Content uploaded by Mori Kogid
Author content
All content in this area was uploaded by Mori Kogid
Content may be subject to copyright.
71
Does Money Lead Prices in Malaysia? A Bivariate and Trivariate Analysis
Mohd Fahmi Ghazali
*,
Mori Kogid
**
and Hanudin Amin
***
Does Money Lead Prices in Malaysia?
A Bivariate and Trivariate Analysis
This paper reinvestigates the relationship between money supply and prices in Malaysia.
The sample includes monthly data of money supply (M1, M2 and M3), Consumer Price
Index (CPI) and Industrial Production Index (IPI) from January 1974 to April 2007. Both
Johansen cointegration and Autoregressive Distributed Lag (ARDL) bounds testing
method suggest that there is a long-run equilibrium relationship between money supply
and prices in Malaysia. Toda-Yamamoto causality test results reveal that there is a
unidirectional causality running from money supply to CPI, and this supports the
quantity theorist's view, even when domestic real activity variable is included as a control
variable.
* Lecturer, Labuan School of International Business and Finance, Universiti Malaysia Sabah, L abuan
International Campus, Jalan Sg. Pagar, 87000 Labuan F T, Malaysia; and is the corresponding author.
E-mail: mohdfahmi_ghazali@yahoo.com
** Lectur er, Uni ve rsiti Malaysia S ab ah, J alan UM S, 88 40 0, Kota Kina ba lu , Sab ah , Malay sia.
E-mail: edy@ums.edu.my
** * Lecturer, Labuan School of International Business and Finance, Universiti Malaysia Sabah, Labuan
International Campus, Jalan Sg. Pagar, 87000 Labuan F T, Malaysia. E-mail: hanudin_zu@yahoo.com
© 2011 IUP. All Rights Reserved.
Introduction
This paper investigates the relationship between money supply and prices in Malaysia, using
monthly data over the period January 1974-April 2007. Nowadays, empirical analysis on money
and prices relationship has received greater attention, as there is a move to assign the single
objective to the central bank. Among likely candidates of monetary policy objective, price
stability is the single most important objective. Assignment of price stability as the single objective
of monetary policy hinges on the empirical strength of money-Consumer Price Index (CPI)
relationship. If empirical result shows a strong and robust relationship between money supply
and CPI, then the central bank can opt for price stability as its single objective.
Over the past two decades, policy makers have become more aware of the social and
economic cost of inflation and are more concerned with a stable price level as the goal of
economic policy. Price stability is desirable because a rise in price level (inflation) creates uncertainty
in the economy, and that may hamper the economy’s growth. Not only do public opinion
surveys indicate that the public is very hostile to inflation, but also a growing body of evidence
suggests that inflation leads to lower economic growth (Fisher, 1993).
There is a strong positive correlation between changes in money supply and aggregate
prices. Some studies have found a strong, if somewhat delayed, link between changes in the
level of money supply and prices. Because of this, many economists like to keep an eye on
The IUP Journal of Applied Economics, Vol. X, No. 1, 2011
72
changes in money supply since it may indicate changes in the rate of inflation later on (Clayton
and Giesbrecht, 2001). The direction of causality, however, has long been a matter of controversy.
The widely accepted quantity theory of money argues that inflation is caused by exogenous
changes in the money supply. A minority ‘structuralist’ view holds that inflation develops from
pressures arising from economic growth in economies with institutional rigidities, particularly in
agriculture and international transactions. Monetary and fiscal authorities choose to expand the
money supply, ratifying the inflationary pressures, rather than face unemployment or disruptions
in consumption and investment. Underdeveloped financial markets and a weakly independent
central bank can contribute to the likelihood of money supply growth. Under this view, money
supply expansion is a consequence of, and therefore caused by, structural inflation (Pinga and
Nelson, 2001).
