IDEC Development Macroeconomics KODIRIN - M090583

Term Paper
The Relationship between Money and Prices in Indonesia
Abstract
Controlling inflation is the main objective of the central bank of Indonesia (Ba
nk Indonesia). According to the quantity theory of money, inflation results sole
ly from a maintained expansion of the money stock at rates in excess of increase
s in the amount of money demanded in the economy. The paper examines the money-p
rice relationship in Indonesia using Autoregression test. The results do not sup
port the quantity theory of money.
Key words: money supply, inflation, autoregression, Indonesia
1. Introduction
Ｉnflation is a persistent rise of prices. An indicator to measure inflation is the
Consumer Price Index (CPI). Price movements of goods and services are reflected
on the changes of the CPI over time. Inflation rate control is essential becaus
e high, unstable inflation will have negative impact on the social and economic
condition of the people. First, high inflation decreases real incomes and livin
g standards so that poor people will suffer more. Second, unstable inflation ma
y hamper economic growth due to uncertainty in the economy. In continuing inflat
ion condition, it is hard to interpret the information conveyed by the prices of
goods and services. So, it complicates decision making for consumers, business
es and government. Third, a higher level of domestic inflation in comparison to
neighbouring countries will make domestic interest rates uncompetitive, which m
ay lead to pressure on the rupiah exchange rate.ã According to quantity theory of mon
ey, increase of the money stock more than the amount of money demanded in the ec
onomy is the cause of inflation.
This paper attempts to estimate money-price relationship in Indonesia. Price sta
bility is one of important objectives of monetary policy. If the empirical resu
lts show a strong and robust relationship between money supply and prices, centr
al bank can control inflation by controlling the supply of money.
The outline of this paper is as follows: (1) introduction; (2) the quantity theo
ry of money; (3) literature review; (4) monetary policy in Indonesia; (5) method
ology; (6) empirical results; and (7) conclusions and policy recommendations.
2. The Quantity Theory of Money
According to the quantity theory of money:
MV = PY (1)
Where:
ï ¬ M is money supply;
ï ¬ V is the income velocity of money;
ï ¬ P is the price level; and
ï ¬ Y is the income level.
The quantity theory of money assumes full employment in the economy. Velocity al
so remains stable at least in the short-run. So, Y and V do not change. Among th
e variables in equation (1), only M and P are not constant. Thus, increase in m
oney supply causes increase in prices.
Equation (1) can be rewritten as:
P = MV/Y (2)
As V and Y are assumed to be constant, we can rewrite equation (2) as:
P = f (M) (3)
The classical view can also be expressed by the Cambridge Cash Balance Equation:
M = kPY (4)
Where, k is the desired cash ratio. With the assumptions that (K) and (Y) are co
nstant, equation (4) also leads to equation (3).
2. Literature Review
Mohamed Benbouziane and Abdelhak Benamar conducted an empirical study on the rel
ationship between money and prices in Algeria, Morocco, and Tunisia. The data co
vers the period from 1975 to 2003 for the case of Algeria and Morocco. All the
variables are taken on annual basis. However, in the case of Algeria the post r
eform period (i.e., 1990-2003) using quarterly data. For Tunisia, Quarterly data
are used for the study and the sample period is 1987-2003.
As far as the Granger causality tests are concerned, the results for both Morocc
o and Tunisia show that the causation is from money (i.e., M1) to prices. This t
ends to support the quantity theoristâ s view that the causal relation between money
and prices is from the former to the latter. For Algeria, however the results sh
ow no causation between money (M1) and prices, i.e., the two variables are indep
endent. One possible explanation of this is that the data for the Consumer Price
Index (CPI) are not reliable mainly. This might be true given that the prices w
hich are reported by the authorities are always lower than those paid by the peo
ple.
The causality tests are also made between M2 and CPI. For Algeria, the results r
eveal different outcomes. There is no causality between money and prices in the
first period; however, after the reform period, M2 and the price level are coint
egrated with significant feedback. Surprisingly, for the case of Morocco and Tun
isia conflicting results are found. There is no causation between M2 and the pri
ce level in Morocco, and we could only accept this hypothesis at the 10 % confid
ence level for the case of Tunisia.
4. Monetary Policy Framework in Indonesia
Bank Indonesia has the authority to conduct monetary policy in Indonesia through
the establishment of monetary targets. According to Act No. 3 of 2004, the goal
of Bank Indonesia is to achieve and maintain the stability of the rupiah. One d
efinition of rupiah stability is the stability of prices for goods and services
reflected in inflation. To achieve this goal, Bank Indonesia adopted the inflati
on targeting framework in 2005.
Under this framework, the inflation targetã is announced explicitly to the public. Th
en, Bank Indonesia conducts monetary policy with a forward-looking approach to a
chieve this target. This means that Bank Indonesia may change some policies to e
nsure that the target will be achieved. Inflation targeting framework leads to
transparency and accountability in conducting monetary policy. At the operation
al level, Bank Indonesia sets BI Rate. This will influence money market rates an
d in turn the deposit rates and lending rates in the banking system. Changes in
these rates will ultimately influence output and inflation.
Figure 1
Monetary Policy Transmission Framework
Source: Bank Indonesia
From the figure we can see that Bank Indonesia does not incorporate money supply
(M1 or M2) in the monetary policy transmission framework.
5. Methodology
5.1. Definitions of variables
Two measures of money supply are used, i.e. narrow money (M1) and broad money (M
2). M1 includes currency held by non-bank public and demand deposit held at the
monetary sector. M2 consists of M1 and time deposits held at commercial banks. B
oth the measures of money supply are used to see their influences on prices. As
far as prices are concerned, we will be using the Consumer Price Index (CPI).
5.2. Database
The source of the data is International Financial Statistic (IFS). I used monthl
y data from year 2000 to 2009.
5.3. Objective
The aim of this research is to examine whether increase in money supply raises p
rices level in Indonesia.
5.4. Research Question
Does increase in money supply raise prices level in Indonesia?
5.5. Hypothesis
Increase in money supply raises prices level in Indonesia?
5.3. Analysis tool.
Eviews program is utilized to conduct Augmented Dicky Fuller unit test r
oot, Johansen cointegration test, Granger causality test, and Autoregression tes
t.
6. Empirical Results
First, to test the unit roots in the logarithm of the price level (LNCPI) and th
e money supplies (LNM1, LNM2), the Augmented Dickey Fuller (ADF) was used. The r
esults of the test are presented in Table (1). The results show that LNCPI, LNM
1, and LNM2 are not stationary at level, but they are stationer at the first dif
ference.ã
Table 1
Statistics for ADF Unit Root Tests
Variable Level First difference
Lag t-ADF Lag t-ADF
LNCPI 0 -1.146381 0 -9.064752**
LNM1 0 -0.457084 0 -14.20359**
LNM2 0 1.196375 0 -11.24372**
Note: The criterion for lag selection is based on the Schwarz info criterion. *,
**, denote rejection at 5 percent and 1 percent respectively
Second, in testing for cointegration, Johansen cointegration test was used. resu
lts are presented in table (2a) and (2b).
Table 2a
Johansen Cointegration Test: LNCPI and LNM1
Unrestricted Cointegration Rank Test (Trace)

