A
fi
1
d
o
d
r
©
J
K
1
m
e
d
n
i
b
t
f
e
e
M
2
h
A
(
h
1
HOSTED BY Available  online  at  www.sciencedirect.com
ScienceDirect
Review of Development Finance 6 (2016) 91–104
Does development finance pose an additional risk to monetary policy?
Haruna Issahaku a,∗, Simon K. Harvey b, Joshua Y. Abor b
a Department of Economics and Entrepreneurship Development, Faculty of Integrated Development Studies, University for Development Studies, Ghana
b Department of Finance, University of Ghana Business School, Legon, Ghana
Available online 3 August 2016
bstract
This study investigates whether remittances entail extra risk for macroeconomic policy management and examines the role (if any) that the
nancial system can play in the interaction between remittances and monetary policy. Employing panel data for 106 developing countries from
970 to 2013, the results from our panel vector autoregressive (PVAR) model reveal that remittance volatility reduces macroeconomic risk in
eveloping countries while simultaneously stimulating a reduction in domestic interest rates. This finding remains robust to alternative specifications
f remittance volatility and monetary policy risk and to variations in the degree of financial development. The key lesson from this study is that
eveloping countries can leverage the positive impact of remittances in reducing macroeconomic instability by implementing policies that induce
emittances. 2016 Africagrowth Institute. Production and hosting by Elsevier B.V. All rights reserved.
EL classification: F33; F34; F35; O11
eywords: Remittances; Monetary policy; Developing countries; Financial development; Panel vector auto regression (PVAR)
m
r
m
2
a
2
d
m
p
p
s
w
b
H
t
A
R
.  Introduction
Remittances have become an important source of develop-
ent finance. Thus, it is not surprising that remittances have
ngaged the attention of researchers, policy makers, global
evelopment financial institutions and other development part-
ers. While policymakers continue to look to researchers for
deas to use remittances more effectively, research in this area has
een clustered around the microeconomic implications of remit-
ances (Ncube and Brixiova, 2013). These micro-level studies
ocus on the role of remittances in poverty reduction (Acosta
t al., 2008, 2007; Adams, 2004; Adams and Page, 2005; Gupta
t al., 2009), child growth (Antón, 2010; Carletto et al., 2011;
ansuri, 2006), employment (Amuedo-Dorantes and Pozo,
006; McCormick and Wahba, 2000; Taylor, 1999), and house-
old expenditures and investment (Adams and Cuecuecha, 2010;
dams, 2006; Yang, 2008), to name a few.∗ rCorresponding author.
E-mail addresses: iharuna@uds.edu.gh (H. Issahaku), sharvey@ug.edu.gh b
S.K. Harvey), joshabor@ug.edu.gh (J.Y. Abor). (
Peer review under responsibility of Africagrowth Institute. t
ttp://dx.doi.org/10.1016/j.rdf.2016.06.001
879-9337/© 2016 Africagrowth Institute. Production and hosting by Elsevier B.V. AThus, a gap remains in the empirical literature regarding the
acroeconomic implications of remittances. Even the limited
esearch on the macro-level impact of remittances has focused
ainly on remittances’ impact on growth (Barajas et al.,
009; Chami et al., 2012; Fayissa and Nsiah, 2010; Ncube
nd Brixiova, 2013; Nsiah and Fayissa, 2011; Pradhan et al.,
008; Waheed, 2004). Nonetheless, for policymakers in both
eveloping and emerging economies, gaining insight into the
acroeconomic influence of remittances is fundamental for
utting their countries on the path towards accelerated and pro-
oor growth (Ncube and Brixiova, 2013).
In particular, the impact of remittances on monetary policy
eems to have eluded the attention of empirical researchers,
hich has resulted in a limited understanding of the relationship
etween remittances and monetary policy (Vacaflores, 2012).
owever, economists have recently begun to test the existence of
he link between remittances and monetary policy (Adenutsi and
hortor, 2008; Chami et al., 2008; Mandelman and Zlate, 2012;
uiz and Vargas-Silva, 2010; Vacaflores, 2012). As limited as the
esearch in this field is, the evidence that has been uncovered has
een rather contradictory. For instance, Ruiz and Vargas-Silva
2010) examine the Mexican context and find no significant rela-
ionship between remittances and domestic monetary policy,
ll rights reserved.
9 velop
a v
s i
r c
w
R t
s f
i c
n w
i n
p w
i t
s p
0 t
M m
G s
d
m
c e
E t
V a
P o
t u
g v
a
c c
i c
f o
( c
w t
r m
o l
v
m t
d a
t e
w r
e
m m
s c
s R
s f
k
t a
m m
a
t w
o a
C l
r P
v o
p
o d
d d
h
a
d
r
2 H. Issahaku et al. / Review of De
lthough Adenutsi and Ahortor (2008) had earlier revealed a
ignificant relationship between monetary policy variables and
emittances in Ghana.
This confusion has been exacerbated by the proposition by
uiz and Vargas-Silva (2010, p. 174) that remittances that are
mall relative to the size of the economy will not have an
mpact on monetary policy. ‘If these flows are not large and/or
ot significant given the total size of the economy, then their
mpact on variables such as inflation, exchange rates and out-
ut will be minimal’. However, if the size of remittances is so
mportant, then why would they matter to monetary policy in a
mall economy, such as Ghana’s, in which they constitute only
.4% of GDP and why would they be rather insignificant in
exico where remittances add up to approximately 2.0% of
DP?
Furthermore, the previous literature on the interaction of
onetary policy and remittances consists mostly of single-
ountry studies: El-Sakka and McNabb (1999) focused on
gypt, Adenutsi and Ahortor (2008) on Ghana, Ruiz and
argas-Silva (2010) on Mexico, and Mandelman (2013) on the
hilippines. The problem with single-country studies is that
hey do not allow for wider applicability of the knowledge they
enerate. The previous literature on the subject on the whole
lso does not allow for the potential moderating effect of finan-
ial development in the remittance-monetary policy nexus. For
nstance, financial markets are known to play an intermediary
unction in the link between capital flows and economic growth
Agbloyor et al., 2014; Osabuohien and Efobi, 2013). However,
ill this moderating role hold in the case of the monetary policy-
emittance link? This question is one of the unresolved issues
n the topic.
Notwithstanding the perceived linkages among macroecono-
ic policy, remittances and the financial system, financial and
evelopment economists have been largely silent on this tripar-
ite nexus. In our literature search in connection with this study,
e have yet to encounter a study that examines the interactive
ffect of monetary policy and remittances on financial develop-
ent and the interactive effect of remittances and the financial
ystem on monetary policy efficiency. Thus, we have been pre-
ented with a fertile opportunity for research, and the present
tudy exploits this opportunity and fills this void.
In this paper, we employ panel vector autoregression (PVAR)
o overcome endogeneity problems; to establish causality among
onetary policy, remittances and other macroeconomic vari-
bles; and to generate orthogonalised impulse responses. We
hen use generalised impulse responses to identify the effects
f remittance shocks on monetary policy. Unlike the usual
holesky impulse responses, the use of generalised impulse
esponses helps us generate shocks that do not vary with the
ariable ordering.
We employ country-level panel data (annual) from 106 devel-
ping countries to analyse the dynamics of monetary policy
ecisions and remittance inflows. In the main, we investigate
ow remittance volatility affects monetary policy volatility. We a
rgue that if remittances flows are indeed countercyclical to the r
omestic economy, then remittance volatility must be negatively w
elated to the monetary policy rate and to monetary policy rate tment Finance 6 (2016) 91–104
olatility. In addition, a contractionary domestic monetary pol-
cy must trigger a remittance inflow that is consistent with the
ountercyclical view of remittances. To test the first hypothesis,
e compute the five-year rolling standard deviation of remit-
ances and the monetary policy rate and model them in a PVAR
ramework. To test the second hypothesis, we simulate monetary
ontraction following the Mundell–Fleming–Dornbusch model
ithin the framework of Cholesky innovations and orthogo-
alised generalised impulse response functions. In so doing,
e document a significant negative relationship between remit-
ances and remittance volatility, on one hand, and monetary
olicy rate and monetary policy volatility, on the other. In addi-
ion, controlling for the level of financial development and the
agnitude of remittances does not nullify this relationship, thus
upporting our claim that remittance volatility reduces both
omestic interest rates and monetary policy risk.
Our paper contributes in a number of ways to the financial
conomics discipline. First, the use of PVAR helps us to analyse
he dynamics of domestic monetary policy and remittances, in
ddition to country-specific fixed effects at the same time. Sec-
nd, the use of orthogonalised impulse responses enables us to
niquely isolate the impact of shocks from each of the system
ariables on the other variables, one at a time.
Our paper further extends the frontiers of knowledge in finan-
ial economics by presenting new evidence showing that a
ontractionary domestic monetary policy will activate the inflow
f remittances. We also add to those recent panel data studies that
onfirm a causal connection between monetary policy and remit-
ances (see, Termos et al., 2013; Vacaflores, 2012). Although
ost previous studies focus on remittances and monetary policy
evels, we take the step further to examine the dynamics in the
olatilities of the two variables. In particular, we find that remit-
ances and remittance volatility reduce the domestic interest rate
nd monetary volatility. Our results are in line with Craigwell
t al. (2010) and Bugamelli and PaternÒ (2011), who find that
emittances reduce receiving countries’ macroeconomic risks.
Our paper also contributes to the recent debate on the inter-
ediary function of financial development in the link between
apital flows and growth (see, Giuliano and Ruiz-Arranz, 2009;
amirez, 2013). This literature shows that remittances substitute
or financial markets in economic growth when capital mar-
ets are shallow. Our results are consistent with this literature
nd scales up the analysis to cover how finance enhances the
itigating impact of remittances on economic policy risk.