The motivation of this study is the need for a further empirical work analyzing the relation
between money supply and prices in Malaysia to reassert that causality exists. More technically,
this research applies the Autoregressive Distributed Lag (ARDL) and Modified-Wald (MWald) test
approach proposed by Toda and Yamamoto (1995), to reinvestigate the cointegration and causal
relations between money supply and prices. This study has also accounted for the possible bias
of using only bivariate framework in causality analysis as highlighted by Al-Yousif (1999), that
bivariate framework is potentially misspecified and may be flawed due to the omission-of-
variable phenomenon. As a result, one would expect biased and mixed results of both causality
and cointegration tests. Thus, this study considers this issue by adding a control variable into the
bivariate VAR between money and prices. Following Husain and Abbas (2000), the included
control variable is domestic real activity, and it is consistent with the well-defined theoretical
structure via money demand (supply) function and inflation model. Conventionally, the volume
of money demand (supply) is determined by interest rate (proxied by inflation rate) and real
income (or domestic real activity). Meanwhile, the inflation rate is a function of money supply
and real income. Finally, this study uses the most appropriate monetary target, M1, M2 and M3
for curbing inflation, applying to a longer and more recent period.
The paper is organized as follows: It reviews some related empirical studies, followed by
a brief discussion on the dataset used and the methodology employed in the paper. Subsequently,
the empirical results are discussed, and finally, the conclusion is offered.
Literature Review
The directions of causality between money and prices have been tested in Malaysia over various
periods of time. The results have yielded conflicting evidence. For example, Tan and Baharumshah
(1999) examined the dynamic causal chain among money (M1, M2 and M3), real output, interest
rate and inflation in Malaysia using monthly data from 1975 to 1995. They found that price
does Granger cause M2 through short-run channel. In addition, the error correction model
provided evidence that real income, interest rate and price do jointly lead M2 in the long run,
and real output, interest rate and M2 do jointly cause price. The study, however, did not directly
state the causal relations between money and price rather than joining the other variables.
Masih and Masih (1998) investigated the causality between money (M1 and M2) and prices
in four Southeast Asian developing countries, namely, Thailand, Malaysia, Singapore, and the
73
Does Money Lead Prices in Malaysia? A Bivariate and Trivariate Analysis
Philippines from January 1961 to April 1990. They found that money supply leads prices, which
is in agreement with the monetarist’s view.
Pinga and Nelson (2001), on the other hand, examined the relationship between money
supply and aggregate prices for 26 countries, and found that no causal relationship exists between
prices and money (M1 and M2) in Malaysia. They also found that aggregate prices cause money
supply in Chile and Sri Lanka, which is in agreement with the structuralist’s view. Evidence of
money supply exogeneity was found to be strongest in Kuwait, Paraguay, and the US. Most
countries exhibited mixed evidence of money supply endogeneity, with bidirectional causation
between money supply and aggregate prices as a common result.
Using MWald test, Tang (2004) examined the causality between money (M2) and prices in
Malaysia from 1970:Q1 to 1998:Q4. He found that there is a unidirectional causality running
from money to prices, that supports the monetarist’s view.
Research on other countries reported both bidirectional and unidirectional causality. Jones
(1989) examined the causality between money and prices in the US over the period 1959:Q1-
1986:Q2. The results, however, show feedback relationship between the measures of money
growth (M1 and M2) and inflation (CPI and Wholesale Price Index, WPI). On the other hand,
Darrat (1986) examined the direction of causation between money and prices for Morocco,
Tunisia and Libya over the period 1960:Q1 and 1980:Q2. The results show a unidirectional
causation running from money to prices without feedback for all the three countries concerned.
He concluded that the results support the monetarist’s view that money causes inflation.
Nepal Rastra Bank (2001) found a feedback interaction in Nepal between money and prices
for the period from 1975:Q3 to 1999:Q2. Recent study by Benbouziane and Benamar (2004) on
the three Maghreb countries found that there is a unidirectional causation from money to prices
in the case of Morocco and Tunisia, supporting Darrat’s (1986) finding. On the other hand, the
results also show the apparent absence of causality between money and prices in the case of
Algeria, which is not easy to explain.
Data and Methodology
Dat a
In order to perform the causality analysis, we employ monthly data for M1, M2, and M3
(in millions of ringgit). The CPI (in 2000 prices) is used as proxy for price. The domestic activity
variable (
Y
) is proxied by industrial production index (in 2000 prices). The dataset was drawn for
the period from January 1974 to April 2007, which comprises 400 observations in total. The
data are obtained from various issues of
International Financial Statistics
(IFS) published by the
International Monetary Fund (IMF) and transformed into natural logarithm scale prior to analysis.