Hypothesized Trace 0.05

No. of CE(s) Eigenvalue Statistic Critical Value Prob.**

None 0.070222 9.735753 15.49471 0.3016

At most 1 0.010348 1.217009 3.841466 0.2699

Trace test indicates no cointegration at the 0.05 level

Table 2a
Johansen Cointegration Test: LNCPI and LNM2
Unrestricted Cointegration Rank Test (Trace)

Hypothesized Trace 0.05

No. of CE(s) Eigenvalue Statistic Critical Value Prob.**

None 0.041477 4.978919 15.49471 0.8111

At most 1 0.000193 0.022578 3.841466 0.8805

Trace test indicates no cointegration at the 0.05 level

The results show that the logarithm of the money supply (LNM1) is not cointegrat
ed with the logarithm of price level (LNCPI). The logarithm of the money supply
(LNM2) is not cointegrated with the logarithm of price level (LNCPI) as well.
A condition for testing for the Granger causality between two variables is that
they must be stationary. In the previous section, we found that the first diffe
rence of the logarithm of the price level (LNCPI) and the first difference of th
e logarithm of the money supply (LNM1, LNM2) are stationary. The results of Gra
nger causality tests are presented in table (3a) and (3b).
Table 3a
Granger Causality Tests: DLNM1 and DLNCPI

Null Hypothesis: Obs F-Statistic Prob.

