[Article]
1
Trapped in Resource Curse After Survived;
Oil Price and Indonesian Macroeconomic Relation before and after Asian Crisis
Ahmad Luthfi1,2
1Graduate
Student
Graduate School for International Development and Cooperation, Hiroshima University,
1-5-1 Kagamiyama, Higashi-Hiroshima 739-8529, Japan
2Fuel
Subsidy Analyst
Ministry of Energy and Mineral Resources of Republic of Indonesia
Jl. Medan Merdeka Selatan No. 10 Jakarta, Indonesia 10910
e-mail: [email protected]
Shinji KANEKO
Professor
Graduate School for International Development and Cooperation, Hiroshima University,
1-5-1 Kagamiyama, Higashi-Hiroshima 739-8529, Japan
e-mail: [email protected]
Abstract
Due to its accelerating economy and growing population, Indonesia consumed more and more energy.
Unfortunately, oil production continued to deplete and Indonesia became oil-importing-country since 2003.
Furthermore, Indonesia energy policy has change drastically since Asian Crisis by giving more funds to fuel subsidy
than development fund.
Previous study about relation between oil price and Indonesia’s economy by Mehrara and Oskoui (2007)
did not consider the structural change due to the crisis, policy change and depletion on oil production. Thus, this paper
examines the oil price impact to the economy in two periods, before (1984-1997) and after crisis (1998 -2012). This
study employs Vector Autoregression Model (VAR) in order to elucidate the role of fuel subsidy. Furthermore, we
also do counterfactual analysis on those periods.
The VAR Model shows that in the first period, the impact of oil was neutral to the Indonesian
macroeconomic indicators. In the first period, the fuel subsidy could be one of factors which successfully protecting
Indonesia from negative impact of the oil price. However, in the second period, Indonesia’s macroeconomic variables
such as GDP Growth, Inflation, and Interest Rate were influenced by Oil Price. Even the amount of fuel subsidy was
almost 9 folds more; it could not shield the economy from the oil price effect. Furthermore, GDP growth in first
(second) period would have been 22% lower (32% higher) if net-oil import share in first (second) period behaved
similarly with the second (first) period.
The policy implication of this paper would be that the Government of Indonesia should consider removing
its fuel subsidy because of its ineffectiveness and invest its oil revenue to development fund or save it in Petroleum
Fund.
Keywords: oil price, fuel subsidy, VAR, counterfactuals
2
1. Introduction
Oil is not only economic commodity, but also political commodity. Ups and downs its price could not simply be
explained by supply and demand factors. For instance, recently The Economist put its headline as “ Sheikhs vs
Shale”, a drastic drop oil phenomenon in the end of 2014 and early 2015. (The Economist, 2014) The OPEC
(Sheikhs) kept the oil production in regular volume even the world was flooded by oil from shale oil technology. By
keeping the volume, they intended to pull down the oil price thus it will halt shale oil production due to its expensive
production cost. As a cartel organization, the OPEC aims to keep the price stable by decreasing oil supply if the price
relatively low and vice versa. However, the OPEC acted contrary on increasing oil supply due to the shale oil. This
event, along with previous political events such as Arab Oil Embargo 1973, Iranian Revolution 1978-1979, Iran Irak
War 1980 – 1988, Gulf War 1991 (Barsky & Kilian, 2004; Hamilton, 2003) and other economic events make research
about “oil and economy relation” always curiosity.
From the economic literature, there are several mechanisms explaining how the oil price impedes the economy. In
the supply side shock, the role of oil price is as an input factor to the production. The declining in oil supply will
hamper the productivity and turns to lessen real wage growth. If wages sticky downward, the economy will decline
and lead to rise unemployment and generate further reduction in the economy. In demand side, rising in oil price will
shift purchasing power from oil-importing-countries to oil-exporting-countries. This phenomenon will boost the
consumer demand in oil exporting countries and vice versa in oil-importing-countries. However, as a net, the effect is
declining in consumer demand and it leads to increasing in world saving. More saving tends to make interest rate
lower and push investment higher and lead to unchanged GDP. However the impact of fall in consumption will lead to
declining of GDP. (Brown & Yücel, 2002)
Moreover, the oil price effect is not only by increasing and decreasing effect, but also volatility effect, particularly
in oil importing country. The uncertainty in oil price as vital production factor makes investors tend to delay their new
investment until the price more stable. (Bernanke, 1983).Moreover, companies tend to hold recruiting new employees
until the condition reactively constant (Hamilton, 1988). However, the effect of oil price seems lighter after 1980s’
and prudent monetary and fiscal policy are two of the possible factors in neutralize the impact.(Hooker, 1996). Thus
Hamilton argues that the oil price impact now has changed to be non-linier by introducing non-linier oil price.
(Hamilton, 1996)
The oil and macroeconomic relation studies begin after the West countries suffering from oil embargo in 1973
and 1979. (Rasche & Tatom, 1981) (Bruno & Sachs, 1982) , (Darby, 1982),(Hamilton, 1983) are among the earlier
researchers about the inverse relation between oil and the economy. Further, the scholars attempted to validate this
phenomenon in other empirical cases. Thus, they analyze it in notable developed or developing countries
(Jiménez-Rodríguez & Sánchez, 2005);(Du, Yanan, & Wei, 2010), and oil exporting countries or oil importing
countries (Mehrara & Oskoui, 2007);(Iwayemi & Fowowe, 2011); (Cavalcanti & Jalles, 2013). They also examines
in relatively small countries, to reassure if the relation is establish from small to large economy.(Jayaraman &
Choong, 2009);(Jbir & Zouari-Ghorbel, 2009; Lorde, Jackman, & Thomas, 2009). Other studies concentrate on the
impact of oil volatility on the economy which also found important results. (Ferderer, 1996; Lee, Ni, & Ratti, 1994;
Rafiq, Salim, & Bloch, 2009)
3
In those countries, the studies found that, in most of the cases, the impact of increasing oil price is negative
(positive) to the oil importing (exporting) countries economy, no matter the size of the economy. Some of the
countries on those studies also experience an increasing inflation and raising the interest rate. For the oil volatility
effect, the studies found that the impact of oil uncertainty also negatively to the economy, particularly to oil importing
countries. Finally, monetary policy and fiscal policy can neutralize the negative impact of oil on macroeconomic
condition. The further questions arise related to the oil impact on the economy, “What would happen to the country
which used to be oil-exporting countries but turned to be importing ?