This paper is also related to Bugamelli and PaternÒ (2011),
ho analyse the impact of remittances on output volatility. These
uthors employ an instrumental variable approach to estab-
ish causality between the two variables. Unlike Bugamelli and
aternÒ (2011), however, we explore the effects of remittances
n interest rates and monetary policy risk. We argue that out-
ut is only an objective of monetary policy and that a more
irect assessment of the effect of remittances on monetary con-
itions is therefore required. In addition, whereas Bugamelli
nd PaternÒ (2011) focus on remittances, we examine both
emittances and remittance volatility. In terms of measurement,
hereas Bugamelli and PaternÒ (2011) measure volatility in
erms of deviations from the mean, we employ five-year rolling
velop
s m
l t
r b
a f
d e
a t
s t
e i
m t
i m
e e
m a
t f
b t
t
d u
m J
i a
d t
l g
fi s
s f
D r
i e
i
d a
p a
v
t P
a
r
m Y
S
2 w
i
2 a
p
2
e
D
s
a Y
l
c
s
k
a
t
o
t
q
b
(
H. Issahaku et al. / Review of De
tandard deviations to diminish the distortionary impact of out-
iers. Craigwell et al. (2010) also assess the association among
emittance, output, investment and consumption volatility using
 panel fixed effects methodology. However, their methodology
oes not allow them to generate impulse responses, which we see
s critical for separating the effects of remittance shocks from
hocks related to economic fundamentals. Unlike Craigwell
t al. (2010) we interact remittances with financial develop-
ent to assess remittances’ impact on macroeconomic policy
mpulses. Within this framework, we uncover a potential mod-
rating role of financial markets in reducing volatilities in both
onetary policy and remittances. We are further able to simulate
he influence of contractionary monetary policy on remittance
ehaviour.
Lastly, from a theoretical standpoint, this study lays the foun-
ation for the development of theory on the tripartite nexus of
onetary policy-remittances-financial development. Uncover-
ng the theoretical underpinnings of this tripartite nexus will help
eveloping countries’ policymakers to devise policies that will
et them get the most out of monetary policy, remittances, and
nancial development for socio-economic advancement. The
tudy seeks to answer the following three main questions. (1)
o remittances pose additional macroeconomic (monetary pol-
cy) risk in developing countries? (2) Do monetary conditions
n the recipient country affect remittance inflows? (3) What role
oes the financial system play in the link between monetary
olicy and remittances?
The remainder of this paper is structured as follows. In Sec-
ion 2, we specify our panel vector autoregressive (PVAR) model
nd describe the variables used. In Section 3, we present our
esults and a discussion on diagnostic exercises, PVAR esti-
ates, and the Cholesky and generalised impulse responses.
ection 4 concludes the paper.
.  Methodological  approach
.1.  The  model
Economists model economic issues in multilateral interde-
endency settings in two main ways (Canova and Ciccarelli,
013). The first option is to develop dynamic stochastic general
quilibrium (DSGE) models. However, although well-specified
SGE models provide precise solutions to policy questions and
implify the welfare implications of economic policy (Canova
nd Ciccarelli, 2013), their restrictive assumptions make them
argely unsuitable for analysing economic issues in a developing
ountry context. In particular, assumptions such as optimal risk
haring, consumption smoothening, homogenous labour mar-
ets, full employment, complete markets and rationality that
nchor a typical DSGE model are largely untenable in the con-
ext of developing countries (Senbeta, 2011). Moreover, certain
f the restrictions of the DSGE are often not consistent with
he distributional characteristics of the dataset, with the conse-
uence that policy recommendations from such models might m
e misleading (Canova and Ciccarelli, 2013). e
The second option is to develop panel vector autoregressive u
PVAR) models that avoid most of the restrictive assumptions [ment Finance 6 (2016) 91–104 93
ade in the DSGE models. The PVAR advantage derives from
he advantages of mother VAR models. First, all variables can
e treated as endogenous, but there is also the added flexibility
or including truly exogenous variables. Thus, PVARs resolve
ndogeneity, one of the most serious problems of econometric
ime series and panel data analysis. Second, PVARs facilitate
he analysis of the impact of innovations, making room for
nteractions among variables and thus producing dynamic solu-
ions that are not often attainable via OLS and other standard
odels (Li et al., 2012). The set of restrictions required in mod-
lling dynamic interdependencies using PVARs is not so limiting
s in DSGE models (Canova and Ciccarelli, 2013). Forecasts
rom VAR models are often more accurate than forecasts from
raditional structural models. PVARs can accommodate mul-
iple cointegration vectors, as opposed to Johansen (1988),
nlike the maximum likelihood cointegration procedure and the
ohansen and Juselius (1990) test for co-integration (Ericsson
nd Irandoust, 2004). PVARs permit the inclusion of fixed effects
hat capture country-specific time-invariant effects as well as
lobal time-invariant effects, and they can effectively handle
hort time dimensions due to extra degrees of freedom gained
rom the inclusion of cross-sections; moreover, by using impulse
esponse functions, PVARs can show delayed effects on (and of)
ach variable in the system (Grossmann et al., 2014).
The PVAR model is a mixture of the conventional VAR
pproach – in which all variables are considered endogenous
 priori – and the panel data approach in which unobserved indi-
idual heterogeneous effects are accommodated. The baseline
VAR model is represented below.
∑p
it =  B0i(t) + αitYit−k +  uit (1)
k=1
here Yit is a vector of K  endogenous variables for each country,
 = 1, .  . ., N  over t = 1, .  . ., T  time periods. In this study, Yit is given
s:
⎡ ⎤
σ
⎢ r
MPRit
⎢ ⎥⎢ σrREMIT ⎥⎢ it ⎥
⎢⎢
⎥
REER ⎥
⎢ it ⎥⎢ TRADE ⎥⎢ it ⎥
= ⎢⎢
⎥
FDI
it ⎢ it ⎥
⎥
⎢ ⎥LCPI ⎥
⎢⎢ it ⎥⎥
⎢⎢ DCPS⎢ it ⎥
⎥
⎥
⎣⎢ LGDP ⎥it ⎦
GDPgit
All variables are defined in Table 1. Boi(t) captures all deter-
inistic components (including constants, seasonal dummies,
tc.), Yit−k are lagged values of the endogenous variables, and
it is a K  × 1 vector of random disturbances given by uit =
u1t , u2t ,  .  . ., uNt]′∼iid(0,  Σ). αit and Boi(t) are allowed to be
94 H. Issahaku et al. / Review of Development Finance 6 (2016) 91–104
Table 1
Description of variables.
Variable Notation Description Data source
Economic openness TRADE Total trade as a ratio of GDP WDI
Financial development DCPS Domestic credit to the private sector as a ratio of GDP WDI
Remittances (2) LREMITT Logarithm of total remittance receipts WDI
Remittances (1) REMIT Personal remittances as a ratio of GDP WDI
Monetary policy rate MPR The central bank’s policy rate IFS
Lending interest rate LRATE Logarithm of the lending interest rate WDI
Inflation rate LCPI Logarithm of the consumer price index (CPI) WDI
Market size LGDP Logarithm of gross domestic product (GDP) WDI
Economic business cycles GDPg Growth rate of GDP WDI
Foreign direct investment FDI Foreign direct investment as a ratio of GDP WDI
Macroeconomic (in)stability (1) σrLCPI Five-year rolling standard deviation of the CPI WDI
Macroeconomic (in)stability (2) σrLCPI Standard deviation of the CPI calculated in the standard manner WDI
Macroeconomic (in)stability (3) σ2  r LCPI Five-year rolling variance of remittances as a ratio of GDP WDI
Macroeconomic (in)stability (4) σ2  s LCPI Variance of remittances as a ratio of GDP calculated in the standard manner WDI
Monetary policy risk (1) σrMPR Five-year rolling standard deviation of the monetary policy rate WDI
Monetary policy risk (2) σsMPR Standard deviation of the monetary policy rate calculated in the standard manner WDI
Monetary policy risk (3) σ2r MPR  Five-year rolling variance of the monetary policy rate WDI
Monetary policy 2 risk (4) σs MPR Variance of the monetary policy rate calculated in the standard manner WDI
Monetary policy risk (5) σ2r LRATE Five-year rolling variance of the lending interest rate WDI
Remittance risk (1) σrREMIT Five-year rolling standard deviation of remittances WDI
Remittance risk (2) σsREMIT Standard deviation of remittances calculated in the standard manner WDI
Remittance risk (3) σ2r REMIT Five-year rolling variance of remittances as a ratio of GDP WDI
Remittance risk (4) σ2s REMIT Variance of remittances calculated in the standard manner WDI
Monetary freedom MONEY FREEDOM Heritage Foundation’s (HF) measure of monetary freedom HF
M broad
N rs; H
c 2
a
i
Y a
e
M
w o
R fi
c
i σ
c
v
a
i
p
a
a
Y
(
W
v
d
oney supply LBMS Logarithm of 
ote: IFS, International Financial Statistics; WDI, World Development Indicato
ross-sectionally dependent. In the event that exogenous vari-
bles are present, Eq. (1) becomes:
∑p
it =  B0i(t) + αitYit−k +  Di(l)Rt +  uit (2)
k=1
here Di,j are K  ×  M  matrices for each lag j = 1, . . ., q, and
t is an M  ×  1 vector of exogenous variables common to all
ountries i.