Jo hanse n Cointegration Tests
A preliminary issue regarding the methodological procedure is related to the fact that the data
generating process for most of the economic series exhibits a unit root. Time series properties,
The IUP Journal of Applied Economics, Vol. X, No. 1, 2011
74
namely, order of integration and cointegration, have been examined by applying the full
information multivariate procedure proposed by Johansen (1988).
The cointegration methodology basically characterizes the existence of a long-run relationship.
According to Johansen (1988), a
p
-dimensional Vector Autoregression (VAR) of order
k
[VAR(
k
)]
can be specified as follows:
TtZZdZ t
kkttt
,...,1...
11
...(1)
We can rewrite this expression as,
tit
k
k
i
iktt ZZdZ
1
1
...(2)
Here is the first difference operator, and
are
p
-by-
p
matrices of unknown parameters
and
t
is a Gaussian error term. Long-run information about the relationship between money
and prices in Malaysia is contained in the impact matrix . When the matrix has full column
rank, it implies that all variables in
Zt
are stationar y. When the matrix has zero column rank,
the expression is a first differenced VAR involving no long-run elements. If, however, the rank of
is intermediate, meaning that 0 < rank () =
r < p
, there will be
r
cointegrating vectors that
make the linear combinations of
Zt
stationary or integrated.
There are two Johansen cointegration tests. First, the maximum likelihood estimation
procedure, which includes a likelihood ratio test, called the trace test, that evaluates the null
hypothesis of, at most
r
cointegrating vectors versus the general null of
p
cointegrating vectors.
A second likelihood ratio test is the maximum eigenvalue test, which evaluates the null hypothesis
of
r
cointegrating vectors against the alternative of (
r
+ 1) cointegrating vectors.
Aut oregressive Distributed Lag Bounds Test
In addition to the Johansen cointegration tests, we have also employed the ARDL bounds
testing approach for cointegration, proposed by Pesaran
et al.
(2001) to test for the robustness
of the results. The ARDL equations are given below:
p
i
q
j
tjtjitittt
LCPILMLCPILMML
1 1
11
...(3)
s
i
t
j
tjtjitittt
LMLCPILMLCPICPIL
1 1
11
...(4)
111
tttt
LIPILCPILMML
p
i
q
j
r
k
tktkjtjiti
LIPILCPILM
1 1 1
...(5)
75
Does Money Lead Prices in Malaysia? A Bivariate and Trivariate Analysis
111
tttt
LIPILMLCPILCPI
s
i
t
j
u
k
tktkjtjiti
LIPIoLMLCPI
1 1 1
...(6)
where
and
are the drift components,
LM
,
LCPI
and
LIPI
are, respectively, the logarithm of
money supply (M1, M2 and M3), general price level and industrial production index,
t
is time
period,
p, q, r, s, t
and
u
are the optimal lag lengths, and
t
and
t
are white noise error.
The first step in the ARDL approach is to estimate Equations (3) to (6) using Ordinary
Least Square (OLS). The second step is to trace the presence of cointegration by restricting all
estim ated coefficients of lagged level variables, i.e.,
LMt
– 1 and
LCPIt
– 1 for bivariate ARDL, and
LMt
– 1,
LCPIt
– 1 and
LIPIt
– 1 for trivariate ARDL to zero. That is, the null hypothesis of no cointegration
(
H0
:
=
= 0,
H0
:
=
= 0,
H0
:
=
=
= 0 and
H0
:
=
=
= 0) is tested against the
alternative (
H1
:
0,
H1
:
0,
H1
:
0 and
H1
:
0). If the null
hypothesis of no cointegrating relationship, is rejected against the alternative then there exists a
long-run relationship between the variables. The standard
F
-statistic normally used for this test
has a nonstandard asymptotic distribution. Two sets of critical values are tabulated. One set
assumes that all the variables are I (1), while the other assumes that all the variables are I (0). The
two sets of appropriate critical values provide a band that is used to classify the variables into
I (0) or I (1). A calculated
F
-statistic that lies above the upper level of the band implies that the
variables are cointegrated, while a calculated
F
-statistic that lies below the lower band implies
that the variables are not cointegrated. If the
F
-statistic lies within the band, the result is
inconclusive.
The main advantage of this approach in testing and estimation is that it can be applied
whether the regressors are I (0) or I (1) and avoids the pretest problems associated with standard
cointegration analysis (Pesaran
et al
., 2001). However, Quattara (2004) argues that in the presence
of I (2) variables, the computed
F
-statistics provided by Pesaran
et al.