2. Indonesia Fuel Subsidy Policy Overview
Indonesia is one of “unlucky” countries which turned to be net importing country after more than a hundred years
net exporting. The peak of oil production in Indonesia occurred in 1977 producing more than 1.638 million barrel per
day which means the net (production – consumption) is about 1.4 million barrel per day. However, industrial
development and population demand more oil but the source of oil is depleting. Thus, since 2003, Indonesia turn to be
net-oil-importing and in 2012 41% of its domestic demand and the amount is increasing (BP - British Petroleum,
2013). In responding the fact, Government of Indonesia decided to withdraw from OPEC in 2009 after almost half
century as one of prominent members.(Pallone, 2009).
Before Asian Crisis (1998), Indonesia is pointed as a role model on managing energy policy for energy exporting
countries. Resource curse, a tendency to develop less in rich-resource-countries; did not happened in Indonesia.
Indonesia had political condition that make the policy elite chose a rational economic policy, for instance, more
openness on trading. Moreover, rich resource in Indonesia tended to bring more foreign direct investment, more
access to technology and also to foreign market. (Basri & Hill, 2004; Rosser, 2007). Indonesia’s prudent monetary
and fiscal policy in that period effectively cushions the impact of oil price shock to macroeconomic variables.
Government of Indonesia (GoI) attempted to keep their currency stable. In the fiscal policy, GoI tended to avoid
budget deficit by cutting energy subsidy and given more fund to infrastructure, education and human
capital.(Mehrara & Oskoui, 2007).
After Asian Crisis, GoI tended to change its fiscal policy. In the graph below we can see the drastic changing of
fuel subsidy policy and capital investment from 1984 -2012 and the natural condition of declining of oil production
and higher domestic demand.
4
50.00%
1,200.00
Oil Net Export, Fuel Subsidy and Capital Investment 1984 - 2012
1,000.00
average 832
40.00%
800.00
36.64%
600.00
30.00%
400.00
20.00%
200.00
13.80%
10.00%
11.03%
0.00
-102
-200.00
1.59%
0.00%
-400.00
1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 -600.00
-10.00%
Asian Crisis
Thoulsand
Barrel
-800.00
Figure 1 Oil Net Export, Fuel Subsidy and Capital Investment 1984 – 2012
(source : Ministry of Finance 2013; BP Statistic 2013)
The figure 1 shown that after crisis, the Government change the fuel subsidy policy by providing more fund
to fuel subsidy. In average, the amount of subsidy between 1984–1997 is only 1.59% from national budget. After
crisis, the amount rockets to 13.08%, almost ninth folds. Moreover, the portion of development expenditure shrink
form 36.64% from government spending to only 11.03% (one third).Not only the policies, but the natural condition
also has changed. The average net-oil production drastically drop from 832 thousand barrel/ day to minus 102
thousand barrel/day. Thus, we should revisit the conclusions from those papers, does it still hold after reform? And
what would be happened if the government change its policy with the same oil price.
There are already several researches about the impact of oil price to Indonesia’s economy. Abeysinghe
(2001) investigates the impact of oil price to several Asian countries including Indonesia. The author examines not
only the immediate impact of oil price but also secondary impact from the trading. The study found that for oil
exporting countries such as Malaysia and Indonesia, the first impact is positive but secondary impact from trading
with partner countries is negative and outnumber previous impact. Thus, the net impact is negative. The second
literature by Mehrara and Oskoui (2007) examines the impact of oil price to four notable oil-exporting-countries,
Saudi Arabiya, Iran, Indonesia and Qatar. The impacts are different among the countries; for the countries which have
invested their income from oil to oil saving institution (Qatar) and imposed prudent monetary and fiscal policy
(Indonesia), the fluctuation in oil price does not have any significant effect on their GDP. The result is contrast to the
other two countries.
However, those studies have not considered structural change from government policy after crisis, thus we
examine the changing and split the time into two period, first period from 1984 Q1 – 1997 Q4 and second period 1998
Q1 2012 Q4.
5
3. Data and Methodology
3.1 Data
This study divide the time into two time frames, before crisis, Q1 1984 – Q4 1997 and after crisis Q1 1998 – Q4
2012. The reason behind choosing the time range is data availability and changing in fuel subsidy in those times, as
mention in the previous part.
To validate the argument, we test whether the subsidy policy variable (in real term) is changing in those time
frames using Bai Perron test (1989) the result as follow:
Figure 2: Bai Perron Test Result of Indonesian Fuel Subsidy Q1 1984 – Q4 2012.
The test reveal that the there is a structural change for fuel subsidy variable, and we can separate the data from Q1
1984 – Q4 1997 and Q1 1998 – Q4 2012.