 DLNM1 does not Granger Cause DLNCPI 117 0.49343 0.6119
DLNCPI does not Granger Cause DLNM1 4.04633 0.0201

Table 3b
Granger Causality Tests: DLNM2 and DLNCPI

Null Hypothesis: Obs F-Statistic Prob.

DLNCPI does not Granger Cause DLNM2 117 0.96729 0.3833

DLNM2 does not Granger Cause DLNCPI 4.34094 0.0153

The results show that the direction of causation are from the first difference o
f the logarithm of the price level (DLNCPI) to the first difference of the logar
ithm of the money supply (DLNM1) and from the first difference of the logarithm
of the money supply DLNM2 to the first difference of the logarithm of the price
level (DLNCPI).ã However, because the variables LNM1 and LNCPI are not cointegrated,
and LNM2 and LNCPI are not cointegrated as well, we cannot rely on these Granger
causality results. Instead, autoregression model should be utilized to test the
direction of causation. The results of Autoregression tests with the logarithm
of CPI as dependent variable and the logarithm of M1 and M2 as the independent v
ariables are presented in Table (4a) and (4b), respectively.
Table 4a
Autoregression: Dependent Variable LNCPI

Variable Coefficient Std. Error t-Statistic Prob.

LNM1 0.023202 0.024909 0.931494 0.3535

C -95.15957 5129.281 -0.018552 0.9852
AR(1) 1.000066 0.003453 289.6604 0.0000

R-squared 0.998713 Mean dependent var 4.573452

Adjusted R-squared 0.998691 S.D. dependent var 0.248967
F-statistic 45001.86 Durbin-Watson stat 1.622507
Table 4a
Autoregression: Dependent Variable LNCPI

Variable Coefficient Std. Error t-Statistic Prob.

LNM2 0.017590 0.054547 0.322466 0.7477

C 5.980845 1.841686 3.247483 0.0015
AR(1) 0.995992 0.003457 288.1187 0.0000

R-squared 0.998719 Mean dependent var 4.573452

Adjusted R-squared 0.998697 S.D. dependent var 0.248967
F-statistic 45227.85 Durbin-Watson stat 1.672053
The results show that neither the logarithm of M1 (LNM1) nor the logarithm of M2
 (LNM2) is the determinant of the logarithm of CPI (LNCPI). These results indic
ate that the Quantity Theory of Money is not applicable in Indonesia.
7. Conclusions and Policy recommendations
This paper examined the causality issue between money and prices in Indonesia fr
om year 2000 to 2009 empirically. The results of the Augmented Dicky Fuller uni
t root test shows that the logarithm of the price level (LNCPI) and the logarith
m of money supplies (LNM1, LNM2) are stationary at the first difference. The Joh
ansen cointegration test shows that there is no cointegration between LNCPI and
LNM1 and between LNCPI and LNM2. This means that the Granger causality test is
not proper to test the direction of causality between LNCPI and LNM1 and between
LNCPI and LNM2. The proper test to examine this direction is the Autoregression
model. The results of the Autoregression test show that neither LNM1 nor LNM2
is the determinant of LNCPI. This conclusion is not in line with the hypothesis
that based on the Quantity theory of money, which proposes that increase in mon
ey supply will raise prices level.
On the other hand, these results support the monetary policy transmission framew
ork implemented by Bank Indonesia. At the operational level, the monetary polic
y stance is reflected in the setting of the policy rate (BI Rate) instead of the
money supply (M1 or M2).
Reference:
Benbouziane, Mohamed and Benamar, Abdelhak. The relationship between money and p
rices in the maghreb countries: a cointegration analysis. Munich Personal RePEc
Archive, 2004.
www.bi.go.id.