.15 0.01 0 0.14 0.14
0.00 0.01 0.03 0 0.04
0.00 0.00 0.01 0.01 0
S.K. Harvey, M.J. Cushing / Review of Development Finance 5 (2015) 53–63 61
Table 3
Relationship between demand shocks.
Gambia Ghana Guinea Nigeria Sierra Leone France Germany Greece Italy Spain
Linear dependence
Gambia
Ghana 0.006
Guinea 0.068 0.512
Nigeria 0.042 0.035 0.015
Sierra Leone 0.024 0.441 0.081 0.079
France 0.054 0.558 0.009 0.019 0.245
Germany 0.037 0.021 0.041 0.013 0.025 0.470
Greece 0.026 0.541 0.012 0.014 0.178 0.101 0.107
Italy 0.094 0.536 0.007 0.011 0.234 0.070 0.155 0.102
Spain 0.074 0.540 0.019 0.008 0.249 0.155 0.139 0.090 0.163
Contemporaneous linear feedback
Gambia
Ghana 0.000
Guinea 0.000 0.000
Nigeria 0.014 0.000 0.002
Sierra Leone 0.011 0.006 0.000 0.070
France 0.000 0.001 0.000 0.001 0.006
Germany 0.000 0.002 0.000 0.000 0.000 0.399
Greece 0.006 0.002 0.000 0.001 0.004 0.021 0.059
Italy 0.014 0.000 0.002 0.001 0.000 0.041 0.061 0.028
Spain 0.010 0.000 0.001 0.000 0.000 0.120 0.072 0.008 0.154
Linear feedback
Gambia 0 0.000 0.007 0.021 0.007 0.049 0.008 0.020 0.051 0.050
Ghana 0.005 0 0.511 0.019 0.399 0.538 0.003 0.527 0.533 0.539
Guinea 0.061 0.001 0 0.001 0.001 0.002 0.021 0.008 0.002 0.014
Nigeria 0.007 0.016 0.012 0 0.005 0.006 0.001 0.004 0.008 0.005
Sierra Leone 0.005 0.036 0.080 0.004 0 0.236 0.021 0.171 0.232 0.243
France 0.006 0.019 0.007 0.011 0.002 0 0.010 0.000 0.016 0.026
Germany 0.028 0.016 0.020 0.012 0.003 0.061 0 0.026 0.048 0.051
Greece 0.001 0.013 0.004 0.008 0.003 0.080 0.023 0 0.054 0.053
I 0.002
S 0.005
l s
a i
e z
r c
o c
a c
t s
a
t d
o
e d
c c
f p
c c
d h
r i
t N
p m
i e
5
e
a
taly 0.029 0.003 0.003 0.002 
pain 0.014 0.002 0.005 0.003 
ook at the coefficients of linear dependence and feedback that
re calculated for these European countries against their current
conomic performance and use that information to discuss the
esults of the WAMZ countries. We mostly use the coefficients
f France and Germany to indicate most convergent economies
nd the coefficient for Greece or Italy with France or Germany as
he least convergent economies. We discuss linear dependence
nd feedback for supply, demand and monetary shocks across
he WAMZ.
Tables 2–4 contain coefficients that measure linear depend-
nce, contemporaneous feedback and feedback between
ountries in both directions. The tables contain the coefficients
or supply, demand and monetary shocks, respectively. The
oefficients at the upper part of the table are measures of linear
ependence; those at the middle part are measures of contempo-
aneous linear feedback. The lower diagonal of the lower part of
he table contains coefficients that measure feedback from sup-
ly shocks of countries in the row to supply shocks of countries
n the column. The upper diagonal does the reverse feedback..3.1. Supply  shocks
In Table 2, France and Italy have the strongest linear depend- 5
nce of 0.500 while France and Germany have 0.358. Spain has
 coefficient of 0.139 with Germany and 0.213 with Italy, this d 0.012 0.047 0.020 0 0.004
 0.009 0.017 0.029 0.005 0
ets an upper limit of 0.500 and a lower limit of 0.139 for judg-
ng the convergence of the WAMZ countries relative to the euro
one. As shown in Table 2, none of the WAMZ countries has a
oefficient with any other that fall within this interval. All these
ountries fall far outside the range, even Ghana and Nigeria’s
oefficient of 0.057, which is the largest in the zone falls far
hort of the interval.