The explanation of variables is as follow:
Oil Price (LIN-O)
We choose Minas as a proxy to crude oil in Indonesia. Minas is the largest oil field in Indonesia, has been
producing since 1955 and was used as one of component of OPEC basket price. The data source is OPEC Statistics,
a monthly data. To eliminate influence of the exchange rate, we transform the Dollar price to Rupiah by the average
monthly exchange rate published by International Financial Statistics IMF.(International Monetary Fund, 2013;
OPEC, 2015)
6
Realized Oil Volatility (VOL)
For measuring oil price volatility, we employ Realized Volatility following Merton (1980), Andersen et al (2003) and
Kliesen and Guo (2005) as follow:
Quarterly Realized Oil Volatility is RV, the sum squared monthly price changes in a quarter q year t :
𝒎=𝟑
𝑹𝑽𝒒𝒕 = ∑ (𝑹𝑬𝑻𝒎 )𝟐
(1)
𝒎=𝟏
where RETm is the changing of Minas Oil Price in month m of quarter q
Net Oil Price Increase (NOPI) and Net Oil Price Decrease (NOPD)
Some scholars argue that measuring oil price using linier oil price is not accurate anymore. Thus, they suggest non
linier measurements such as Net Oil Price Increase (NOPI) and Net Oil Price Decrease (NOPD). One of the scholars
is Hamilton who argues that current oil prices should be compared with prices over a year and not the previous
quarter. Therefore, the net oil price increase is defined as the percentage increase in the current price of oil over the
price in the previous 4 quarters, if it is positive, and zero otherwise. For measuring NOPI we follow (Hamilton 1996)
and formulated as
NOPIt = MAX [0,log (LIN-Ot) – log (MAX (LIN-Ot-1, LIN-O t-2,…, LIN-Ot-4))]
(2)
NOPDt = MIN [0,log (LIN-Ot) – log (MIN (LIN-O t-1, LIN-O t-2,…, LIN-Ot-4))]
(3)
Growth of Gross Domestic Product (GGDP)
The source of data is from Indonesian Statistical Body, a quarterly data from 1984 Q1 to Q4 2012 Q1. The original
data are deflated to 2005 basis. (Indonesian Statistical Body, 2013)
Inflation Rate (IF)
The data are derived from International Financial Statistics IMF, a quarterly data from Q1 1984 – Q4 2012.
Unemployment (Unemploy)
The data are derived from Indonesian Statistical Body, a yearly data from 1984 –2012, converted to quarterly basis by
adopting (Lisman, J H C and Sandee, 1964)
Fuel Subsidy (Fuel Subsidy)
In other studies, the authors employ some macroeconomic data to proxy the subsidy policy. Rafiq (2009) using
budget deficit as a proxy of fuel subsidy for Thailand case and Jbir & Zouari-Ghorbel (2009) in Tunisian case using
government spending to be the proxy. In Indonesian case, because Ministry of the Government paid the fuel subsidy
7
in the end of the year, thus quarterly data not represents the real situation. For example, the fuel subsidy for Q1, Q2
and Q3 usually zero and suddenly the total amount cumulative in Q4. One way to estimate subsidy policy is by using
PSE method (IEA, 2000). It simply defines subsidy as the gap between the reference energy price and the local energy
price. We follow this approach with some adjustments. Basically our approach distributes the yearly fuel subsidy
amount to quarterly, based on the gap between reference price and the retail subsidized price in Indonesia. We assume
that the volume of the fuel does not changing in a year, which is still plausible. We can denote the formula as follow:
(𝑭𝑹𝑷𝒒𝒕 – 𝑭𝑺𝑷𝒒𝒕 )
𝑺𝑼𝑩𝒒𝒕 = (∑𝟒
𝒒=𝟏(𝑭𝑹𝑷𝒒𝒕
) 𝒙 𝑺𝑼𝑩𝒕
– 𝑭𝑺𝑷𝒒𝒕 )
SUBqt
= Amount of Fuel1 Subsidy in quarter q year t (Local currency)
FRPqt
= Mean of Fuel Reference Price in quarter q year t (Local currency)
FSPqt
= Mean of Fuel Subsidy Price in quarter q year t (Local currency)
SUBt
= Total Fuel Subsidy in year t (Local currency)
(4)
Real Interest Rate (Interest Rate)
The data are derived from International Financial Statistics IMF, a quarterly data and converted to real term by
subtract to inflation rate.
3.2. Methodology
To examine the relationship between oil price and macroeconomic variables, we utilize VAR (Vector
Autregression). This method is pioneered by Sims (1980) and has become very popular among other methods in
analyzing oil and economy relation. In mathematical formula, the equation can be written as:
𝒑
𝒀𝒕 = 𝒄 + ∑ 𝚽𝒊 𝐘𝒕−𝒊 + 𝜺𝒕
(5)
𝒊=𝟏
Where 𝒀𝑡 = (Y1t,Y2t…Ynt)’ is a n x 1 vector of endogenous variables, while 𝐘𝑡−𝑖 is the corresponding lag terms
of order i. Φ𝑖 is the n x n matrix of autoregressive coefficient of vector Yt-i for i=1,2,…p. c = (c1,c2,…cn)’ is the n x
intercept vector of the VAR model. 𝜺𝑡 = (ε1t, ε2t,… εnt)’ is the n x 1 vector of White Noise Process.
In matrix form, we can denote as:
𝐴11 (𝐿)
𝐴10
𝑂𝑖𝑙 𝑃𝑟𝑖𝑐𝑒 𝑡
𝐴12 (𝐿)
𝐴20
𝐺𝐺𝐷𝑃𝑡
𝐼𝐹𝑡
= 𝐴30 + 𝐴13 (𝐿)
𝐼𝑅𝑡
𝐴40
𝐴14 (𝐿)
[𝑈𝑛𝑒𝑚𝑝𝑙𝑜𝑦𝑡] [𝐴50 ] [𝐴15 (𝐿)
𝐴21 (𝐿)
𝐴22 (𝐿)
𝐴23 (𝐿)
𝐴24 (𝐿)
𝐴25 (𝐿)
𝐴31 (𝐿)
𝐴32 (𝐿)
𝐴33 (𝐿)
𝐴34 (𝐿)
𝐴35 (𝐿)
𝐴41 (𝐿)
𝐴42 (𝐿)
𝐴43 (𝐿)
𝐴44 (𝐿)
𝐴45 (𝐿)
𝐴51 (𝐿) 𝑂𝑖𝑙 𝑃𝑟𝑖𝑐𝑒 𝑡
𝑒1𝑡
𝑒
𝐴52 (𝐿)
𝐺𝐺𝐷𝑃𝑡
2𝑡
𝐴53 (𝐿)
𝐼𝐹𝑡
+ 𝑒1𝑡
𝑒4𝑡
𝐼𝑅𝑡
𝐴54 (𝐿)
[
[
𝑈𝑛𝑒𝑚𝑝𝑙𝑜𝑦
]
𝐴55 (𝐿)]
𝑡 𝑒5𝑡 ]
1 Fuel Price is calculated as weighted average of 3 main fuel commodities, Gasoline Octane 88, Diesel Cetane 48 and Kerosene
(6)
8
4. Empirical Result
In this section we analyze the empirical result of the VAR model describe in previous section, including
Granger causality test, impulse response function and variance decomposition.