(2001) are no more valid
because they are based on the assumption that the variables are either I (0) or I (1); therefore, the
implementation of unit root tests in the ARDL procedure might still be necessary in order to
ensure that none of the variables are integrated of order two or beyond. This technique is also
appropriate for small or finite sample size (Pesaran
et al.
, 2001).
Granger Causality Tests Based on Toda-Yamamoto Level VAR
To examine the issue of causation, we have employed a modified version of the Granger causality
test, which is robust for the cointegration features of the process. This procedure was suggested
by Toda and Yamamoto (1995), with the objective to overcome the problem of invalid asymptotic
critical values when causality tests are performed in the presence of non-stationary series. Zapata
and Rambaldi (1997) explained that the advantage of using the Toda-Yamamoto procedure is
that in order to test Granger causality in the VAR framework, it is not necessary to pretest the
The IUP Journal of Applied Economics, Vol. X, No. 1, 2011
76
variables for the integration and cointegration properties, provided the maximal order of integration
of the process does not exceed the true lag length of the VAR model. The Toda-Yamamoto
procedure, however, does not substitute the conventional unit root and cointegration properties
pretesting in time series analysis. They are considered as complementar y to each other.
The Toda-Yamamoto procedure basically involves the estimation of an augmented VAR(
k+dmax
)
model, where
k
is the optimal lag length in the original VAR system, and
dmax
is the maximal
order of integration of the variables in the VAR system. The Toda-Yamamoto procedure uses
MWald test for zero restrictions on the parameters of the original VAR(
k
) model. The remaining
dmax
autoregressive parameters are regarded as zeros and ignored in the VAR(
k
) model. This test
has an asymptotic chi-squared distribution with
k
degrees of freedom in the limit when a
VAR(
k
+
d
max) is estimated. The dynamic causal relationship between prices and money supply
would be as follows:
dk
i
dl
j
tjtjitit
uLCPILMLM
1 1
...(7)
dw
i
dx
j
tjtjitit
LMcLCPIbaLCPI
1 1
...(8)
dk
i
dl
j
t
dm
k
ktkjtjitit
uLIPILCPILMLM
111
...(9)
dw
i
dx
j
t
dy
k
ktkjtjitit
LIPILMcLCPIbaLCPI
1 1 1
...(10)
where
k, l, m, w, x
and
y
are the optimal lag lengths,
d
is the maximal order of integration of the
series in the system, and
u
and are the error terms that are assumed to be white noise. The
initial lag lengths
k, l, m, w, x
and
y
are chosen using the Schwarz Information Criterion (SIC).
However, the initial lag lengths are augmented with extra lag(s) depending on the likely order of
integration (
d
) of the series,
LMt, LCPI t
and
LIPIt
. If
LMt
, LCPI t
and
LIPIt
are likely to be I (1) (as it
is with most macroeconomic data) then one extra lag is added to each variable in Equations (7)
and (8) for bivariate causality and Equations (9) and (10) for trivariate causality. If all variables are
assumed to be I (0), no extra lag is added in the equation, and the Toda-Yamamoto test is
equivalent to the Granger causality test. Wald tests are then used to test the direction of causality.
For example, in Equations (7) and (9), the lags of
LCPIt
, excluding the extra lag added to capture
the maximum order of integration, are tested for their significance. If the null hypothesis that the
lags are jointly equal to zero is accepted, then
LCPIt
does not cause
LMt
. Testing for the joint
significance of
LMt
, excluding the extra lag added, in Equations (8) and (10) allows tests for
unidirectional or bidirectional causality.
77
Does Money Lead Prices in Malaysia? A Bivariate and Trivariate Analysis
Estimation Results
The first stage involves establishing the order of integration using the Augmented Dickey-Fuller
(ADF), with and without a deterministic trend. Table 1 presents the results of the unit root tests
for the four variables,
LM
1,
LM
2,
LM
3,
LCPI
and
LIPI
. The results indicate that all the variables
are non-stationary at their levels. On the other hand, all are stationary at first difference, therefore
indicating that all variables are I (1).