Eqs. (1) and (2) have three main distinguishing character-
stics. First, they have Dynamic  Interdependencies, which are
aptured by incorporating the lagged values of the endogenous
ariables. Second, they have Static  Interdependencies, where uit
re allowed to be correlated with the cross-sectional dimension
. Cross-sectional  Heterogeneity, where the intercept and slope
arameters and the variances of the shocks are permitted to vary
cross units (countries).
Alternatively, based on Love and Zicchino (2006), we might
lso specify the PVAR in reduced form as follows:
∑p
it = αitYit−k +  τ2Ri +  fi +  dc,t +  eit (3)
k=1
The inclusion of exogenous variables (Ri) differentiates Eq.
3) above from the specification by Love and Zicchino (2006).hereas fi captures fixed effects – country-specific unobser-
able time-invariant effects, dc,t captures country-specific time i
ummies that represent macro shocks specific to each country. i money supply as a ratio of GDP WDI
F, Heritage Foundation.
.2.  Empirical  specification  of  the  model
Based on Eqs. (1) and (2), we specify the model equations
nvolving remittance and monetary policy in this section, as they
re the two most important variables in this study. The model
quations involving these two variables are specified below.
onetary policy risk can be specified as a function of the lags
f endogenous variables while controlling for country-specific
xed and time specific effects as follows:
∑p ∑p
rMPRit = ∅1jσrMPRit−j + ∅2jσrREMITit−j
j=1 j=1
∑p ∑p
+ ∅3jLGDPit−j + ∅4jTRADEit−j
j=1 j=1
∑p ∑p
+ ∅5jGDPgit−j + ∅6jLCPIit−j
j=1 j=1
∑p ∑p
+ ∅7jDCPSit−j + ∅8jREERit−j
j=1 j=1
∑p ∑p
+ ∅9jMPR.FDit−j + ∅10jREMIT.FDit−j
j=1 j=1
+  fi +  dt +  eit (4)
 is the country subscript while t is a time subscript; σrMPRit
s the monetary policy risk for country i at time t; σrREMITit−j
velop
i o
r i
o i
l i
i G
i
fi m
fi i
a a
R c
fi a
c m
d p
u
d t
c
a
σ
n
r
y
t
fi
t
a
C
r
(
t
f
v
T
3
w
m
m
2 s
c
i
o v
n R
( t
i o
t p
i b
p 4
W
T
s
‘
i
H. Issahaku et al. / Review of De
s the lag of remittance volatility; REERit−j is the lag of the
eal effective exchange rate; TRADEit−j is the lag of economic
penness, proxied by the share of trade in GDP; LCPIit−j is the
ag of inflation, proxied by the logarithm of the consumer price
ndex; LGDPit−j is the lag of the log of real GDP; LGDPgit−j
s the lag of the real GDP growth rate; DCPSit−j is the lag of
nancial development, proxied by total credit provided by the
nancial sector as a proportion of GDP; MPR.FDit−j is an inter-
ction term between monetary policy and financial development;
EMIT.FDit−j is an interaction term between remittances and
nancial development; fi captures the country i-specific inter-
ept representing country-specific fixed effects; dt captures time
ummies; and eit is the noise error term.
Similarly, remittance volatility can be specified as the main
ependent variable as follows.
∑p ∑p
rREMITit = θ1jσrMPRit−j + θ2jσrREMITit−j
j=1 j=1
∑p ∑p
+ θ3jREERit−j + θ4jTRADEit−j
j=1 j=1
∑p ∑p
+ θ5jLGDPit−j + θ6jGDPgit−j
j=1 j=1
∑p ∑p
+ θ7jLCPIit−j + θ8jDCPSit−j
j=1 j=1
∑p
+ θ9jMPR.FDit−j
j=1
∑p
+ θ10jREMIT.FDit−j
j=1
+ o fi + do o t +  eit (5)
here all variables are as defined under Eq. (4) above.
.3.  Data  and  variable  selection
Apart from the monetary policy rate (MPR), which was
btained from the International Monetary Fund’s (IMF) Inter-
ational Financial Statistics (IFS), and Monetary Freedom
MONEY FREEDOM), which was obtained from the Her-
tage Foundation (HF), all other variables were sourced from
he World Bank’s World Development Indicators (WDI). We
nclude 106 developing countries around the world in our sam-
le, and these countries are listed in Table A1 in the appendix.
e use an unbalanced panel (annual data) from 1970 to 2013.
wo main factors informed our selection of countries for the
tudy. First and foremost, in deciding which countries are in the 3
Developing Country’ category we used the IMF list of develop-
ng countries,1 which is the most widely accepted classification
U
(
1 We used the list captured in the 2013 World Economic Outlook reports. iment Finance 6 (2016) 91–104 95
f countries. Secondly, for a country to be selected for the study,
t must have sufficient data for the main variables for the study,
ncluding remittances, monetary policy rate and/or the lending
nterest, and financial development (private credit as a ratio of
DP).
The central bank’s monetary policy rate (MPR) is used as the
ain measure of monetary policy. We use this variable because
t reflects the reactions of the monetary authorities to domestic
nd international economic conditions. The policy rate is also
onsidered the indicative interest rate in the domestic economy,
nd all other interest rates are fixed with respect to it. To capture
onetary policy risk we compute the standard deviations of the
olicy rate (σrMPR) with a five-year rolling window and also
se the normal standard deviation (σsMPR) (deviations from
he mean) for robustness checks. Further robustness checks are
onducted later using the five-year moving variance (σ 2 r MPR)
nd normal variance (σ 2s MPR) in the MPR.
We measure remittances (REMIT) as the share of total inter-
ational remittance inflows in GDP. Analogously, we measure
emittance risk (volatility) in four similar ways – as the five-
ear moving standard deviation of remittances (σrREMIT), as
he normal standard deviation of remittances (σsREMIT), as the
ve-year rolling variance 2 of remittances (σr REMIT), and as
he normal variance of remittances (σ 2s REMIT). Standard devi-
tions of remittances have been employed in previous studies by
raigwell et al. (2010) and Bugamelli and PaternÒ (2011).
Inflation is proxied by the log of the CPI, and the five-year
olling standard deviation of CPI is used to proxy for economic
in)stability. We use the log of GDP to measure market size and
he growth rate of GDP as a measure of changes in economic
ortunes (business cycle effects). The description of all of the
ariables, data sources and associated notations are reported in
able 1.
.  Results  and  discussion
Descriptive statistics are reported in Table 2. Because the
ean is susceptible to distortions from outliers, we use the
edian of the distribution for our discussion. Median con-
umer inflation (CPI) is quite high (46.44%), which signals high
ommodity prices in developing countries. The measure of the
nterest rate, the monetary policy rate (MPR), has a high median
alue, indicating the high cost of funds in the developing world.
emittances as a percentage of GDP is 1.86, which signals
he increasing significance of remittances as a source of devel-
pment finance in developing economies. When channelled
roperly, these receipts could facilitate economic development
y increasing GDP growth (GDPg) above the median value of
.28%.
.1.  Model  selection  and  estimationThe criteria for model selection is presented in Table 3.
sing the model selection criteria suggested by Andrews and Lu
2001), the preferred model is a first-order panel VAR because
t yields the minimum values for MBIC, MAIC and MHQ. On
96 H. Issahaku et al. / Review of Development Finance 6 (2016) 91–104
Table 2
Descriptive statistics.
Mean Median Maximum Minimum Std. dev. Jarque–Bera Prob. Obs.
CPI 48.64 46.44 288.65 0.00 37.59 119.66 0.00 3471
DCPS 30.06 23.72 165.72 0.80 23.89 6464.97 0.00 3727
FDI 3.10 1.62 53.81 0.06 5.05 90,601.72 0.00 3608
MPR 12.62 8.99 200.00 0.020 16.33 111,015.40 0.00 654
REMIT 4.69 1.86 106.48 0.00 9.02 247,713.10 0.00 3197
GDPg 3.85 4.28 88.96 -5.0E + 01 5.84 69,195.00 0.00 3902
TRADE 75.26 68.59 375.38 6.32 40.12 2080.32 0.00 3715
Note: MPR, monetary policy rate; REMIT, remittances; GDPg, gross domestic produ
total trade; DCPS, domestic credit provided to the private sector.
Table 3 f
Selection order criteria. o
Lag MBIC MAIC MHQ
1 −273.4472 −85.9615 −161.5667 3
2 −184.3375 −59.3469 −109.7505
3 −94.8601 −32.3649 −57.5666
N v
c e
r
p
t d
w i
p
s
c
3
t
(
3 d
r
p
e o
t a
t T
a m
t r
t l
p B
(
s s
a i
v m
t f
e h
d a
e p
e s
r
s t
I s
r
e tively related to remittances, which suggests that FDI acts as an
a
p
l
ote: MBIC, modified Bayesian criteria, MAIC, modified Akaike information
riteria; MHQ, modified Hannan–Quinn information criteria.
he basis of the results of the model selection criteria (Table 3),
e fit a first-order panel VAR.
Our PVAR models are all exactly identified, and for that rea-
on, Hansen’s J statistic of over-identifying restrictions is not
omputed. Monte Carlo simulation with 1000 repetitions is used
o produce 5% error bands for impulse response functions.
.2.  Results  of  panel  unit  root  test
In time series and panel data analyses, it is important to
xplore the order of variable integration. The stationarity sta-
us (the order of integration) of the variables helps to choose
he appropriate model for estimating the coefficients. There are
dvantages to deploying panel unit root tests over individual
ime series-based unit root tests. First, panel data-based unit root
ests have more statistical power than their univariate counter-
arts. In a panel setting, the traditional Augmented Dicky–Fuller
ADF) has low power identifying stationarity, particularly in
hort panels. Second, panel unit root tests are less restrictive
nd allow for fixed effects at the country level as well as time
ariations in the parameters across panels. Moreover, panel data
echniques provide a suite of estimation options ranging from
stimation with no trend and no constant, to estimations with a
eterministic trend and a constant, and testing for common time
ffects. These techniques provide a high degree of flexibility in
stimating parameters.