Unit Root Test
Since the data are times series data, it is essential to make sure that the data are stationer. The Augmented
Dickey Fuller (ADF) and Philips Perron were applied to test the data stationary. Table 1 below resumes the
result of both two tests. These tests were based on the three following models: (i) without intercept, (ii) with an
intercept and (iii) with an intercept and trend.
From the test we can conclude that all variables were stationer except for Oil Price, Subsidy and
Unemployment. Those variables were integrated in order one (I(1)) and stationer in first difference as shown in
Table 2. We differentiate and take natural log for Oil Price and Unemployment and for Fuel Subsidy, we only
differentiate due to some negative values.
Table 1: Unit Root Test
I
Level
ii
Iii
I
ADF
OIL
VOLATILITY
GGDP
IR
IF
UNP
SUB
NOPI
NODI
-0.94
-7.78***
-3.57***
-2.83*
-5.74***
-1.41
-2.11
-8.31***
-8.11***
-3.09
-7.77***
-3.56**
-2.99
-5.73***
-0.62
-4.44***
-8.30***
-8.15***
0.21
-6.14***
-2.22**
-1.90*
-4.21***
0.10
-1.21
-7.25***
-7.47***
-9.69***
-18.02***
-24.63***
-11.10***
-8.16***
-6.68***
-9.11***
-11.62***
-9.38***
-9.70***
-17.95***
-24.51***
-11.05***
-8.12***
-6.82***
-9.10***
-11.58***
-9.34***
-9.69***
-18.10***
-24.74***
-11.15***
-8.19***
-6.65***
-9.11***
-11.68***
-9.42***
PP
OIL
VOLATILITY
GGDP
IR
IF
UNP
SUB
NOPI
NODI
-0.83
-7.86***
-14.54***
-3.07**
-11.15***
-1.42
-2.26
-8.05***
-7.80***
-3.00
-7.85***
-14.46***
-3.25*
-5.45***
-0.78
-3.86**
-8.04***
-7.87***
0.60
-6.54***
-12.49***
-1.92
-4.21***
0.00
-1.10
-7.23***
-7.51***
-9.74***
-44.66***
-59.93***
-11.11***
-19.51***
-3.06**
-10.25***
-60.00***
-65.98***
-10.72***
-45.87***
-59.71***
-11.06***
-19.44***
-3.07
-10.23***
-66.15***
-68.16***
-9.35***
-44.93***
-60.32***
-11.15***
-19.64***
-3.09***
-10.15***
-60.41***
-66.51***
Variables
*,** and *** denote significant at 10%,5%, and 1 % respectively
First Different
Ii
Iii
9
Granger-Causality Test
To investigate if those variables have causal relationships to other variables in the VAR model we employ
Granger Causality test. Variable x is said to Granger-cause variable y if the inclusion of past values of variable x
helps in better prediction of variable y.
In 1984 Q1 – 2012 Q2, Indonesia experienced The Asian Financial Crisis. It begins in summer 1997, hit
Thailand by devaluating its currency. Furthermore it rapidly spread to other Asian and Latin countries in fall
1997 and arrive in Russia and Brazil around summer 1998. In the beginning of the crisis, Indonesia is predicted
as the least affected country due to its quick and concerted action. Unfortunately, because of many problem and
weaknesses before the crisis, it becomes worsen. The economic drastically drop from growing 5 % in 1996,
decline to -4% in 1997 and the worst in 1998 by growing – 13.37%.(Indonesian Statistical Body, 2013)
Table 2: Granger Causality Test (1984 – 1997)
Null Hypothesis Ho
OIL does not Granger Cause
GGDP
SUB
IF
IR
UNP
OIL
VOL
NOPI
NOPD
χ2-Statistic (p-value)
χ2-Statistic (p-value)
χ2-Statistic (p-value)
χ2-Statistic (p-value)
0.28 (0.60)
9.03*** (0.01)
0.02 (0.88)
0.16 (0.69)
1.03 (0.31)
0.10 (0.75)
0.07 (0.79)
0.14 (0.71)
0.03 (0.87)
0.08 (0.78)
1.21 (0.27)
0.45 (0.50)
0.07 (0.80)
0.24 (0.63)
0.75 (0.39)
0.79 (0.38)
1.14 (0.29)
0.84 (0.36)
1.34 (0.25)
0.32 (0.57)
Table 3: Granger Causality Test (1998 – 2012)
Null Hypothesis Ho
OIL does not Granger Cause
GGDP
SUB
IF
IR
UNP
OIL
VOL
NOPI
NOPD
χ2-Statistic (p-value)
χ2-Statistic (p-value)
χ2-Statistic (p-value)
χ2-Statistic (p-value)
7.38*** (0.01)
1.54 (0.21)
0.52 (0.47)
1.29 (0.26)
0.51 (0.48)
10.66*** (0.01)
0.48 (0.49)
3.55* (0.06)
3.90** (0.05)
0.02 (0.87)
14.20*** (0.01)
1.19 (0.28)
13.49*** (0.01)
2.75* (0. 10)
2.38 (0.12)
0.91 (0.34)
0.38 (0.54)
0.11 (0.74)
0.07 (0.79)
0.67 (0.41)
The result of Granger Causality Tests are presented in the table 2 and 3. It can be seen that in period Q1
1984 – Q4 1997, Linier Oil Price Shock only Granger-Cause Fuel Subsidy. Moreover, the result shows that there
is no significant impact from Oil Volatility, Net Oil Price Increase and Net Oil Price Decrease to
Macroeconomic indicators in period 1984 Q1 – 1997 Q4. This result confirms some studies (Rossen 2007)
(Mehrara 2007) (Basri 2005) that conclude Indonesian energy policy in this period was very effective. As a
result, Indonesia is one of exporting countries that do not suffer from resource curse (Rosser, 2007), secure from
oil shock and successfully divert the income source to others than oil. (Basri,2005) and relatively does not
exposed by negative effect of changing in oil price.(Mehrara, 2007)
Furthermore this test reveals that fuel
10
subsidy policy in this period was efficient, even only 1-2% from national budget, but effective in limiting the
negative impact of oil shock.