Given the variables are I (1), the cointegration hypothesis between the variables is examined
using the methodology developed by Johansen (1991) in order to specify the long-run relationship
between the variables. The results of the bivariate and trivariate cointegration tests are reported
in Tables 2 and 3. The null hypothesis of no cointegrating vector (
r =
0) is rejected. Thus, the
Va r i a bl es
ADF Unit Root Tests at Level ADF Unit Root Tests at First Differen ce
LM
1 –0.6077 (16) –2.8887 (16) –4.2059 (15)*** –4.2118 (15)***
LM
2 –1.2402 (15) –2.7310 (12) –3.4561 (11)*** –3.7387 (14)**
LM
3 –2.1681 (16) –1.8825 (16) –3.0750 (15)** –3.6375 (15)**
LCPI
–2.0517 (16) –1.3591 (13) –3.8418 (12)*** –3.9980 (12)***
LIPI
–1.0253 (16) –2.6693 (15) –5.2644 (15)*** –5.3031 (15)***
Table 1: Result s of the Unit Root Tests
N o t e : The null hypothesis is that the series is non-stationary, or contains a unit root. The rejection of the null
hypothesis for ADF test is based on the McKinnon critical values. Values in parentheses are optimal lag
lengths.
and
are constant and trend and constant, respectively. *** and ** denote that a test
statistic is significant at 1% and 5% significance levels, respectively.
Table 2: Test ing fo r Biv ari ate Cointegration
N o t e : VAR is order of the variance; *** denotes statistically significant at 1% level; and
r
denotes the number
of cointegrating vectors.
N u ll E i g e n - Tr a c e 5% C ritical Max-Eigenvalue 5 % Crit i c a l VAR
H y po t h e s i s v a l u e St a t i st i c s Val u e S t at i s t i cs Va l u e
LM
1
-LCP I
r
= 0 0.2147 99.7141*** 19.96 95.9423*** 15. 67 2
r
= 1 0.0095 3.7717 9.24 3.7717 9.24
LM
2
-LCP I
r
= 0 0.2167 100.4335*** 19.96 96.9408*** 15.67 2
r
= 1 0.0088 3.4927 9.24 3.4927 9.24
LM
3
-LCP I
r
= 0 0.2227 103.8645*** 19.96 100.0059*** 15.67 2
r
= 1 0.0097 3.8586 9.24 3.8586 9.24
The IUP Journal of Applied Economics, Vol. X, No. 1, 2011
78
money supply (M1, M 2 and M 3) and the price level (CPI) are cointegrated, indicating that there
is a long-run relationship between them.
Table 3: Test ing fo r Trivariat e Coin teg rat ion
N o t e : VAR is order of the variance; *** denotes statistically significant at the 1% level; and
r
denotes the
number of cointegrating vectors.
N u ll E i g e n - Tr a c e 5% C ritical Ma x-Ei g enva lue 5% Cr i t ical VAR
H y po t h e s i s v a l u e S t a ti s t i cs Va l u e S ta t i s ti c s Va l u e
LM
1
-LCP I
r
= 0 0.2544 135.2575*** 34.91 116.5524*** 22.00 2
r
= 1 0.0358 18.7051 19.96 14.4611 15.67
r
= 2 0.0106 4.2439 9.24 4.2439 9.24
LM
2
-LCP I
r
= 0 0.2603 135.3297*** 34.91 119.7149*** 22.00 2
r
= 1 0.0278 15.6148 19.96 11.2101 15.67
r
= 2 0.0110 4.4046 9.24 4.4046 9.24
LM
3
-LCP I
r
= 0 0.2641 143.4559*** 34.91 121.7619*** 22.00 2
r
= 1 0.0425 21.6940 19.96 17.2431 15.67
r
= 2 0.0111 4.4509 9.24 4.4509 9.24
The Johansen-based results are compared to those obtained from the ARDL method, to
investigate the robustness of long-run relationship between the variables. The computed
F
-statistics are presented in Table 4 for bivariate framework and Table 5 for trivariate framework
along with the critical values. The results of the bounds cointegration test show that the null
hypothesis of
=
= 0,
=
= 0,
=
=
= 0 and
=
=
= 0 against the alternative
0,
0,
0 and
0 can be rejected and there exists a steady-state
equilibrium between the variables. As per the results reported in Tables 4 and 5,
F
-statistics
Table 4: Test ing for Coi nte gra tio n
Using the ARD L Appr oac h – Bivariate Framew ork
Depend e n t
Va r i a b l es Variabl e s Added
Jo int Test of Zero Re str icti on of Var iables Ad ded
in Column s 2 and 3
F
-Statist i c s
F
-Crit i c a l
Lo w e r Bo u n d Up per Bo u n d
LM
1
-LCP I
LM
1
LM
1
t –
1
LCPI
t –
10.1864 4.042c4.788c
LCPI LCPI
t –
1
LM
1
t –
110.1507 7.057a7.815a
LM
2
-LCP I
LM
2
LM
2
t –
1
LCPI
t –
12.5923 4.042c4.788c
LCPI LCPI
t –
1
LM
2
t –
17.5853 4.934b5.764b
79
Does Money Lead Prices in Malaysia? A Bivariate and Trivariate Analysis
exceed the upper level of the band respectively at 1%, 5% and 10% levels of significance for
LCPI
equation. The results of the ARDL approach confirm those of the Johansen approach.
