The Journal of World Economic Review; Vol. 6 • No. 2 • (July-December 2011) pp. 123-142

Oil Price Shocks and Macroeconomic Activities in Malaysia Mukhriz Izraf Azman Aziz* & Nor’Aznin Abu Bakar** *

**

Senior lecturer, College of Business. Universiti Utara Malaysia. 06010 Sintok, Kedah, Malaysia. E-mail: [email protected]. Associate Professor, College of Business, Universiti Utara Malaysia. 06010 Sintok, Kedah, Malaysia. E-mail: [email protected]

ABSTRACT

This paper investigates the asymmetric effects of oil price shocks on real economic activities in Malaysia from 1991 to 2007. Using an unrestricted Vector Auto Regressive (VAR) method, mixed results are obtained. Evidence of a symmetric relationship between oil prices and economic activities is obtained from the impulse response function (IRFs). However, the variance decomposition analyses VAR suggest that oil prices have different impacts on economic activities when they increase than when they fall. JEL Codes: F11, N50, C32. Keywords: Oil price shocks, International trade, Unit root test, GDP growth, Export.

1. INTRODUCTION Among the non-OPEC countries, Malaysia is a net oil exporter. While manufacturing products currently account for 75% of total Malaysian exports (and 30% of nominal GDP), 1 crude oil exports contributed 43 billion Malaysian ringgits (RM) of export revenue in 2008, or 6% of Malaysian total exports, and on average 40% of government revenues in annual budgets (Central Bank of Malaysia, 2008). The rise in crude oil prices in recent years has contributed significantly to the increase in revenue for the Malaysian government. The oil revenue to the government paid by PETRONAS2 increased from RM13.6 billion in 2000 to RM51.2 billion in 2007 (Bank Negara Malaysia Report, 2008). This higher oil-based revenue has allowed the government to undertake development spending on infrastructure, education, and healthcare, thus enhancing the country’s long-term productive capacity. As a net exporter of oil, Malaysia benefits from higher crude oil prices in the short term because of better terms of trade. However, high oil prices are also a double-edged sword for Malaysia. Rising oil prices will impact on world growth, which will affect world consumption and income. Thus, oil price shocks may impede the growth of trade between Malaysia and its trading partners, especially for oil importers such as the US, China, Japan, and Europe. Being an open economy, Malaysia is susceptible to adverse cyclic effects of this shock. Economic slowdown in these countries will limit internal consumer demand and thus affect Malaysia’s exports of goods and services. The recent global economic turmoil had resulted in lower GDP growth and rising inflation for Malaysia. In 2008, GDP growth was recorded at 4.6%, down from 6.7% in 2007, while the inflation rate surged to 8.5%, its highest level since 1982, when inflation was 7.2% (Department of Statistics, Malaysia). Studies on Malaysia have shown that high oil prices are not always beneficial to the economy. Abeysinghe (2001) concluded that, although the direct impact of high oil prices on Malaysia is