The result in the second period is different. In the second period, Q1 1998 – Q4 2012, all Oil Price
Change variables are Granger- cause Macroeconomic indicators except Net Oil Price Decrease. The Impact of
Linier Oil Price Shock is significant for GDP Growth, and for Oil Price Volatility and Net Oil Price Increase, the
impact are significant to GDP Growth, Inflation Rate and Real Interest Rate.
Impulse Response Function
Figures 3 – 8 contain the impulse response functions for the responses of the macroeconomic variables to
different oil price change variables. Each figure traces the effect of a one-time shock to the measures of oil shocks on
the current and future values of each of the macroeconomic variables. In period 1984 Q1 – 1997 Q4, only Linier Oil
Price imposes significant effect, thus we only focus on that variable. For period 1998 Q1 – 2012 Q4, we describe
Impulse Response Function of all variables. Figure 3 contains impulse response of macroeconomics variables to
Linier Oil Price Shock between 1984Q1 – 1997Q4, and Figures 4 - 7 depict impulse response to both linier and non
linier oil price shocks between 1998Q1 – 2012Q4.
Response to Cholesky One S.D. Innovations ± 2 S.E.
Response of D(FUEL_SUBSIDY) to D(LOG(OIL_PRICE_SHOCK) )
Response of GDP_GROWTH to D(LOG(OIL_PRICE_SHOCK) )
20
.008
.004
15
.000
10
-.004
5
-.008
-.012
0
-.016
-5
-.020
1
2
3
4
5
6
7
8
9
10
1
Response of INFLATION to D(LOG(OIL_PRICE_SHOCK) )
2
3
4
5
6
7
8
9
10
Response of INTEREST_RATE to D( LOG( OIL_PRICE_SHOCK))
.008
.03
.006
.02
.004
.01
.002
.00
.000
-.01
-.002
-.004
-.02
1
2
3
4
5
6
7
8
9
10
1
2
3
4
Response of D(LOG(UNEMPLOYMENT)) to D(LOG(OIL_PRICE_SHOCK) )
.03
.02
.01
.00
-.01
-.02
1
2
3
4
5
6
7
8
9
10
Figure 3: Impulse Response Function of Linier Oil Price Shock 1984 – 1997
5
6
7
8
9
10
11
Figure 3 comprises the response of the macroeconomic variables to a shock in the linier oil price shocks between
1984 Q1 -1997 Q4. The response of GDP Growth to shocks in the linier oil price shock is negative until the 3 periods,
thus indicating that linier oil price shocks have a negative effect on output, even though the amount is relatively small.
For fuel subsidy and inflation, a linier oil price shock first results in a positive response, which lasts for different
periods, ranging from the second for inflation and to third period for fuel subsidy. For interest rate and
unemployment, the response are almost zero.
Response to Cholesky One S.D. Innovations ± 2 S.E.
Response of D(FUEL_SUBSIDY) to D(LOG(OIL_PRICE_SHOCK) )
Response of GDP_GROWTH to D(LOG(OIL_PRICE_SHOCK) )
80
.01
60
40
.00
20
-.01
0
-20
-.02
1
2
3
4
5
6
7
8
9
10
1
Response of INFLATION to D(LOG(OIL_PRICE_SHOCK) )
2
3
4
5
6
7
8
9
10
Response of INTEREST_RATE to D( LOG( OIL_PRICE_SHOCK))
.020
.04
.015
.03
.010
.02
.005
.01
.000
.00
-.005
-.010
-.01
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
Response of D(LOG(UNEMPLOYMENT)) to D(LOG(OIL_PRICE_SHOCK) )
.010
.005
.000
-.005
-.010
1
2
3
4
5
6
7
8
9
10
Figure 4 : Impulse Response Function of Linier Oil Price Shock 1998 - 2012
Figure 4 contains impulse response of macroeconomics variables to linier oil price shock between 1998 Q1
– 2012 Q4. The Fuel Subsidy response to the linier oil price shock is positive until third quarter. For GDP growth, the
response is positive in the first quarter, turn to be negative in second quarter and diminishing in fifth or sixth period.
The pattern of fuel subsidy and GDP growth is similar if we compare between period 1984 Q1 – 1997 Q4 and 1998
Q1– 2012 Q4 but the magnitude is higher in the last period (see fig. 3). For inflation rate and interest rate, the response
to the linier oil price shock is positive until 10th quarter. For the unemployment rate, the response is almost zero.
This result deduces that the government responds the shock in oil price by giving more fuel subsidy to
hinder the negative effects in both periods. The different are, the amount of subsidy is greater in second period and
higher impact hits in second period instead of the first period. One of the possible explanations is that the government
in first period is better in managing their oil revenue by investing to the economy. In the second period, Indonesia turn
10
12
to be net oil exporting country which has less profit from selling the oil and have to finance huge amount of fuel
subsidy.(Pallone, 2009). The amount of fuel subsidy in first period is 8 times higher than second period, drastically
increase from 1.59% to 13.09% from national budget.
Response to Cholesky One S.D. Innovations ± 2 S.E.