Using similar arguments for establishing intervals for linear
ependence, the intervals for contemporaneous linear feedback
f the supply shocks is 0.474–0.054. Clearly, from the mid-
le segment of Table 2, all the WAMZ countries have feedback
oefficients that are below the lower limit. These weak contem-
oraneous feedbacks imply that policies implemented in each
ountry will have no effect in other and common policies will
ave different effects. For example, high rate of unemployment
n Ghana will not be reduced by increased industrial activity in
igeria and an expansionary monetary policy across the zone
ight be inflationary in Ghana and contained by the increased
conomic activity in Nigeria..3.2.  Demand  shocks
In Table 3, France and Germany have the strongest linear
ependence of 0.470 while France and Italy have the lowest of
62 S.K. Harvey, M.J. Cushing / Review of Development Finance 5 (2015) 53–63
Table 4
Relationship between monetary shocks.
Gambia Ghana Guinea Nigeria Sierra Leone France Germany Greece Italy Spain
Linear dependence
Gambia
Ghana 0.007
Guinea 0.034 0.106
Nigeria 0.031 0.083 0.018
Sierra Leone 0.012 0.080 0.027 0.035
France 0.130 0.172 0.035 0.040 0.061
Germany 0.076 0.079 0.025 0.037 0.016 0.219
Greece 0.017 0.143 0.005 0.012 0.073 0.052 0.089
Italy 0.136 0.120 0.018 0.036 0.048 0.312 0.262 0.073
Spain 0.055 0.115 0.034 0.010 0.049 0.214 0.072 0.073 0.226
Contemporaneous linear feedback
Gambia
Ghana 0.002
Guinea 0.002 0.001
Nigeria 0.018 0.001 0.010
Sierra Leone 0.001 0.001 0.002 0.023
France 0.062 0.059 0.007 0.022 0.002
Germany 0.029 0.028 0.006 0.022 0.003 0.160
Greece 0.006 0.002 0.000 0.001 0.016 0.000 0.043
Italy 0.070 0.010 0.000 0.024 0.000 0.280 0.193 0.009
Spain 0.010 0.002 0.003 0.001 0.000 0.209 0.035 0.006 0.206
Linear feedback
Gambia 0 0.005 0.003 0.012 0.005 0.044 0.000 0.011 0.046 0.041
Ghana 0.000 0 0.102 0.033 0.079 0.110 0.027 0.128 0.107 0.114
Guinea 0.030 0.003 0 0.002 0.002 0.024 0.006 0.002 0.010 0.025
Nigeria 0.001 0.049 0.006 0 0.002 0.015 0.002 0.003 0.007 0.008
Sierra Leone 0.006 0.000 0.023 0.010 0 0.044 0.011 0.045 0.045 0.045
France 0.024 0.003 0.003 0.003 0.014 0 0.006 0.002 0.028 0.000
Germany 0.046 0.023 0.013 0.013 0.002 0.053 0 0.012 0.058 0.033
Greece 0.001 0.013 0.003 0.008 0.012 0.050 0.033 0 0.058 0.045
I 0.002
S 0.004
0 o
G t
o W
t m
d o
s t
c a
l c
E t
c
5 n
r
l G
c s
c s
e r
W m
i a
T
6 t
(
taly 0.021 0.003 0.008 0.005 
pain 0.003 0.000 0.006 0.001 
.070 that give the upper and lower limits, respectively. Ghana-
uinea and Ghana-Sierra Leone have strong linear dependence
f 0.512 and 0.441, respectively. These are stronger than all
he European countries including France-Germany. In terms of
emand shocks, Ghana and Guinea seem to have high shock
ymmetry.
Contemporaneous linear feedback is weak among the WAMZ
ountries and is virtually zero for all the WAMZ countries. The
inear feedback is not much different between WAMZ and the
uropean countries.
.3.3.  Monetary  shocks
An equivalent interval derived from Table 4 for Gewekey
inear dependence is 0.312–0.052. Only Ghana-Guinea coeffi-
ient of 0.101 falls within this interval. None of the WAMZ
ountries fall within the interval for the contemporaneous lin-
ar dependence. The feedback between monetary shocks of the
AMZ countries is as strong as that of the European countries,
ncluding France and Germany..  Conclusions
a
This paper investigates whether West African Monetary Zone u
WAMZ) is a common currency area by using a structural VAR m 0.004 0.011 0.006 0 0.009
 0.004 0.004 0.023 0.011 0
f real growth, real exchange rates and price level of five of
he six countries in the West African Monetary Zone (WAMZ).