The results from Table 4 show that, apart from the Con-
umer Price Index (CPI), all variables are integrated of order
(0). The CPI is integrated of order I(1). In addition, the loga-
ithmic (logs) transformation of CPI is stationary at level. We
mploy the logs of CPI in our estimation, which implies that
ll variables used for our estimations do not follow a unit root a
rocess and suggests that it is unlikely that a unique state of fl
ong-run equilibrium for the system variables exists. The results bct growth; CPI, consumer price index; FDI, foreign direct investment; TRADE,
rom unreported cointegration tests confirm the non-existence
f a unique long-run relationship.
.3.  Monetary  policy  and  remittances
We present the results of the PVAR in Table 5. The dependent
ariable for Model 1 is remittance as a ratio of GDP; the depend-
nt variable for Models 2–8 is the central bank’s monetary policy
ate (MPR) used to capture the monetary policy stance and the
revailing interest rate. We include the five-year rolling standard
eviation of consumer inflation instead of the CPI, as we view
t as a better measure of macroeconomic (in)stability. Table 1
rovides the description of variables.
.3.1. Macroeconomic  determinants  of  remittances
Model (1) in Table 5 reveals that financial development
DCPS) is negatively related to remittances. This finding
oes not necessarily imply that financial development reduces
emittance inflows. We offer two interpretations. The first inter-
retation is that a financial sector that is not well developed
bstructs the flow of remittances by increasing both the monetary
nd non-monetary costs of sending and receiving remittances.
he second interpretation is that remittances and financial
arkets play substitute roles in growth, which occurs when
emittance recipients rely on migrants for ‘credit’ instead of the
ocal financial system. This latter interpretation concurs with
rown et al. (2013).
Remittances are largely self-driven, which is shown by the
ignificance of the lag of remittances. Once migrants start send-
ng money home, they have the propensity to continue sending
oney because they feel obliged to promote the welfare of the
amily and friends they left behind. In addition, monies sent back
ome to undertake projects are usually delivered incrementally
nd not in bulk. We further find that the size of the economy
ositively impacts the flow of remittances. In addition, our mea-
ure of economic business cycles, growth in GDP, has an inverse
elationship with the inflow of remittances supporting the coun-
ercyclical view of remittances. However, this coefficient is not
ignificant.
Our alternative measure of foreign inflows, FDI, is nega-lternative source of international finance in reality. These two
ows are underpinned by different characteristics, as described
y Chami et al. (2008). Unlike other capital flows, remittances
H. Issahaku et al. / Review of Development Finance 6 (2016) 91–104 97
Table 4
Panel unit root test.
MPR REMIT LREMITT σ 2s REMIT LGDP CPI FDI
Level
LLC −11.00*** 7.97*** −10.65*** −46.66*** −0.49828 47.57*** −11.18***
IPS −6.80*** −4.8*** −2.53*** −15.64*** 12.2972 54.97*** −13.35
ADF 185.24*** 330.44*** 349.76*** 476.36*** 158.073 39.76*** 619.74***
PP 204.75*** 372.82*** 342.32*** 743.92*** 194.544 37.34*** 609.03***
First difference
LLC −33.94*** −12.32***
IPS −34.29*** −14.65***
ADF 1525.28*** 770.79***
PP 1529.36*** 971.55***
LRATE REER RIR LCPI TRADE
Level
LLC −17,646.2*** −6.51*** −32.34*** −20.15*** −3.74***
IPS −3707.32*** −5.17*** −28.03*** −19.35*** −4.87***
ADF 959.38*** 219.44*** 1048.70*** 1445.44*** 330.57***
PP 1115.92*** 225.12*** 1132.33*** 1143.66*** 337.35***
Note 1: LLC, Levine–Lin–Chu statistics; IPS, Im, Pesaran and Shin statistics; ADF, Augmented Dickey Fuller Fisher Chi-square statistics; PP, Phillips Perron
statistics.
N
N e 5% 
i 2
m o
t
i i
r
i
3
m
e
p
a
t
r
s
fi
l
i
T
i
r
l
t 3
t v
r
n a
t w
a m
M o
r A
a r
1 t
a
r i
y i
n n
t t
t
m
v
ote 2: All variables are described in Table 1.
ote 3: *** shows significance at the 1% level, and ** shows significance at th
gnite family bonds. Second, these ignited familial relationships
ake remittances respond more to the needs of family members
han standard private capital flows, which are largely driven by
nvestment motives.
.3.2.  Remittances  and  monetary  policy  –  dissecting  the
vidence
There is strong confirmation of a negative impact of remit-
ances on the monetary policy rate that is evidenced by the
tatistical significance as well as the negative coefficient of the
ag of remittances in models 2, 3 and 5, as shown in Table 5.
here are two explanations for this finding. First, an increase in
emittances boosts the quantity of loanable funds available for
ending in the economy, which may then lead to a decline in
he interest rate. Second, when households receive remittances,
heir demand for formal credit will decline if the remittance
eceived is large enough to meet their welfare and investment
eeds, which will cause interest rates to decline. This revela-
ion is consistent with the prevailing wisdom based on single-
nd cross-country studies. For instance, using a DSGE model,
andelman (2013) finds that remittance inflows reduced interest
ate in the Philippines. In addition, Vacaflores (2012) employs
 DSGE model and comes to the same conclusion in a panel of
1 Latin American countries.
Our finding that remittances reduce domestic interest rate
emains robust when remittances are measured in terms of five-
ear rolling standard deviation, normal standard deviation, and
ormal variance. This finding implies that the volatility of remit-
ances helps to ease domestic interest conditions thereby helping
o stabilise the macroeconomy. h
The ability of remittances to ensure output and macroecono- b
ic stability stems from the capability of remittances to reduce t
olatilities in consumption and investment (Craigwell et al., mlevel.
010). We will further discuss the macroeconomic implications
f remittances in the next section.
As expected, a rise in each of the following causes the pol-
cy rate to fall: financial development, real effective exchange
ate, economic openness and size. This finding suggests that
f developing countries can improve and sustain macroecono-
ic gains, they can improve the effectiveness of their monetary
olicies. An effective monetary policy will promote the growth
nd income of the populace. As the income of the citizenry
ises, their demand increases for goods and services, including
nancial assets, which opens up more space for monetary pol-
cy management. In addition, a more favourable exchange rate
s conducive for monetary policy management.
.3.3.  Does  monetary  policy  volatility  affect  remittance
olatility?
The volatility of monetary policy or interest rates has an
dverse impact on economic growth. Therefore, central banks
orldwide seek to stabilise monetary conditions to ensure
acroeconomic stability. Model 1 under Table 6 shows the effect
f policy rate volatility on the variation in remittance inflows.
n increase in monetary policy volatility tends to decrease
emittance volatility. This finding is consistent with the coun-
ercyclical properties of remittances, which are derived from the
ltruism theory of remittance. When macroeconomic conditions
n the receiving country are unfavourable, we expect an increase
n remittance inflows, and we expect the reverse when macroeco-
omic conditions improve. Migrants are considered sensitive to
he plight of their families back home and often offer a helping
and when conditions in the home country hit their family mem-
ers hard. This finding also confirms the widely held view that
he macroeconomic environment in the receiving country affects
igrants’ remitting behaviour. The countercyclical properties
98 H. Issahaku et al. / Review of Development Finance 6 (2016) 91–104
Table 5
Monetary policy and remittances.
REMIT (1) MPR (2) MPR (3) MPR (4) MPR (5) MPR (6) MPR (7) MPR (8)
REMIT(−1) 1.0208*** −0.1671*** −0.1075* −0.2456 −0.1783***
(0.0148) (0.0572) (0.0601) (0.0484) (0.0498)
MPR(−1) −0.0093 0.2364*** 0.2235*** 0.1906*** 0.1748*** 0.6806*** 0.1330*** 0.1872***
(0.0078) (0.0301) (0.0300) (0.0256) (0.0251) (0.0363) (0.0258) (0.0253)
σrLCPI(−1) 0.3623 −4.4956** −5.5559*** −4.1206*** −5.3386*** 1.7243 −0.0510*** 0.0005**
(0.5232) (2.0267) (2.0303) (1.6930) (1.6674) (2.0462) (0.0158) (0.0003)
DCPS(−1) −0.3778** −3.0653*** −2.9369*** −5.4937*** −5.3942*** −1.5309*** −4.0263*** −5.1504***
(0.1433) (0.5551) (0.5484) (0.5244) (0.5083) (0.3973) (0.5339) (0.5107)
TRADE(−1) 0.0093*** −0.0304*** −0.0242** −0.0140 −0.0065 −0.0061 −0.0073 −0.0112
(0.0028) (0.0108) (0.0109) (0.0092) (0.0091) (0.0059) (0.0083) (0.0087)
REER(−1) −0.0005 −0.0023 −0.0022 −0.0039*** −0.0039*** −0.0005 −0.0042*** −0.0042***
(0.0004) (0.0015) (0.0015) (0.0013) (0.0013) (0.0009) (0.0013) (0.0014)
GDPg(−1) −0.0264 −0.0007 −0.0519 −0.0954 −0.1565* −0.0366 −0.0337 −0.0531
(0.0292) (0.1129) (0.1127) (0.0948) (0.0931) (0.0605) (0.0897) (0.0940)
LGDP(−1) 0.1109** −0.6886*** −0.7199*** −0.9090*** −0.9497*** −0.5200*** −0.7170*** −0.9086***
(0.0450) (0.1744) (0.1721) (0.1473) (0.1430) (0.0934) (0.1465) (0.1482)
FDI(−1) −0.0716*** −0.0767 −0.0013 0.0279 0.1172 0.0080 0.1376* −0.0075
(0.0243) (0.0943) (0.0966) (0.0794) (0.0801) (0.0504) (0.0799) (0.0807)
REMIT.FD −0.0597*** −0.0689*** −0.01625 −0.0477*** 0.7545***
(0.0210) (0.0173) (0.0109) (0.0168) (0.0771)
MPR.FD 0.7021*** 0.7163*** 0.2858*** 0.5865***
(0.0718) (0.0687) (0.0508) (0.0759)
σrREMIT −0.2685*
(0.1454)
σsREMIT −0.1212***
(0.0508)
σ 2s REMIT −0.0064***
(0.0019)
R-squared 0.9609 0.4939 0.5115 0.6486 0.6720 0.8788 0.6705 0.6252
Adj. R-squared 0.9594 0.4735 0.4896 0.6328 0.6558 0.8725 0.6545 0.6087
F-statistic 612.1135*** 24.2859*** 23.3487*** 41.1511*** 41.3551*** 138.4670*** 41.8167*** 37.8723***
Note 1: Remittances is the dependent variable for model (1), while the central bank’s monetary policy rate is the dependent variable for models (2) to (8).