Response of D( FUEL_SUBSIDY) to OIL_VOLATILITY
Response of GDP_GROWTH to OIL_VOLATILITY
30
.005
20
.000
10
0
-.005
-10
-.010
-20
-.015
-30
-40
-.020
1
2
3
4
5
6
7
8
9
10
1
Response of INFLATION to OIL_VOLATILITY
.05
.020
.04
.015
.03
.010
.02
.005
.01
.000
.00
-.005
-.01
2
3
4
5
6
7
8
3
4
5
6
7
8
9
9
10
1
2
3
4
5
6
7
8
9
Response of D( LOG( UNEMPLOYMENT) ) to OIL_VOLATILITY
.012
.008
.004
.000
-.004
-.008
1
2
3
4
5
6
7
8
9
10
Figure 5: Impulse Response Function of Oil Price Volatility 1998 - 2012
Figure 5 depicts the response of macroeconomics variables to oil price volatility. The response of fuel
subsidy and GDP Growth are negative until second quarter and 10th quarter. For inflation rate and interest rate, both
responses are positive until 10th quarter. Lastly, the unemployment response is positive but very small. This result
confirms the result from Granger Causality test that uncertainty of oil price will reduce GDP Growth, rise inflation
and interest rate.
10
Response of INTEREST_RATE to OIL_VOLATILITY
.025
1
2
10
13
Response to Cholesky One S.D. Innovations ± 2 S.E.
Response of D( FUEL_SUBSIDY) to NET_OIL_PRICE_INCREASE
Response of GDP_GROWTH to NET_OIL_PRICE_INCREASE
50
.010
40
.005
30
.000
20
-.005
10
-.010
0
-.015
-10
-20
-.020
1
2
3
4
5
6
7
8
9
10
1
Response of INFLATION to NET_OIL_PRICE_INCREASE
2
3
4
5
6
7
8
9
10
Response of INTEREST_RATE to NET_OIL_PRICE_INCREASE
.025
.05
.020
.04
.015
.03
.010
.02
.005
.01
.000
.00
-.005
-.01
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
Response of D( LOG(UNEMPLOYMENT) ) to NET_OIL_PRICE_INCREASE
.015
.010
.005
.000
-.005
-.010
1
2
3
4
5
6
7
8
9
10
Figure 6 : Impulse Response Function of Net Oil Price Increase (NOPI) 1998 - 2012
The Figure 6 shows the response of macroeconomic variables to Net Oil Price Increase. The response of
Fuel Subsidy is increasing until third quarter, and dies out. For GDP growth, the response are negative, peak in the
second quarter and still negative until 10th quarter. For inflation, interest rate and unemployment, the responses are
positive and still positive after 10th quarter. The peak for inflation, interest rate and unemployment are second, fourth
and third quarter, respectively.
10
14
Response to Cholesky One S.D. Innovations ± 2 S.E.
Response of D( FUEL_SUBSIDY) to NET_OIL_PRICE_DECREASE
60
Response of GDP_GROWTH to NET_OIL_PRICE_DECREASE
.015
.010
40
.005
20
.000
-.005
0
-.010
-20
-.015
1
2
3
4
5
6
7
8
9
10
1
Response of INFLATION to NET_OIL_PRICE_DECREASE
2
3
4
5
6
7
8
9
10
Response of INTEREST_RATE to NET_OIL_PRICE_DECREASE
.016
.03
.012
.02
.008
.01
.004
.00
.000
-.004
-.01
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
Response of D( LOG( UNEMPLOYMENT) ) to NET_OIL_PRICE_DECREASE
.010
.005
.000
-.005
-.010
1
2
3
4
5
6
7
8
9
10
Figure 7 :Impulse Response Function of Net Oil Price Increase (NOPI) 1998 - 2012
The last figure is the response of macroeconomic variables to Net Oil Price Decrease. It is surprise that the
responses of macroeconomic variables are similar with the Net Oil Price Increase. However, due to insignificant
result in Granger Causality test, this figure only explains that the responses are very small.
Variance Decomposition
Another instrument to reveal the VAR result is by employing variance decomposition analysis.
Variance decomposition gives the proportion of the movement in the dependent variables that are due to their
own shocks versus shock to the other variables.
10
15
Table 4: Variance Decomposition to Oil Price 1984 – 1997
Dependent Variable
Period
LIN-O
VOL
NOPI
NOPD
1
5
10
1
5
10
1
5
10
1
5
10
1
5
10
50.13
44.01
43.34
0.72
3.06
3.05
10.60
10.35
10.33
0.00
0.12
0.13
0.00
0.74
0.77
0.02
0.50
0.51
6.01
6.01
7.12
7.11
3.60
2.97
3.60
3.60
1.61
0.86
1.56
1.61
0.02
2.00
1.23
1.14
29.00
26.61
26.36
1.51
1.28
1.29
5.07
5.09
5.07
0.40
3.01
3.30
0.14
2.34
2.70
9.85
10.46
10.35
1.24
2.26
2.26
1.60
2.70
2.70
0.02
1.17
1.28
2.54
1.30
1.21
Subsidy
GGDP
Inflation
Interest
UNP
Table 5:Variance Decomposition to Oil Price 1998 – 2012
Dependent Variable
Subsidy
GGDP
Inflation
Interest
UNP
Period
LIN-O
VOL
NOPI
NOPD
1
5
10
1
5
10
1
5
10
1
5
10
1
5
10
61.11
58.08
58.03
2.23
12.95
13.18
0.56
14.49
15.06
4.74
17.63
20.34
0.22
0.50
1.15
7.95
7.81
7.81
12.07
17.15
17.67
12.95
23.57
24.70
3.87
33.23
36.00
0.61
3.23
4.57
7.95
16.53
17.09
17.11
0.19
19.28
19.76
3.90
39.33
40.36
3.63
44.54
51.88
1.10
4.21
5.52
31.42
29.96
29.94
0.83
6.33
6.47
2.07
6.87
7.14
2.71
8.70
9.64
0.28
0.24
0.48
The fourth column of Table 5 contains the variance decompositions for oil price volatility and it is seen that
oil price volatility contributed only 0.02% to variation in output in the first period and slightly increase to 0.50%
in the 5th period and end up to 0.51 % in the 10th period. The oil price volatility account only less than 1% for fuel
subsidy, about 7 % for GDP growth but account only very small portion for others variables.