AMZ is a smaller group of countries within Economic Com-
unity of West African States (ECOWAS) that is in the process
f introducing a single currency the eco. The identification of
he structural shocks is based on Blachard and Quah (1989).
The evidence from the variance decomposition of the vari-
bles in the SVAR suggests that the region does not have
ommon sources of shock. Also, the impulses response func-
ions and the analysis of the structural shocks suggest that the
ountries in the region do not respond symmetrically to all exter-
al shocks. These suggest lack of ex-ante convergence in the
egion to form an optimum currency area. However, Ghana and
uinea seem to be close, both in commonness of sources of
hock and symmetry of shocks, and may be able to cope with a
ingle currency since the sources of shocks and shock recovery
ate between them is somehow similar. A piecemeal approach to
onetary union may be adopted where Ghana and Guinea adopt
 single currency and the other countries ascend to it over time.
his arrangement, however, may have serious implications for
he other countries that are not in the union on the onset because
s shown by Bayoumi (1994), while the gains from the monetary
nion in the form of lower transaction costs are limited to the
embers, the losses from the union in the form of lower output
of De
a B
B
w B
s
B
C
d
S B
A
c B
t
m B
o
C
e
O
e D
m
D
t E
e F
t
D F
p
G
s
t I
s
i K
K
g K
i
f
n M
t
M
M
R
O
A
A O
A
O
A
S
B
S
B International Payments system within ECOWAS Countries, held in Accra-
B
S.K. Harvey, M.J. Cushing / Review 
ffect every country in the region. Unlike the Eurozone where
ayoumi and Eichengreen (1997) find a core group of countries
ithin the union that is a common currency area, the findings
uggest that WAMZ does not have such a group.
These results confirm some previous studies on Economic
ommunity of West African States (ECOWAS) and contra-
ict others. The results of Bayoumi and Ostry (1997) on
ub-Saharan Africa “indicate little evidence that Sub-Saharan
frican countries would benefit in the near future from larger
urrency unions” but Debrun et al. (2003) conclude that mone-
ary union in ECOWAS might be beneficial for a number of the
ember states. Debrun et al. (2005) also conclude that because
f the fiscal heterogeneity of the countries in the union, Nigeria
specially might not be compatible with the rest of the countries.
gunkola (2005) also concludes that further convergence of the
conomies in the region is required for a stable region-wide
onetary union in West Africa.
The results, however, contradict Balogun (2007) that argues
hat the countries of the WAMZ are better off surrendering their
conomies to a common monetary policy. This directly suggests
hat these countries are better of with a common currency. Also,
ebrun et al. (2005) argue that asymmetric shocks are not the
roblem but lack of fiscal convergence. Even though the current
tudy is on a sub set of ECOWAS, we can interpret the results
ogether with Debrun et al. (2005) to mean that both asymmetric
hocks and lack of fiscal convergence are the obstacles to the
ntroduction of the common currency.
Lessons from the current euro crisis, suggest that fiscal inte-
ration should precede the introduction of a single currency even
f the region satisfies all the other criteria, which is not the case
or WAMZ. There is also the need for further policy harmo-
ization and removal of barriers to factor mobility to enable
ransmission of shocks through these economies to synchronize.
eferences
lesina, A., Barro, R.J., 2002. Currency unions. Q. J. Econ. 117 (2), 409–436.
lesina, A., Barro, R.J., Tenreyro, S., 2002. Optimal Currency Areas. NBER
Working Paper, 9072.
misano, G., Giannini, C., 1997. Topics in Structural VAR Econometrics.
Springer.
sante, R.D., Masson, P.R., 2001. Point/counterpoint: the pros and cons of
expanded monetary union in West Africa. Finance and development. Q.
Mag. IMF 38 (1).
alogun, E.D., 2009. Inflation Differential in the West African Monetary Zone
(WAMZ) Area: Implications for Unionization MPRA Paper No. 13045,
Online at http://mpra.ub.uni-muenchen.de/13045/
alogun, E.D., 2007. Monetary Policy and Economic Performance of
West African Monetary Zone Countries. MPRA Paper, 4308, Online at
http://mpra.ub.uni-muenchen.de/4308/ W
ayoumi, T., 1994. A formal model of optimum currency areas. Staff Pap. Int.