Next, the Toda-Yamamoto causality tests are employed to explore the direction of causality.
Since all the variables are at levels, the results provide information about the long-run causal
relationships among non-stationar y variables in the system. The bivariate causality results are
reported in Table 6. The results indicate that the null hypothesis that CPI does not Granger cause
M1, M2 and M3 cannot be rejected. This suggests that money supply does not respond to
lagged changes in CPI in the system. On the other hand, the hypothesis that money supply does
not Granger cause CPI can be rejected at 1% significance level for M1 and M3 and 5% significance
level for M2. This indicates that there is a unidirectional causality running from money supply to
CPI in the case of Malaysia, but not the reverse.
Table 4 (Cont.)
Depend e n t
Va r i a b l es Variabl e s Added
Jo int Test of Zero Re str icti on of Var iables Ad ded
in Column s 2 and 3
F
-Statist i c s
F
-Crit i c a l
Lo w e r Bo u n d Up per Bo u n d
N o t e : denotes first difference; the lag length selection was based on SIC (not reported in this table); a, b, and
c denote
F
-critical values at 1%, 5% and 10% levels respectively.
LM
3-
LC PI
LM
3
LM
3
t –
1
LCPI
t –
18.7167 7.057a7.815a
LCPI LCPI
t –
1
LM
3
t –
17.3130 4.934b5.764b
Table 5: Test ing for Coi nte gra tio n
Us ing the AR DL App roac h – Trivar iat e Fram ework
Depend e n t
Va r i a bl es Variabl e s Added
Jo int Test of Ze ro Res tri ctio n of Vari a ble s
Added in Col umn s 2, 3 and 4
F
-Statist i c s
F
-Crit i c a l
Lo w e r Bo u n d Up per Bo u n d
N o t e : denotes first difference; the lag length selection was based on SIC (not reported in this table); a, b, and
c denote
F
-critical values at 1%, 5% and 10% levels respectively.
LM
1-
LC P I
LM
1
LM
1
t –
1
LIPI
t –
1
LCPI
t –
13.6327 3.182c4.126c
LCPI LCPI
t –
1
LIPI
t –
1
LM
1
t –
15.2287 3.793b4.855b
LM
2
-LCP I
LM
2
LM
2
t –
1
LIPI
t –
1
LCPI
t –
13.218180 3.182c4.126c
LCPI LCPI
t –
1
LIPI
t –
1
LM
2
t –
14.805525 3.182c4.126c
LM
3-
LC P I
LM
3
LM
3t
–
1
LIPI
t –
1
LCPI
t –
17.868306 5.288a6.309a
LCPI LCPI
t –
1
LIPI
t –
1
LM
3t
–
14.785271 3.182c4.126c
The IUP Journal of Applied Economics, Vol. X, No. 1, 2011
80
Finally, Toda-Yamamoto trivariate causality results are presented in Table 7. The table shows
the causal relationship between money and prices conditional on the presence of domestic real
activity (real income). In can be seen that the results are similar to those obtained in the bivariate
case, i.e., strong unidirectional causality from money to prices. Therefore, the empirical evidence
for Malaysia supports the quantity theorist’s view.