Furthermore, in the most cases for the first period (1984 – 1997), only small portion of macroeconomic
indicators can be explained by changing in the linier oil price, NOPI and NOPD except for fuel subsidy. Linier
oil price shock , NOPI and NOPD account for 40% ,26% and 10% variation in this variable, respectively.
The situation is totally different for the second period. In the most cases, almost all variation in the
macroeconomic indicators can be explained by changing in the oil price variables. For instance, linier oil price
16
shock and NOPD account for about 58% and 30 % variation in the fuel subsidy, respectively. Furthermore, other
examples are oil volatility and NOPI account for about 36% and 52% of variation in interest rate respectively. In
short, except for unemployment, all the oil price variables can be explained variation in the macroeconomic
variables. Again, we find evidence that effect of Oil Price in the second period had a more pronounced effect on
macroeconomic variables than the first period.
Counterfactual Analysis
In this part, we would like to investigate, what would happen to the real output growth if net import share in the
first period behaved similarly to the second period and vice versa. As we see, the Indonesian oil import is
growing and in the second period, Indonesia becomes net-oil-importing. To bring the net import share to the
equation, we follow (Hamilton, 2003) by defining a new measurement of NOPI.
̂ = 𝑵𝒆𝒕 𝑶𝒊𝒍 𝑰𝒎𝒑𝒐𝒓𝒕 𝑺𝒉𝒂𝒓𝒆2 𝒙 𝑶𝒊𝒍 𝑷𝒓𝒊𝒄𝒆
𝑵𝑶𝑷𝑰
𝒏
̂ 𝑷𝒆𝒓𝒊𝒐𝒅
Thus we can define 𝑵𝑶𝑷𝑰
= 𝑵𝒆𝒕 𝑶𝒊𝒍 𝑰𝒎𝒑𝒐𝒓𝒕 𝑺𝒉𝒂𝒓𝒆𝑷𝒆𝒓𝒊𝒐𝒅 𝒏 𝒙 𝑶𝒊𝒍 𝑷𝒓𝒊𝒄𝒆𝑷𝒆𝒓𝒊𝒐𝒅 𝒏
𝒕
(7)
(8)
Furthermore, we run regression following (Cavalcanti & Jalles, 2013) to investigate the relation between
Indonesian economic growth and oil price in both period.
𝟒
𝐆𝐫𝐨𝐰𝐭𝐡𝒕 = 𝜶 + ∑
𝒋=𝟏
𝟒
𝜸𝒋 𝑮𝒓𝒐𝒘𝒕𝒉 𝒕−𝒋 + ∑
𝒋=𝟏
̂ 𝒕−𝒋
𝜷𝒋 𝑵𝑶𝑷𝑰
(9)
(7)
For Indonesian case, we believe that the appropriate counterfactual analysis is to assume that the import share
and oil price is interchangeable on those periods. Thus, we model the counterfactual as
1. For basis, we calculate the Growtht on both periods
2. For first counterfactual, we calculate the Growth t in period 1 by using oil price in period 2 and vice
versa
3. For the second counterfactual, we calculate the Growth t in period 1 by using oil import share in period
2 and vice versa.
4. We compare all the results to the basis.
17
Table 6 : Counterfactual exercises : the Role of Oil Dependence
Data Predicted
Experiment 1
̂
𝑃𝑒𝑟𝑖𝑜𝑑2
𝛽𝑗𝑃𝑒𝑟𝑖𝑜𝑑1 𝑁𝑂𝑃𝐼
𝑖−𝑗
Experiment 2
̂
𝑃𝑒𝑟𝑖𝑜𝑑1
𝛽𝑗𝑃𝑒𝑟𝑖𝑜𝑑2 𝑁𝑂𝑃𝐼
𝑖−𝑗
Experiment 3
̂
𝑃𝑒𝑟𝑖𝑜𝑑1
𝛽𝑗𝑃𝑒𝑟𝑖𝑜𝑑2 𝑁𝑂𝑃𝐼
𝑖−𝑗
Experiment 4
̂
𝑃𝑒𝑟𝑖𝑜𝑑2
𝛽𝑗𝑃𝑒𝑟𝑖𝑜𝑑1 𝑁𝑂𝑃𝐼
𝑖−𝑗
Y end of period
1.6842
1.2296
Y exp/Y predicted
NOPIt
standard deviation
0.0318
0.0227
Std exp/Std predicted
1.0831
0.9518
0.7725
0.9344
0.7453
0.6311
1.3155
1.1133
5. Conclusion and Policy Implication
From Ganger causality, we conclude that for the first period, only fuel subsidy is Granger caused by linier oil
price shock. However, in the second period we can infer that GDP growth is Granger caused by linier oil price shock,
oil price volatility and NOPI. Moreover, interest rate and inflation rate are Granger caused by oil price volatility and
NOPI.
Based on impulse response function, we can infer that, in the most cases in both periods, impact of oil price
variables are positive to fuel subsidy, negative to GDP growth, inflation and interest rate. However, the oil price
variables only can explain very small variation of unemployment rate. Moreover, the oil price variables in the second
period have stronger and longer effect to the macroeconomic variables. Furthermore, the result of variance
decomposition confirms Granger causality test and impulse response function that the impact of different type of oil
price variables are more powerful in the second period.
The policy implication of this result would be that the government should reconsider its fuel subsidy policy
because even providing with quite huge amount of money, the negative effects of oil price variables still exist.