124

THE JOURNAL OF WORLD ECONOMIC REVIEW

positive, it cannot escape the contractionary effect through its trading partners on growth, and in the long run Malaysia would also lose out. Cunado & de Gracia (2005) found that oil price shocks have a significant effect on inflation, although the oil prices–macroeconomy relationship seems to be less significant. Jaafar et al., (2008) examined the impacts of oil price shocks on the Malaysian economy using computable general equilibrium analysis. Results from the simulation showed that a 5% increase in oil price reduces Malaysian real GDP (RGDP) by 0.35% and fixed capital investment by 8%. According to Abeysinghe (2001), the magnitude to which the volatility of oil prices affects open economies depends on whether the economy is a net importer or exporter of oil. Early literature on oil price shocks on the oil-importing US economy found a statistically significant negative linear relationship with output. These studies include Rasche and Tatom (1981), Darby (1982), Hamilton (1983), Burbidge and Harrison (1984), and Gisser and Goodwin (1986). By the mid-1980s, however, the estimated linear relationship between oil prices and real activity began to lose significance. In fact, the declines in oil prices that occurred over the second half of the 1980s had smaller positive effects on economic activity than predicted by linear models. Thus, Mork (1989), Lee et al., (1995), and Hamilton (1996) introduced non-linear transformations of oil prices and established an asymmetric (non-linear) relationship between increases in oil prices and output growth. More recently, Hamilton (2003) and Jiménez-Rodríguez (2004) also found evidence of a non-linear relationship between the two variables for the US and several OECD economies respectively. Despite the non-linear (asymmetric) relationship found between oil price shocks and macroeconomic variables in oil-importing economies, studies on oil-exporting countries have shown that this relationship is in fact symmetric (linear). Eltony and Al-Awadi (2001) found evidence that symmetric oil price shocks are important in explaining fluctuations in macroeconomic variables in Kuwait. Their results showed the importance of oil price shocks on government expenditures, which are the major determinant for the level of economic activity in Kuwait. Berument et al., (2010) studied the effects of symmetric oil price shocks on output (proxied by industrial production) for a group of Middle East and North African countries. Their impulse response analyses suggest that the effects of world oil price on GDP of Middle Eastern countries are positive and statistically significant. In a recent study by Jbir and ZouariGhorbel (2009), no evidence was found of asymmetric effects of oil price shocks on the Tunisian economy, as no difference in terms of results was discovered between the linear and non-linear vector autoregressive (VAR) models. This paper therefore aims to determine whether the macroeconomic volatility in Malaysia is due to fluctuations in oil prices. Specifically, the paper attempts to establish whether the impact of oil price shocks on Malaysia’s macroeconomic variables are symmetric or asymmetric. This is achieved using a six-variable VAR model with quarterly data from 1991:1 to 2007:4. The paper will employ simulation techniques (impulse response functions) to determine what impacts an oil price shock would have on the variables in the model, how long such effects would last, and when the maximum repercussions could be expected. In addition, an out-ofsample forecast is carried out to measure the accuracy of the VAR model for six quarters from 2008:1 to 2009:2. To take into account the asymmetric effect of oil prices, the paper estimates the relationship between oil prices and macroeconomic variables using the standard linear and

OIL PRICE SHOCKS AND MACROECONOMIC ACTIVITIES IN MALAYSIA

125

three non-linear approaches. These are (1) quarterly oil price changes, (2) separate oil price variables for price increases and decreases, as in Mork (1989), (3) scaled oil price increases and decreases, as in Lee at al., (1995), and (4) net specification as in Hamilton (1996), which considers the amount by which oil prices have risen over the previous year. The present study extends the existing empirical literature on the Malaysian economy in two directions. First, it complements the methodologies employed in Abeysinghe (2001) and Cunado & de Gracia (2005) on the use of VAR with different oil price shock specifications (both linear and non-linear). Although Abeysinghe employed a VAR model, he did not examine the asymmetric effect of oil price shocks on the Malaysian economy. In contrast, Cunado & de Gracia (2005) tested for asymmetric effects of oil prices, but the estimation was carried out using an autoregressive distributed lag (ADL) model. Therefore, the present study combines elements of both papers to obtain a new set of results for the Malaysian economy. Second, unlike Abeysinghe (2001) and Cunado & de Gracia (2005), who specifically examined the effects of oil price shocks on output growth and inflation, the present paper extends the analysis by investigating five macroeconomic variables, namely GDP growth, exchange rate, government expenditure, export, and inflation. This should enable a better understanding of how oil price shocks affect the Malaysian economy. The remainder of this paper is organised as follows. Section 2 provides some background to Malaysia as an oil-exporting country, while section 3 discusses the theoretical framework. Section 4 considers the methodology, followed by the results in section 5. Section 6 concludes the paper. 2. BACKGROUND: MALAYSIA AS AN OIL EXPORTER This section presents some background to Malaysia as a net oil exporter and recent developments in its production of oil and fuel prices. Over the last two decades, Malaysia as a non-OPEC country has been a net exporter of natural gas and crude petroleum. Malaysia is important to world energy markets because, Malaysia has the world’s 10 th largest natural-gas reserves and the 27th largest crude oil reserves (EIA, 2008). Malaysia has six oil refineries, with a total capacity of 753,700 barrels per day. Oil reserves in the country continued to increase in 2008, rising to 4 billion barrels, or 22 years of lifespan (EIA, 2008), following the discovery of several deepwater oil fields in offshore Sabah. These fields now account for about one-fourth of its oil reserves and have led to higher investments in the sector, particularly from multinational oil companies and domestic service providers related to oil and gas. The recent rise in the oil price (of both crude oil and products) is one of a series of large shifts in price that have occurred during the last 30 years. Since 1983, the Malaysian government has been using the Automatic Pricing Mechanism to regulate and fix the domestic fuel prices. Its function is to stabilize the price of petrol and diesel in the country to a certain extent, via a variable amount sales tax and subsidy, so the retail price only has to be changed if the difference in price exceeds the threshold of the tax and subsidy, at the discretion of the government. From 1990-1999, the price for research octane number (RON) 97 was fixed at RM1.10 per litre. However, retail fuel prices were gradually increased from 1999 onwards following the rapid increase in crude oil prices. From RM1.10 per litre in 1999 for RON 97, it escalated to