Tabl e 6: Test of Cau sality, Toda-Yamamoto Ap proach – Bivaria te Cas e
Depend e n t
Va r i a bl es
In d e p e n d ent
Va r i a b l es
Jo int Test of Zero
Re s t ricti o n of Vari a bles
Added in Colu mn 2
La g
Structure
MWa l d - S t a t i s t i c s
p
-Valu es
N o t e : The [
k
+
dmax
]th order level VAR was estimated with
dmax
=1 since the order of integration is 1; the lag
length selection was based on SIC (not reported in this paper). *** and ** denote statistically significant
at 1% and 5% levels respectively.
VAR
O r d e r
LM
1-
LC PI
LM
1
LCPI
1 (2) 0.110403 0.7399
LCPI LM
1 4 (5) 5.047628*** 0.0006
LM
2-
LC P I
LM
2
LCPI
1 (2) 1.850955 0.1745
LCPI LM
2 2 (3) 4.550242** 0.0111
LM
3-
LC P I
LM
3
LCPI
2 (3) 2.307815 0.1008
LCPI LM
3 2 (3) 5.256416*** 0. 0056
Table 7: Test of Caus ality, Toda-Yamamo to App roa ch – Trivariat e Case
Depend e n t
Va r i a bl es
In d e p e n d ent
Va r i a b l es
Jo int Test of Zero
Re s t ricti o n of Vari a bles
Added in Colu mn 2
La g
Structure
MWa l d - S t a t i s t i c s
p
-Valu es
N o t e : The [
k+dmax
]th order level VAR was estimated with
dmax
=1 since the order of integration is 1; the lag
length selection was based on SIC (not reported in this paper); *** denotes statistically significant at 1%
level.
VAR
O r d e r
LM
1-
LC PI
LM
1
LCPI
1 (2) 0.334183 0.5635
LCPI LM
1 2 (3) 9.777780*** 0.0001
LM
2-
LCPI
LM
2
LCPI
1 (2) 2.715890 0.1002
LCPI LM
2 2 (3) 4.926900*** 0.0077
LM
3
-LCP I
LM
3
LCPI
2 (3) 2.058677 0.1521
LCPI LM
3 2 (3) 5.535062*** 0.0043
81
Does Money Lead Prices in Malaysia? A Bivariate and Trivariate Analysis
Co n clusion
The paper empirically re-examines the relationship between prices and money in Malaysia, including
domestic real activity variable as a control variable. It employs monthly data and tests for
cointegration using the Johansen and ARDL approach, and applies the Toda-Yamamoto causality
approach to analyze the money and prices interaction in Malaysia. Using Johansen and ARDL
cointegration approach, it is obtained that long-run association exists between prices and money,
which is in line with the findings of previous researches in other countries (Husain and Abbas,
2000; and Benbouziane and Benamar, 2004). This implies that prices and money move together
in the long run. Using the powerful causality testing procedure developed by Toda and Yamamoto
(1995), evidence of a unidirectional link from money to prices is found without significant
feedback. This tends to support the quantity theorist’s view that the causal relation between
money and prices is from the former to the latter, which is in line with Masih and Masih (1998)
and Tang’s (2004) findings.
Thus, the monetary authorities should control money supply (M1, M2, and M3) to influence
and control inflation. It is well-known that the objective of the Malaysian monetary policy is to
maintain price stability in the form of low inflation in order to create a stable environment for
sustainable economic growth. As suggested by monetarists, this can be best achieved by
maintaining a steady rate of growth of the money supply, roughly corresponding to the long-
run growth of the real output. That is, the central bank should pursue a constant policy that
accommodates real growth but not inflation.
This study also suggests that the monetary policy should conduct with care. Money supply
can cause domestic price, and the latter variable plays an important role in determining Malaysia’s
import demand in the long run. An increase in domestic price might make imported goods
cheaper. Consequently, it might deteriorate Malaysia’s external balance because of increasing
demand for imported goods.
References
1. Al-Yousif Y K (1999), “On the Role of Exports in the Economic Growth of Malaysia:
A Multivariate Analysis”,
International Economic Journal
, Vol. 13, No. 3, pp. 67-75.
2. Benbouziane M and Benamar A (2004), “The Relationship Between Money and Prices in
the Maghreb Countries: A Cointegration Analysis”, The 11th Annual Conference of the
Economic Research Forum (ERF), December 16-28, Beirut, Lebanon.
3. Brillembourg A and Khan M S (1979), “The Relationship Between Money, Income and
Prices: Has Money Mattered Historically”,
Journal of Money, Credit and Banking,
Vol. 11,
No. 3, pp. 358-365.