Note 2: MPR, monetary policy rate; REMIT, remittances as a ratio of GDP; GDPg, growth rate of GDP; σrREMIT, five-year rolling standard deviation of remittance
inflows; σsREMIT, the (normal) standard deviation of remittances; 2 σs REMIT, the (normal) variance of remittances; LGDP, the log of gross domestic product;
σrLCPI, the five-year rolling standard deviation of the consumer price index; FDI, foreign direct investment; REER, the real effective exchange rate; TRADE,
total trade as a ratio of GDP; DCPS, domestic credit to private sector; REMIT.FD in an interactive term between remittances and financial development; MPR.FD,
an interaction term between monetary policy (MPR) and financial development; σrMPR, five-year rolling standard deviation of MPR; σsMPR, (normal) standard
deviation of MPR; and (−1) placed after a variable indicates the lag of the variable.
N ures i
o fl
B p
b
r i
n a
H (
i r
v
3 s
m i
i
p (
a v
2 o
e
p
i
r
ote 3: ***, **, * represents significance at 1%, 5% and 10%, respectively. Fig
f remittances have been confirmed by Craigwell et al. (2010),
ugamelli and PaternÒ (2011), and Adenutsi (2014).
We further find that an advanced financial system reduces
emittance volatility. In addition, an increase in economic open-
ess tends to decrease the variability in remittance flows.
owever, as the domestic economy expands, remittance volatil-
ty also increases.
.3.4. Do  remittances  constitute  an  additional
acroeconomic  risk?
We report the impact of remittance uncertainty on monetary
olicy risk (measured as the rolling and normal standard devi-
tion of the policy rate) in Table 6. The results from models
 to 6 are provided with monetary policy risk as the depend-
nt variable. Remittance volatility tends to reduce 9 monetary
olicy e riskiness. The finding is fairly consistent in the major-
ty s of our models and is consistent with one of the established
egularities in the empirical literature, that unlike other capital fln parentheses are standard errors.
ows such as official development assistance, FDI and private
ortfolio flows, remittances are countercyclical and can act as a
uffer for macroeconomic stability for that matter. By smoothen-
ng consumption, for instance, remittances help raise economic
ctivity during hard times and reduce business cycle effects
Singer, 2010). The macroeconomic risk-mitigating impact of
emittances remains robust, whether we measure remittance
olatility as a five-year moving standard deviation or as normal
tandard deviation. Previous research on the macroeconomic
mplications of remittances reached similar conclusions. For
nstance, in a study of 69 economies, Bugamelli and PaternÒ
2011) confirm a negative link between remittances and output
olatility. In addition, Craigwell et al. (2010) support the role
f remittances in taming macroeconomic shocks in a panel of
5 countries. The ability of remittances to ameliorate macro-
conomic risk arises from the low procyclical nature, increasing
ize and stability of remittances relative to other types of capital
ows.
H. Issahaku et al. / Review of Development Finance 6 (2016) 91–104 99
Table 6
Remittance risk and monetary policy risk.
σrREMIT (1) σrMPR (2) σrMPR (3) σrMPR (4) σrMPR (5) σsMPR (6)
σrREMIT(−1) 0.8961*** −0.1939** −0.1002 −0.1923** −0.1016
(0.0374) (0.0929) (0.0921) (0.0937) (0.0927)
σrMPR(−1) 0.0417*** 0.0022 0.0106 0.0019 0.0111
(0.0133) (0.0331) (0.0318) (0.0333) (0.0319)
σrLCPI(−1) −0.7117 4.1068** 2.4999 4.1168** 2.4803
(0.7613) (1.8898) (1.8558) (1.8965) (1.8639)
DCPS(−1) −0.1798** −0.6976*** −0.7489*** −0.6779*** −0.7719*** −4.0262***
(0.0858) (0.2131) (0.2045) (0.2436) (0.2346) (0.5339)
TRADE(−1) 0.0049*** −0.0008 0.0035 −0.0008 0.0036 −0.0074
(0.0016) (0.0041) (0.0040) (0.0041) (0.0041) (0.0083)
REER(−1) 2.70E−05 −0.0012** −0.0012* −0.0012** −0.0011* −0.0042***
(0.0002) (0.0006) (0.0006) (0.0006) (0.0006) (0.0013)
GDPg(−1) −0.0077 −0.0767* −0.1155** −0.0758** −0.1167*** −0.0337
(0.0163) (0.0405) (0.0399) (0.0409) (0.0405) (0.0897)
LGDP(−1) 0.0536** 0.1078 0.0942 0.1079 0.0939 −0.7171***
(0.0276) (0.0685) (0.0657) (0.0687) (0.0659) (0.1465)
FDI(−1) −0.0194 −0.0187 0.0242 −0.0197 0.0256 0.1376*
(0.0144) (0.0356) (0.0358) (0.0363) (0.0366) (0.0799)
σsREMIT(−1) −0.1212***
(0.051)
σsMPR(−1) 0.1330***
(0.0258)
σsLCPI(−1) −0.0510***
(0.0158)
REMIT.FD −0.0293*** −0.0295*** −0.04767***
(0.0075) (0.0075) (0.0168)
MPR.FD −0.0055 0.0064 0.5865***
(0.0328) (0.0316) (0.0759)
R-squared 0.8327 0.2922 0.3545 0.2922 0.3547 0.6705
Adj. R-squared 0.8234 0.2526 0.3142 0.2481 0.3100 0.6545
F-statistic 89.0788*** 7.3838*** 8.7877*** 6.6082*** 7.9447*** 41.8167***
N ors ar
3 t
r fi
a
e e
a a
v i
I T
d n
t s
t t
n
m 3
m r
w
t
t s
d i
t i
r e
r u
n
T
e
ote: Refer to notes under Table 5 for the description of variables. Standard err
.3.5.  Dissecting  the  role  of  financial  development  in  the
emittance-monetary  policy  nexus
Financial markets contribute to economic progress by
nhancing efficiency and risk sharing, monitoring managerial
ctions to prevent fraud, harnessing and channelling savings to
iable projects, and by reducing the cost of access to financing.
f these properties of financial markets hold, then our financial
evelopment variable must be negatively related to the mone-
ary policy rate or to the domestic interest rate. Table 5 shows
hat the financial development variable (DCPS) is consistently
egative and significant for models 2, 3, 4, 5, 6 and 8, which
eans that a well-developed financial sector will lead to a lower
onetary policy rate and hence a lower domestic interest rate. A
ell-developed financial system offers a wider scope for mone-
ary policy than an immature system. This finding dovetails with
he findings by Krause and Rioja (2006) that financial market
evelopment promotes monetary policy efficiency. Table 6 fur-
her shows that financial development lessens macroeconomic
isk by reducing volatility in the policy rate. This result is quite
obust, as it is consistent in all of the model specifications.Additionally, the remittance-finance interactive term is sig- a
ificant and has a negative sign in models 3, 5, 7 and 8 in t
able 5, which means that finance complements the stabilising i
ffect of remittances on macroeconomic variables. According ae in parentheses.
o Agbloyor et al. (2014) and Osabuohien and Efobi (2013),
nancial markets play a moderating role between capital flows
nd growth. In doing so, financial markets augment the positive
ffects of capital flows on the economy while hindering any neg-
tive impact. This finding highlights the need for policy reform
n developing countries to make financial markets more efficient.
he interactive term between remittances and finance is also sig-
ificant in minimising macroeconomic risk (policy volatility) as
hown in the results presented in Table 6. The robust nature of
his finding should be of consequence to macroeconomic policy.
.4.  The  effects  of  a  contractionary  monetary  policy  on
emittance inflows
A key unresolved issue in measuring monetary shocks is the
pecification of a contractionary or expansionary monetary pol-
cy. Conventionally, a rise in the short-term interest rate or a fall
n monetary aggregates is interpreted as a contractionary mon-
tary policy. In this regard, the recursive Cholesky approach is
sed to identify monetary shocks. However, Ho and Yeh (2010)
rgue that this identification may be suitable only with respect
o a closed economy. They argue that in a closed economy, the
nterest rate is the main instrument of monetary policy, such that
 policy tightening may cause the short-run interest rate to fall.