126

THE JOURNAL OF WORLD ECONOMIC REVIEW

RM1.92 per litre in 2006. The domestic fuel price increased to a record high of RM2.70 per litre in June 2008 (RON97) when world oil price escalated to US$131 per barrel. Although Malaysia produces and exports oil, it is not a member of OPEC or a major oil-producing country. Thus, Malaysia has no influence on how the price of oil is determined internationally. If there is a large increase in oil prices on the world market, this affects the price of petroleum products such as diesel, petrol, and cooking gas (LPG) in Malaysia. 3. THEORETICAL FRAMEWORK 3.1 Demand and Supply Sides’ Impacts of Oil Price Shocks Fluctuations in oil prices have negative repercussions on the aggregate economy, as shown by economic literature. An oil price shock, as an example of an adverse supply shock, results in a rise in price level and a reduction in output and employment (Dornbusch et al., 2001). On the other hand, aggregate demand decreases as higher commodity prices translate to lower demand for goods and services, resulting in contraction in aggregate output and employment levels. The macroeconomic effects of oil shocks are transmitted via supply and demand side channels and are potentially minimised by economic policy reactions. Since oil is a factor of production in most sectors and industries, a rise in oil prices increases companies’ production costs and thus stimulates contraction in output (Jimenez-Rodriguez and Sanchez, 2004). Given a firm’s resource constraints, the increase in the price of oil as an input of production reduces the quantity it can produce. Hunt et al., (2001) add that an increase in input costs can drive down non-oil potential output supplied in the short run given existing capital stock and sticky wages. Moreover, workers and producers may respond to the decline in their real wages and profit margins by putting upward pressure on unit labour costs and the prices of finished goods and services. As discussed earlier, oil price increases translate to higher production costs, leading to commodity price increases. Higher commodity prices then translate to lower demand for goods and services, therefore shrinking aggregate output and employment. Furthermore, higher oil prices affect aggregate demand and consumption in the economy. The transfer of income and resources from oil-importing to oil-exporting economies is projected to reduce worldwide demand as demand in the former is likely to decline more than it will rise in the latter (Hunt et al., 2001). This is particularly true when the marginal propensity for oil importers to consume is higher than that of oil exporters. The resulting lower purchasing power of the oil-importing economy translates to a lower demand for goods and services. In sum, an increase in oil prices causes a leftward shift in both the demand and the supply curve, resulting in higher prices and lower output. 3.2 Why Does Asymmetry Effects of Oil Shock Arise? The linearity or symmetric assumption between real oil prices and macroeconomic variables (for oil importing countries) implies that if oil price increases causes economic recession, then oil price declines must cause an economic expansion with the same magnitude, although in reverse direction. On the other hand, the asymmetric effects of oil price shocks assume that an oil price decrease may actually lower future GDP growth through other channels. Hamilton (1988) suggests that asymmetry could be the result of adjustment costs to changing oil prices.