18
References
Barsky, R. B., & Kilian, L. (2004). Oil and the Macroeconomy Since the 1970s. Journal of Economic
Perspectives, 18(4), 115–134. doi:10.1257/0895330042632708
Basri, M. C., & Hill, H. (2004). Ideas, Interest and Oil Prices- The Political Economy of Trade Reform
During Soeharto’s Indonesia. The World Economy, Volume 27, Issue 5, pages 633–655.
Bernanke, B. S. (1983). Irreversibility, Uncertainty, and Cyclical Investment. Quarterly Journal of
Economics, (February).
BP - British Petroleum. (2013). Statistical review of world energy 2013. London. Retrieved from
http://www.bp.com/content/dam/bp/excel/Statistical-Review/statistical_review_of_world_energ
y_2013_workbook.xlsx
Brown, S. P. a, & Yücel, M. K. (2002). Energy prices and aggregate economic activity: An
interpretative survey. Quarterly Review of Economics and Finance, 42, 193–208.
doi:10.1016/S1062-9769(02)00138-2
Bruno, M., & Sachs, J. (1982). Input Price Shocks and the Slowdown in Economic Growth: The Case
of U.K. Manufacturing. Review of Economic Studies, 49, 679–705. Retrieved from
http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=4622694&site=ehost-live&sc
ope=site
Cavalcanti, T., & Jalles, J. T. (2013). Macroeconomic effects of oil price shocks in Brazil and in the
United States. Applied Energy, 104, 475–486. doi:10.1016/j.apenergy.2012.10.039
Darby, M. R. (1982). The Price of Oil and World Inflation and Recession. World, 72(4), 738–751.
Du, L., Yanan, H., & Wei, C. (2010). The relationship between oil price shocks and China’s
macro-economy: An empirical analysis. Energy Policy, 38(8), 4142–4151.
doi:10.1016/j.enpol.2010.03.042
Ferderer, J. P. (1996). Oil price volatility and the macroeconomy. Journal of Macroeconomics, 18(1),
1–26. doi:10.1016/S0164-0704(96)80001-2
Hamilton, J. D. (1983). Oil and the Macroeconomy since World War II. Journal of Political Economy,
91(2), 228. doi:10.1086/261140
Hamilton, J. D. (1988). A Neoclassical Model of Unemployment and the Business Cycle. Journal of
Political Economy, 96(3), 593. doi:10.1086/261553
Hamilton, J. D. (1996). This is what happened to the oil price-macroeconomy relationship. Journal of
Monetary Economics, 38, 221–222. doi:10.1016/S0304-3932(96)01283-4
Hamilton, J. D. (2003). What is an oil shock? Journal of Econometrics, 113, 363–398.
doi:10.1016/S0304-4076(02)00207-5
19
Hooker, M. a. (1996). This is what happened to the oil price-macroeconomy relationship: Reply.
Journal of Monetary Economics, 38, 221–222. doi:10.1016/S0304-3932(96)01283-4
IEA. (2000). World Energy Outlook 1999: Looking at Energy Subsidies: Getting the Prices Right.
Paris: International Energy Agency. doi:http://dx.doi.org/10.1787/weo-1999-en
Indonesian Statistical Body. (2013). comparison 2 era subsidy. Jakarta.
International Monetary Fund. (2013). International Financial Statistics.
Iwayemi, A., & Fowowe, B. (2011). Impact of oil price shocks on selected macroeconomic variables
in Nigeria. Energy Policy, 39(2), 603–612. doi:10.1016/j.enpol.2010.10.033
Jayaraman, T. K., & Choong, C. K. (2009). Growth and oil price: A study of causal relationships in
small Pacific Island countries. Energy Policy, 37, 2182–2189. doi:10.1016/j.enpol.2009.01.025
Jbir, R., & Zouari-Ghorbel, S. (2009). Recent oil price shock and Tunisian economy. Energy Policy,
37, 1041–1051. doi:10.1016/j.enpol.2008.10.044
Jiménez-Rodríguez, R., & Sánchez, M. (2005). Oil price shocks and real GDP growth: empirical
evidence for some OECD countries. Applied Economics, 37(1989), 201–228.
doi:10.1080/0003684042000281561
Lee, K., Ni, S., & Ratti, R. A. (1994). Kiseok Lee, Shawn Ni, and Ronald A. Ratti ∗, (May).
Lisman, J H C and Sandee, J. (1964). Derivation of Quarterly Figures from Annual Data. Journal of
the Royal Statistical Society, 13(2), 87–90.
Lorde, T., Jackman, M., & Thomas, C. (2009). The macroeconomic effects of oil price fluctuations on
a small open oil-producing country: The case of Trinidad and Tobago. Energy Policy, 37, 2708–
2716. doi:10.1016/j.enpol.2009.03.004
Mehrara, M., & Oskoui, K. N. (2007). The sources of macroeconomic fluctuations in oil exporting
countries: A comparative study. Economic Modelling, 24, 365–379.
doi:10.1016/j.econmod.2006.08.005
OPEC. (2015). OPEC and Other regional price.
Pallone, S. (2009). Indonesia ’ s Oil Crisis : How Indonesia Became a Net Oil Importer. The Journal of
International Policy Solution, 10.
Rafiq, S., Salim, R., & Bloch, H. (2009). Impact of crude oil price volatility on economic activities: An
empirical investigation in the Thai economy. Resources Policy, 34, 121–132.
doi:10.1016/j.resourpol.2008.09.001
Rasche, R. H., & Tatom, J. a. (1981). Energy price shocks, aggregate supply and monetary policy: The
theory and the international evidence. Carnegie-Rochester Conference Series on Public Policy,
14, 9–93. doi:10.1016/0167-2231(81)90002-6
20
Rosser, A. (2007). Escaping the resource curse: The case of Indonesia. Journal of Contemporary Asia,
37(April 2015), 38–58. doi:10.1080/00472330601104557
The Economist. (2014). Sheikhs v shale. London.