1 velop
H a
v s
a c
Y
c t
t c
s s
c t
t r
t s
L
t s
a p
T c
i
t
i
00 H. Issahaku et al. / Review of De
owever, for an open economy in which there are large inter-
entions in the forex market, a tight policy may be captured by
 rise in interest rates or a reduction in foreign reserves.
A sign restriction methodology, as proposed by Uhlig (2005),
an be employed to identify different contractionary mone-
ary policy identification schemes. Alternative sign restrictions
chemes have been implemented with varying degrees of suc-
ess. First, Bernanke and Blinder (1992) implement a scheme
hat assumes that when there is a contractionary monetary shock,
he short-term interest rate will not fall. Second, Gordon and
eeper (1994) use a scheme based on the assumption that a con-
ractionary monetary policy will not lead to a rise in monetary
ggregates. A third identification scheme combines the first two.
he fourth scheme views monetary contraction as innovationsn both the interest rate and the exchange rate. The fifth alterna- f
ive scheme captures monetary policy innovations as a decrease m
n money supply, an appreciation of the domestic currency, and t
Respo nse of  REMIT_ to Cholesky
One S.D. In novatio ns
6 6
5
4
4
2
3
2 0
1
-2
0
-4
-1
-2 -6
1 2 3 4 5 6 7 8 9 10
REMIT_ MPR REER
TRADE LCPI DCPS
GDP_GROWTH LBMS LGDP
Fig. 1. Response of remittances to co
Ac cumulated Response of REMIT_ t o C holesky
One S.D. In nova tions
40
40
30 30
20
20
10
10 0
-10
0
-20
-10 -30
1 2 3 4 5 6 7 8 9 10
REMIT_ MPR REER
TRADE LCPI DCPS
GDP_GROWTH LBMS LGDP
Fig. 2. Response of remittances to monetaryment Finance 6 (2016) 91–104
n increase in the interest rate (Mountford, 2005). The sixth
cheme posits that a tightening of monetary policy will not
ause interest rates to fall or foreign reserves to rise (Ho and
eh, 2010). Rafiq and Mallick (2008) use the seventh alterna-
ive identification scheme by employing data for three European
ountries, and the restrictions in this scheme are based on the
tandard Mundell–Fleming–Dornbusch model, which stipulates
hat tight monetary policy will cause interest rates to rise and the
eal exchange rate to appreciate, while causing prices, money
upply, and real output to fall.
Ho and Yeh (2010) find that identification schemes one to five
uffer from one or more of price, liquidity and/or exchange rate
uzzles. The price puzzle arises when a tight monetary policy
auses the price level to rise instead of causing the price level to
all. In the case of the liquidity puzzle, positive innovations in
onetary policy cause interest rates to rise instead of depressing
hem. With the exchange rate puzzle, a tight monetary policy
Res ponse of REMIT_ to Ge neralized One
S.D. Innovations
1 2 3 4 5 6 7 8 9 10
REMIT_ MPR R EER
TRADE LCPI DCPS
GDP_GROWTH LBMS LGDP
ntractionary monetary policy.
Accumulate d Res ponse of REMIT_ to Ge neraliz ed One
S.D. In nova tions
1 2 3 4 5 6 7 8 9 10
REMIT_ MPR RE ER
TRADE LCPI LDCPS
GDP_GROWTH LBMS LGDP
 contraction – accumulated response.
velop
s l
t c
( fi
r
( s
t f
e r
o
G 3
p
g
c m
I p
m m
t i
t a
i t
t
m
r
a
T
R
σ
σ
σ
T
F
D
R
M
D
F
C
R
A
S
S
F
L
A
S
N
D
e
r
m
H. Issahaku et al. / Review of De
hock leads to a depreciation – instead of an appreciation – of
he currency. Only schemes six (Ho and Yeh, 2010) and seven
Rafiq and Mallick, 2008) avoid all of the puzzles.
Based on the foregoing discussion, we follow scheme seven
the Mundell–Flemin–Dorbusch model) and specify a contrac-
ionary monetary policy as a one-unit positive shock to the inter-
st rate (MPR), a one-unit positive shock to the exchange rate, a
ne-unit negative shock to inflation, a one-unit negative shock to
DP, a one-unit negative shock to money supply, and a one-unit
ositive shock to GDP growth. The inclusion of a shock to GDP
rowth is to control for supply shocks to prevent misidentifi-
ation. The impulse responses from Cholesky and Generalised
mpulse Responses are shown in Fig. 1. The associated accu-
ulated responses are shown in Fig. 2. Both the Cholesky and
he Generalised Impulse Responses in Fig. 1 show that a con-
ractionary monetary shock leads to a steady rise in remittance
nflows. This finding implies that remittances can frustrate con-
ractionary monetary policies if not properly anticipated.If properly anticipated, remittances can serve as pseudo auto- b
atic stabilisers and can substitute for monetary policy. This d
esult is consistent with Singer (2010), who argues that in o
 trilemma policy framework, remittances can substitute for m
able 7
emittances and monetary policy (lending interest rate).
σ 2r REMITT σ 2 2 r LRATE σr LCPI TRADE
2
r REMITT(−1) 0.6066*** −0.2957* 4.93E−06 −0.030
(0.0294) (0.1559) (9.6E−05) (0.0510
2
r LRATE(−1) −0.0003 0.1620*** 4.45E−06*** −0.000
(0.0004) (0.0019) (1.2E−06) (0.0006
2
r LCPI(−1) −3.5819 17.9678 0.8603*** 7.4905
(4.6751) (24.7993) (0.0153) (8.1154
RADE(−1) 0.0183*** 0.0139 −8.46E−06 0.9749
(0.0058) (0.0307) (1.9E−05) (0.0100
DI(−1) −0.0159 −0.0199 9.26E−05 −0.000
(0.0429) (0.2275) (0.0001) (0.0744
CPS(−1) −0.0119 −0.5817 0.0005 −0.472
(0.4401) (2.3346) (0.0014) (0.7639
EER(−1) −0.0347*** −0.0798 −0.0001*** 0.0098
(0.0097) (0.0517) (3.2E−05) (0.0169
ONEY FREEDOM(−1) −0.0554 −0.8664*** 0.0006*** 0.0463
(0.0339) (0.1801) (0.0001) (0.0589
REMITM 1.5210*** −2.9587 −3.07E−05 −0.156
(0.4547) (2.4123) (0.0015) (0.7894
INCDEVM −0.9702 −7.3643* −0.0016 0.9350
(0.7151) (3.7933) (0.0023) (1.2413
 6.6956** 88.5848*** −0.0349*** −0.590
(2.7809) (14.7516) (0.0091) (4.8274
-squared 0.6033 0.9411 0.9218 0.9628
dj. R-squared 0.5950 0.9399 0.9201 0.9620
um sq. resides 9501.0430 267,341.6 0.1018 28,629
.E. equation 4.4629 23.6741 0.0146 7.7472
-statistic 72.5522*** 762.1170*** 562.0185*** 1235.0
og likelihood −1416.839 −2231.098 1375.394 −1685
kaike AIC 5.8518 9.18893 −5.5918 6.9548
chwarz SC 5.9463 9.2834 −5.4973 7.0493
ote: σ 2s LRATE, the five-year rolling variance of the lending interest rate; REER, 
CPS, domestic credit to private sector; FINCDEVM, a financial development dumm
xceeds the median level of financial development and zero otherwise (low financia
eceiving country) when a country’s remittance receipts exceed the median level an
onetary freedom; and (−1) placed after a variable indicates the lag of the variable. ment Finance 6 (2016) 91–104 101
oss of monetary independence based on their stabilising and
ountercyclical properties and allow economies to implement
xed exchange rate regimes. The results from the accumulated
esponses in Fig. 2 are more definite. A contractionary monetary
hock causes a persistent rise in remittance inflows. It is there-
ore safe to conclude that monetary tightening causes a rise in
emittance inflows.
.5.  Further  robustness  checks
We performed further robustness checks against measure-
ent error and misspecification. First, instead of the monetary
olicy rate, we used the lending rate as an alternative proxy for
onetary policy because the lending rate responds to changes
n the policy rate. The correlation between the two variables is
pproximately 73.54%. Second, we used the log of total remit-
ances instead of remittances as a proportion of GDP.