Monet. Fund 41 (4), 537–554.velopment Finance 5 (2015) 53–63 63
ayoumi, T., Eichengreen, B., 1992. Shocking Aspects of Monetary Unification.
NBER Working Paper, 3949.
ayoumi, T., Eichengreen, B., 1997. Ever closer to heaven? An optimum-
currency-area index for European Countries. Eur. Econ. Rev. 41 (1997),
761–770.
ayoumi, T., Ostry, J.D., 2010. Macroeconomic shocks and trade flows within
Sub-Saharan Africa: implications for optimum currency arrangements. J.
Afr. Econ. 6 (3), 412–444.
ayoumi, T., Ostry, J.D., 1997. Macroeconomic shocks and trade flows within
Sub-Saharan Africa: implications for optimum currency arrangements. J.
Afr. Econ. 6 (3), 412–444.
lachard, O.J., Quah, D., 1989. The dynamic effects of aggregate demand and
supply disturbances. Am. Econ. Rev. 79 (4), 655–673.
uigut, Valev, 2005. East and Southern African monetary integration: a structural
vector autoregression analysis. World Dev. 33 (12), 2119–2133.
larida, R., Gali, J., 1994. Sources of real exchange rate fluctuations: how impor-
tant are nominal shocks? Carnegie-Rochester Conf. Ser. Pub. Policy 41,
1–56.
ebrun, X., Masson, P., Pattilo, C., 2005. Monetary union in West Africa: who
might gain, who might lose, and why? Can. J. Econ. 38 (2), 454–481.
ebrun, X., Masson, P., Patillo, C., 2003. West African currency unions: ratio-
nale and sustainability. CESifo Econ. Stud. 49 (3), 381–413.
ichengreen, B., 1991. Is Europe a Common Currency Area? NBER, 3579.
rankel, J., Rose, A., 1996. A panel project on purchasing power parity: mean
reversion within and between countries. J. Int. Econ. 40, 209–224.
rankel, J., Rose, A., 1997. Is EMU more justifiable ex post than ex ante? Eur.
Econ. Rev. 41, 753–760.
eweke, J., 1982. Measurement of linear dependence and feedback between
multiple time series. J. Am. Stat. Assoc. 77 (378), 304–313.
yare, S.O., Lorde, T., Francis, B., 2005. West African States (Ecowas). West
Afr. J. Monet. Econ. Integr. 5 (2(a)).
empa, B., 2002. Is Europe converging to optimality? On dynamic aspects of
optimum currency areas. J. Econ. Stud. 29 (2), 109–120.
enen, P.B., 1969. The Theory of Optimum Currency: An Eclectic View.
ouparitas, 2001. Is the United States an Optimum Currency Area? An Empirical
Analysis of Regional Business Cycles. Federal Reserve Bank of Chicago,
WP-01-22.
asson, P., Patillo, C., 2003. Monetary union in West Africa: an agency of
restraint for fiscal policies? J. Afr. Econ. 11 (3), 387–412.
cKinnon, R.I., 1963. Optimum currency areas. Am. Econ. Rev. 53, 717–725.
undell, R.A., 1961. A theory of optimum currency areas. Am. Econ. Rev. 51,
657–665.
baseki, P.J., 2005. The future of the West African Monetary Zone (WAMZ)
Programme. West Afr. J. Monet. Econ. Integr. 5 (2(a)).
gunkola, O., 2005. An Evaluation of the Viability of a Single Monetary Zone
in ECOWAS. African Economic Research Consortium (AERC), Nairobi,
Research Paper 147.
jo, M.O., 2005. Towards a common currency in West Africa: progress, lessons
and prospects. West Afr. J. Monet. Econ. Integr. 5 (2(a)).
agbamah, 2005. Perspectives on the European monetary union: lessons for
the Economic Community of West African States (ECOWAS). West Afr. J.
Monet. Econ. Integr. 5 (2(a)).
oyibo, A., 1998. International payments system: West Africa ‘common cur-
rency’ is it possible or necessary? In: Presentation at a workshop onGhana, March 30–31, 1998 in CEPA.
AMI, 2002. Questions and Answers on the West African Monetary Zone.
Accra, Ghana.