4. Clayton G E and Giesbrecht M G (2001),
A Guide to Ever yday Economic Statistics
,
5th Edition, McGraw-Hill/Irwin, New York.
5. Darrat A F (1986), “Money, Inflation and Causality in the North African Countries: An
Empirical Investigation”,
Journal of Macroeconomics,
Vol. 8, No. 1, pp. 87-103.
6. Fisher S (1993), “The Role of Macroeconomic Factors in Growth”,
Journal of Monetary
Economics,
Vol. 32, No. 3, pp. 485-512.
The IUP Journal of Applied Economics, Vol. X, No. 1, 2011
82
7. Husain F and Abbas K (2000), “Money, Income, Prices, and Causality in Pakistan:
A Trivariate Analysis”, Research Report, No. 178, pp. 1-10, Pakistan Institute of Development
Economics.
8. Johansen S (1988), “Statistical Analysis of Cointegration Vectors”,
Journal of Economic
Dynamics and Control,
Vol. 12, Nos. 2 & 3, pp. 231-254.
9. Johansen S (1991), “Estimation and Hypothesis Testing of Cointegration Vectors in Gaussian
Vector Autoregressive Models”,
Econometrica,
Vol. 59, No. 6, pp. 1551-1580.
10. Jones J D (1989), “A Comparison of Lag-Length Selection Techniques in Tests of Granger
Causality Between Money Growth and Inflation: Evidence for the US, 1959-86”,
Applied
Economics,
Vol. 21, No. 6, pp. 809-822.
11. Jones J D and Uri N (1987), “Money, Inflation and Causality: Another Look at the Empirical
Evidence for the USA, 1953-84”,
Applied Economics,
Vol. 19, No. 5, pp. 619-634.
12. Masih A M M and Masih R (1998), “Does Money Cause Prices, or the Other Way Around?
Multi-Country Econometric Evidence Including Error-Correction Modelling from South-
East Asia”,
Journal of Economic Studies,
Vol. 25, No. 3, pp. 138-160.
13. Nepal Rastra Bank (2001), “Money and Price Relationship in Nepal: A Revisit”,
Economic
Review: Occasional Paper,
Vol. 13, pp. 50-65, Research Department, Monetary Division.
14. Pesaran M H, Shin Y and Smith R J (2001), “Bounds Testing Approaches to the Analysis of
Level Relationships”,
Journal of Applied Econometrics
, Vol. 16, No. 3, pp. 289-326.
15. Pinga V E B and Nelson G C (2001), “Money, Prices and Causality: Monetarist Versus
Structuralist Explanations Using Pooled Country Evidence”,
Applied Economics,
Vol. 33,
No. 10, pp. 1271-1281.
16. Quattara B (2004), “The Impact of Project Aid and Programme Aid on Domestic Savings:
A Case Study of
Côte d’Ivoire
”, Centre for the Study of African Economies Conference on
Growth, Poverty Reduction and Human Development in Africa, April.
17. Tan H B and Baharumshah A Z (1999), “Dynamic Causal Chain of Money, Output, Interest
Rate and Prices in Malaysia: Evidence Based on Vector Error-Correction Modeling Analysis”,
International Economic Journal
, Vol. 13, No. 1, pp. 103-120.
18. Tang T C (2004), “Causality Between Money and Price in Malaysia: A Revisit”,
Labuan
Bulletin of International Business & Finance
, Vol. 2, No. 1, pp. 71-81.
19. Toda H Y and Yamamoto T (1995), “Statistical Inference in Vector Autoregressions with
Possibly Integrated Pro ce sses”,
Journal of Econometrics,
Vol. 66, Nos. 1 & 2,
pp. 225-250.
20. Turnovsky S J and Wohar M E (1984), “Monetarism and the Aggregate Economy: Some
Long Run Evidence”,
Review of Economics and Statistics,
Vol. 66, No. 4, pp. 619-629.
21. Zapata H O and Rambaldi A N (1997), “Monte Carlo Evidence on Cointegration and
Causation”,
Oxford Bulletin of Economics and Statistics,
Vol. 52, No. 2, pp. 285-298.
Reference # 05J-2011-01-05-01
Copyright of IUP Journal of Applied Economics is the property of IUP Publications and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.