Using the logs helps reduce variability and minimises possi-
le heteroscedastic tendencies. Third, instead of using standard
eviation we used a five-year rolling variance as a measure
f risk. Fourth, we included a dummy for financial develop-
ent (FINCDEVM) based on the median level of financial
 FDI DCPS REER MONEY FREEDOM
4 0.0069 0.0015* 0.0153 −0.0266
) (0.0255) (0.0009) (0.0727) (0.0257)
3 −0.0001 6.34E−06 −0.0006 0.00027
) (0.0003) (1.1E−05) (0.0009) (0.0003)
 −0.7893 −0.1960 27.0516*** −33.6473***
) (4.0479) (0.1374) (11.5601) (4.0783)
*** 0.0074 0.0002 0.0046 0.0043
) (0.0050) (0.0002) (0.0143) (0.0050)
9 0.58262*** 0.0001 0.201052* −0.0315
) (0.0371) (0.0013) (0.1060) (0.0374)
1 −0.2016 0.9175*** −0.6407 −0.0559
) (0.3811) (0.0129) (1.0882) (0.3839)
 −0.0005 0.0003 0.8319*** 0.00716
) (0.0084) (0.0004) (0.0241) (0.0085)
 −0.0035 −0.0004 0.0762 0.6573***
) (0.0294) (0.0010) (0.0839) (0.0296)
2 0.34435 −0.0158 −0.3477 0.4131
) (0.3937) (0.0134) (1.1244) (0.3968)
 0.5603 0.1056*** 0.6661 0.1823
) (0.6192) (0.0210) (1.7682) (0.6238)
3 1.5654 0.2408*** 10.7789 25.4876***
) (2.4079) (0.0817) (6.8764) (2.4259)
 0.3804 0.9714 0.7407 0.7804
 0.3674 0.9708 0.7352 0.7758
.31 7122.687 8.2044 58,091.36 7230.140
 3.864228 0.1311 11.0356 3.8933
67*** 29.28925*** 1620.008*** 136.2276*** 169.4904***
.979 −1346.539 304.4558 −1858.629 −1350.192
 5.56368 −1.2027 7.6624 5.5787
 5.6581 −1.1082 7.7569 5.6731
the real effective exchange rate; TRADE, total trade as a proportion of GDP;
y equal to 1 (high financial development) if a country’s financial development
l development); DREMITM, a remittance dummy equal to 1 (high remittance
d zero otherwise (low remittance receiving country); MONEY FREEDOM is
The figures in parentheses are standard errors.
1 velop
d
m c
m t
( t
t e
r k
r 2
r n
t t
m m
b c
e i
t
d s
i
c c
i o
c f
e b
m h
a d
a
o c
t c
T g
s t
( s
c
i
i o
a t
p fi
m
i n
r (
c b
t o
t v
p
4
a
t
i t
o t
n p
r T
b b
m p
fi l
i r
t
i
t A
m
r
c
c
02 H. Issahaku et al. / Review of De
evelopment. FINCDEVM equals 1 (high financial develop-
ent) when a country’s financial development is above the
edian level of financial development and zero otherwise
low financial development). In addition, we examined whether
he amount of remittances received matters by including a
emittance dummy (DREMITM). DREMITM equals 1 (high
emittance receiving countries) when a country’s remittance
eceipts exceed the median level and zero otherwise (low remit-
ance receiving countries). Finally, we included a new variable,
onetary freedom (MONEY FREEDOM), to test for possi-
le omitted variable bias. We report the results of the PVAR
stimation in Table 7.
In the first column in which the variance of remittances is the
ependent variable, the financial development variable (DCPS)
s no longer significant after accounting for the level of finan-
ial development. However, the financial development dummy
s negative and significant. This finding remains robust after
ontrolling for the amount of remittance received. This can be
xplained by noting that in countries with shallow financial
arkets–where the cost of credit in the formal circuit is high
nd access is limited–households rely upon remittances as an
lternative mode of finance. This understanding dovetails with
ur previous conclusion that remittances can serve as a substi-
ute for bank credit when the financial system is underdeveloped.
he economic openness variable (TRADE) remains positive and
ignificant as shown earlier. The real effective exchange rate
REER) also remains significantly negative.
In the second column in which the variance of the lending rate
s the dependent variable, the variance of remittances is signif-
cant and negative after accounting for the level of remittances
nd the level of financial advancement. This result supports the
revious finding that remittances help to mitigate macroecono-
ic volatility. In addition, the remittance dummy is not signif-
cant, implying that the macroeconomic smoothening effect of
emittances pertains in both low- and high-remittance receiving
ountries. From our robustness checks, we can fairly conclude
hat the results of this study are robust to alternative specifica-
ions of remittances, monetary policy and financial development.
.  Conclusions
Remittances continue to play an increasingly important role
n developing countries and are becoming a dominant source
f development finance, which has implications for macroeco-
omic policy. We find a complex web of relationships among
emittances, monetary policy and financial markets. Notably,
oth remittances and remittance volatility tend to reduce both the
onetary policy rate and monetary policy volatility. First, this
nding implies that in the presence of remittances, the domestic
nterest rate becomes downward biased; in other words, remit-
ance inflows will lead to favourable reductions in domestic
nterest rates, thereby reducing financing costs. Second, remit-
ances are countercyclical and have a smoothening effect onacroeconomic magnitudes, which means that the presence of
emittances can reduce macroeconomic fluctuations, thereby
reating favourable economic conditions for the pursuit of poli- e
ies that deliver shared prosperity. mment Finance 6 (2016) 91–104
This paper highlights the important role played by the finan-
ial sector in the remittance-monetary policy nexus. We find
hat financial development helps to reduce monetary policy risk
hrough its interaction with remittances. This finding supports
arlier studies that endorse the moderating role of financial mar-
ets in the finance-growth relationship (see, Agbloyor et al.,
014; Osabuohien and Efobi, 2013). However, we establish a
egative association between financial development and remit-
ances. Our robustness checks help us explain this finding to
ean that in countries with weak financial systems, the high
ost of sending and receiving remittances obstructs remittance
nflows. In addition, in undeveloped financial markets, domes-
ic residents rely on their offshore benefactors as an alternative
ource of income.
Our simulation of contractionary monetary policy reveals that
ontractionary monetary impulses engineer a persistent inflow
f remittances. We believe this finding is relevant in terms of
ormulating monetary policy. Central banks ought to factor this
ehaviour of remittances into their policy decisions and may
ave to think about sterilisation (when required) to achieve the
esired policy outcomes.
These findings imply that one of the ways developing
ountries can diminish monetary policy risks is to pursue poli-
ies that facilitate remittance inflows. Such policies should be
eared towards reducing the cost of sending and receiving remit-
ances by providing innovative financial products for remittance
enders and recipients alike and by encouraging the use of formal
hannels for transmitting remittances.
Our findings are largely robust to an alternative specification
f remittances and monetary policy, when additional explana-
ory variables are included and after controlling for the level of
nancial development and the level of remittances received.
This work corroborates earlier studies on the finance-growth
exus by Bugamelli and PaternÒ (2011) and Craigwell et al.
2010). However, although these studies establish a relationship
etween remittance volatility and output volatility (an indirect
utcome of monetary policy), we assess the impact of remittance
olatility on a direct measure of monetary policy – the monetary
olicy rate and its volatility.
Our paper extends the literature on international capital flows
nd macroeconomic stability by using a panel vector approach
o establish the impact of remittance and its volatility on domes-
ic monetary conditions. We contribute to the advancement of
heory by simulating the impact of a contractionary monetary
olicy based on the Mundell–Fleming–Dornbush hypothesis.
he impulse responses generated allowed us to understand the
ehaviour of remittances in the presence of domestic monetary
olicy shocks. In conclusion, this study, while supporting ear-
ier findings, offers new insights into the link between migrant
emittances and macroeconomic stability.
cknowledgmentsThe authors acknowledge very useful comments from the
ditors of the Journal and two anonymous referees. Their com-
ents have enriched the paper.
A
T
L
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
R
A
A
A
A
A
A
A
A
A
A
A
A
B
B
H. Issahaku et al. / Review of Development Finance 6 (2016) 91–104 103
ppendix.
able A1
ist of countries included in the study.
. Algeria 21. China 41. Guyana 61. Moldova 81. Samoa 101. Uganda
. Antigua and Barbuda 22. Colombia 42. Honduras 62. Mongolia 82. Sao Tome and Principe 102. Ukraine
. Argentina 23. Congo Republic 43. Hungary 63. Morocco 83. Senegal 103. Vanuatu
. Armenia 24. Costa Rica 44. India 64. Mozambique 84. Seychelles 104. Venezuela, RB
. Azerbajan 25. Cote d’Ivoire 45. Indonesia 65. Namibia 85. Sierra Leone 105. Vietnam
. Bangladesh 26. Croatia 46. Iran 66. Nepal 86. Solomon Islands 106. Yemen
. Barbados 27. Djibouti 47. Jamaica 67. Nicaragua 87. South Africa
. Belarus 28. Dominica 48. Jordan 68. Niger 88. Sri Lanka
. Belize 29. Dominican Republic 49. Kazakhstan 69. Nigeria 89. St. Lucia
0. Benin 30. Equador 50. Kenya 70. Oman 90. St. Vincent and the Grenadines
1. Bolivia 31. Egypt 51. Kyrgyz Republic 71. Pakistan 91. Sudan
2. Bosnia and Herzegovina 32. El Salvador 52. Lao PDR 72. Panama 92. Suriname
3. Botswana 33. Ethiopia 53. Latvia 73. Papua New Guinea 93. Tajikistan
4. Brazil 34. Fiji 54. Lesotho 74. Paraguay 94. Tanzania
5. Bulgaria 35. Georgia 55. Macedonia, FYR 75. Peru 95. Thailand
6. Burkina Faso 36. Ghana 56. Malawi 76. Philippines 96. Togo
7. Burundi 37. Grenada 57. Malaysia 77. Poland 97. Tonga
8. Cabo Verde 38. Guatemala 58. Maldives 78. Romania 98. Trinidad and Tobago
9. Cambodia 39. Guinea 59. Mali 79. Russia 99. Tunisia
0. Cameroon 40. Guinea-Bissau 60. Mexico 80. Rwanda 100. Turkey
eferences Brown, R.P.C., Carmignani, F., Fayad, G., 2013. Migrants’ remittances
and financial development: macro- and micro-level evidence of a per-
costa, P., Calderon, C., Fajnzylber, P., Lopez, H., 2008. What is the impact of verse relationship. World Econ. 36 (5), 636–660, http://dx.doi.org/
international remittances on poverty and inequality in Latin America? World 10.1111/twec.12016.
Dev. 36 (1), 89–114. Bugamelli, M., PaternÒ, F., 2011. Output growth volatility and remit-
costa, P.A., Fajnzylber, P., Lopez, H., 2007. The impact of remittances on tances. Economica 78 (311), 480–500, http://dx.doi.org/10.1111/
poverty and human capital: evidence from Latin American household sur- j.1468-0335.2009.00838.x.
veys. World Bank Policy Research Working Paper 4247. Canova, F., Ciccarelli, M., 2013. Panel Vector Autoregressive Models: A Survey.
dams Jr., R.H., 2004. Remittances and poverty in Guatemala. World Bank Emerald Group Publishing Limited.
Policy Research Working Paper 3418. Carletto, C., Covarrubias, K., Maluccio, J.A., 2011. Migration and child growth
dams, R.H., 2006. International remittances and the household: analysis and in rural Guatemala. Food Policy 36 (1), 16–27.
review of global evidence. J. Afr. Econ. 15 (2), 396–425. Chami, R., Barajas, A., Cosimano, T., Fullenkamp, C., Gapen, M., Montiel, P.,
dams, R.H., Cuecuecha, A., 2010. Remittances, household expenditure and 2008. Macroeconomic Consequences of Remittances. International Mone-
investment in Guatemala. World Dev. 38 (11), 1626–1641. tary Fund, Washington, DC.
dams, R.H., Page, J., 2005. Do international migration and remittances reduce Chami, R., Hakura, D.S., Montiel, P.J., 2012. Do worker remittances reduce
poverty in developing countries? World Dev. 33 (10), 1645–1669. output volatility in developing countries? J. Glob. Dev. 3 (1), 1–25.
denutsi, D.E., 2014. Macroeconomic determinants of workers’ remittances Craigwell, R., Jackman, M., Moore, W., 2010. Economic volatility and
and compensation of employees in Sub-Saharan Africa. J. Dev. Areas 48 remittances. Int. J. Dev. Issues 9 (1), 25–42, http://dx.doi.org/10.1108/
(1), 337–360, http://dx.doi.org/10.1353/jda.2014.0015. 14468951011033789.
denutsi, D.E., Ahortor, C.R.K., 2008. Remittances, exchange rate, and mone- El-Sakka, M.I.T., McNabb, R., 1999. The macroeconomic determinants of emi-
tary policy in Ghana. West Afr. J. Monet. Econ. Integr. 8 (2), 1–42. grant remittances. World Dev. 27 (8), 1493–1502.
gbloyor, E.K., Abor, J.Y., Adjasi, C.K.D., Yawson, A., 2014. Private capital Ericsson, J., Irandoust, M., 2004. The productivity-bias hypothesis and
flows and economic growth in Africa: the role of domestic financial markets. the PPP theorem: new evidence from panel vector autoregressive mod-
J. Int. Financ. Mark. Inst. Money 30, 137–152. els. Japan World Econ. 16 (2), 121–138, http://dx.doi.org/10.1016/
muedo-Dorantes, C., Pozo, S., 2006. Migration, remittances, and male and s0922-1425(03)00015-x.
female employment patterns. Am. Econ. Rev. 96 (2), 222–226. Fayissa, B., Nsiah, C., 2010. The impact of remittances on economic growth
ndrews, D.W.K., Lu, B., 2001. Consistent model and moment selec- and development in Africa. Am. Econ. 92–103.
tion procedures for GMM estimation with application to dynamic panel Giuliano, P., Ruiz-Arranz, M., 2009. Remittances, financial development,
data models. J. Econom. 101 (1), 123–164, http://dx.doi.org/10.1016/ and growth. J. Dev. Econ. 90 (1), 144–152, http://dx.doi.org/10.1016/
S0304-4076(00)00077-4. j.jdeveco.2008.10.005.
ntón, J., 2010. The impact of remittances on nutritional status of children in Gordon, D.B., Leeper, E.M., 1994. The dynamic impacts of monetary policy:
Ecuador. Int. Migr. Rev. 44 (2), 269–299.
arajas, A., Chami, R., Fullenkamp, C., Gapen, M., Montiel, P.J., 2009. Do G
workers’ remittances promote economic growth? IMF Working Papers
09/153.
ernanke, B.S., Blinder, A.S., 1992. The federal funds rate and the channels of G
monetary transmission. Am. Econ. Rev. 82 (4), 901–921.an exercise in tentative identification. J. Polit. Econ. 106 (2), 1228–1247.
rossmann, A., Love, I., Orlov, A.G., 2014. The dynamics of exchange rate
volatility: a panel VAR approach. J. Int. Financ. Mark. Inst. Money 33,
1–27, http://dx.doi.org/10.1016/j.intfin.2014.07.008.
upta, S., Pattillo, C.A., Wagh, S., 2009. Effect of remittances on poverty and
financial development in Sub-Saharan Africa. World Dev. 37 (1), 104–115.
1 velop
H O
P
J
R
J
R
K
L
R
L S
S
M
T
M T
M U
M
V
M
N W
N
04 H. Issahaku et al. / Review of De
o, T., Yeh, K., 2010. Measuring monetary policy in a small open economy
with managed exchange rates: the case of Taiwan. South. Econ. J. 76 (3),
811–826.
ohansen, S., 1988. Statistical analysis of cointegration vectors. J. Econ. Dyn.
Control 12 (2), 231–254.
ohansen, S., Juselius, K., 1990. Maximum likelihood estimation and inference
on cointegration—with applications to the demand for money. Oxf. Bull.
Econ. Stat. 52 (2), 169–210.
rause, S., Rioja, F., 2006. Financial Development and Monetary Policy Effi-
ciency. Department of Economics, Georgia State University.
i, L., Zhang, N., Willett, T.D., 2012. Measuring macroeconomic and financial
market interdependence: a critical survey. J. Financ. Econ. Policy 4 (2),
128–145.
ove, I., Zicchino, L., 2006. Financial development and dynamic investment
behavior: evidence from panel VAR. Q. Rev. Econ. Financ. 46 (2), 190–210,
http://dx.doi.org/10.1016/j.qref.2005.11.007.
andelman, F.S., 2013. Monetary and exchange rate policy under remit-
tance fluctuations. J. Dev. Econ. 102, 128–147, http://dx.doi.org/10.1016/
j.jdeveco.2012.02.006.
andelman, F.S., Zlate, A., 2012. Immigration, remittances and busi-
ness cycles. J. Monet. Econ. 59, 196–213, http://dx.doi.org/10.1016/
j.jmoneco.2012.01.004.
ansuri, G., 2006. Migration, School Attainment, and Child Labor: Evidence
from Rural Pakistan. World Bank Policy Research Working Paper 3946.
cCormick, B., Wahba, J., 2000. Overseas employment and remittances to a
dual economy. Econ. J. 110 (463), 509–534.
ountford, A., 2005. Leaning into the wind: a structural VAR investigation of
UK monetary policy. Oxf. Bull. Econ. Stat. 67 (5), 597–621.
cube, M., Brixiova, Z., 2013. Remittances and their Macroeconomic Impact:
Evidence from Africa. African Development Bank Working Paper 188.siah, C., Fayissa, B., 2011. Remittances and economic growth in Africa,
Asia, and Latin American-Caribbean countries: a panel unit root Y
and panel cointegration analysis. J. Econ. Finance 37 (3), 424–441,
http://dx.doi.org/10.1007/s12197-011-9195-6.ment Finance 6 (2016) 91–104
sabuohien, E.S., Efobi, U.R., 2013. Africa’s money in Africa. S. Afr. J. Econ.
81 (2), 292–306.
radhan, G., Upadhyay, M., Upadhyaya, K., 2008. Remittances and economic
growth in developing countries. Eur. J. Dev. Res. 20 (3), 497–506.
afiq, M.S., Mallick, S.K., 2008. The effect of monetary policy on out-
put in EMU3: a sign restriction approach. J. Macroecon. 30 (4),
1756–1791.
amirez, M.D., 2013. Do financial and institutional variables enhance the impact
of remittances on economic growth in Latin America and the Caribbean?
A panel cointegration analysis. Int. Adv. Econ. Res. 19 (3), 273–288,
http://dx.doi.org/10.1007/s11294-013-9407-2.
uiz, I., Vargas-Silva, C., 2010. Monetary policy and international remittances.
J. Dev. Areas 43 (2), 173–186.
enbeta, S., 2011. How applicable are the new Keynesian DSGE models to a
typical low-income economy? MPRA Paper 30931.
inger, D.A., 2010. Migrant remittances and exchange rate regimes in the devel-
oping world. Am. Polit. Sci. Rev. 104 (02), 307–323.
aylor, E.J., 1999. The new economics of labour migration and the role of
remittances in the migration process. Int. Migr. 37 (1), 63–88.
ermos, A., Naufal, G., Genc, I., 2013. Remittance outflows and inflation: the
case of the GCC countries. Econ. Lett. 120 (1), 45–47, http://dx.doi.org/
10.1016/j.econlet.2013.03.037.
hlig, H., 2005. What are the effects of monetary policy on output?
Results from an agnostic identification procedure. J. Monet. Econ. 52 (2),
381–419.
acaflores, D.E., 2012. Remittances, monetary policy, and partial ster-
ilization. South. Econ. J. 79 (2), 367–387, http://dx.doi.org/10.4284/
0038-4038-2011.147.
aheed, A., 2004. Foreign capital inflows and economic growth of developing
countries: a critical survey of selected empirical studies. J. Econ. Coop. 25
(1), 1–36.
ang, D., 2008. International migration, remittances and household investment:
evidence from Philippine migrants’ exchange rate shocks. Econ. J. 118 (528),
591–630.