OIL PRICE SHOCKS AND MACROECONOMIC ACTIVITIES IN MALAYSIA

127

Falling oil prices stimulate economic activity, and rising oil prices impede economic activity, but the costs of adjusting to changing oil prices also impede economic activity. Combining these elements, it can be seen that rising oil prices would present two negative effects for economic activity. Falling oil prices would present both a negative and a positive effect which would tend to be offsetting. Another possibility is that monetary policy may account for the asymmetric response of aggregate economic activity. The way monetary policy is conducted plays a significant role in how the consequence of oil-price shocks play out in the economy. Bohi (1991), Bernanke et al., (1997), Leduc and Sill (2004) and Jumah and Pastuszyn (2007) among others, argue that contractionary monetary policy following an oil price increase accounts for the decline in aggregate economic activity. Ferderer (1996) suggests a third possibility. Uncertainty and financial pressure brought on by changing oil prices could amplify the negative effects of rising oil prices and offset to some degree the positive effects of falling oil prices. 4. DATA AND METHOD 4.1 The Data This paper uses quarterly data for the period 1991:1 to 2007:4 for five macroeconomic variables (defined below) and four measures of oil price shocks. The four measures of oil price shocks are quarterly oil price changes (∆ ROIL), Hamilton’s (1996) Net Oil Price Increase (NOPI), Lee et al.’s (1995) Scaled Oil Price Increase (SOPI), and Mork’s (1989) Oil Price Increase and Decrease. All macroeconomic variables and oil price variables are expressed in logs. The data sets were obtained from International Finance Statistics (IFS), the Economic Planning Unit (EPU), and the Statistics Department of Malaysia. The six variables used in this paper are defined as follows: i. Real gross domestic product (RGDP) is a measure of total output for the Malaysian economy. RGDP is expressed in RM. ii. Real effective exchange rate (REER) index of the RM is the nominal effective exchange rate index (NEERI) of the RM adjusted for inflation rate differentials with countries whose currencies comprise the NEERI basket. REER is defined such that an increase means a real appreciation of the currency considered. An appreciation of the real exchange rate is expected to harm the country’s external competitiveness. iii. Real government expenditure (GOVT) measures the total Malaysian government public expenditures (e.g. payments of governmental employees and subsidies). The GOVT variable partly reflects the role of oil subsidy policy undertaken by the government to regulate fuel prices in Malaysia. GOVT is expressed in RM. iv. Real exports (EXPORT) measure the importance of goods exports which includes oil exports in Malaysia. EXPORT is expressed in RM. v. Inflation (CPI variable) is measured via the consumer price index in Malaysia. vi. Oil price shock variables are divided into two specifications: linear and non-linear. The choice of five macroeconomic variables is based on work by Farzanegan and Markwardt (2009). However, real imports and real industrial GDP variables as employed in Farzanegan

128

THE JOURNAL OF WORLD ECONOMIC REVIEW

and Markwardt (2009) are substituted with real export and RGDP in this paper. The use of RGDP is consistent with previous literature (see Lee et al., 1995, and Hamilton, 2003, for examples), while the use of real exports instead of real imports reflects the significant contribution of the exports sector to Malaysian GDP. 4.2 Oil Price Shock Variables An important question is how to incorporate oil prices into the model. A number of studies have used different oil price variables to account for the effects of these shocks on economic activity. According to studies on the US (Hamilton, 1996; Mork, 1989; Lee et al., 1995), the effect of oil prices on growth is asymmetric. Thus, by defining the real oil price in time t as roilt = log (O*Et /PPIt), where O is Dubai crude oil price in US$, Et is the exchange rate of Malaysia against US$, and PPIt is the producer price index of Malaysia, this paper estimates the effects of oil price shocks using linear and non-linear specifications of a VAR model, which is discussed in the next section. The linear specification of oil price is based on quarterly changes in real oil prices, that is, the conventional first difference transformation of oil price variables (in logs): ∆ roilt = roilt – roilt – 1 (1) where roilt is the real oil price (in logs) in local currency, i.e. RM, as defined above. On the basis of asymmetric oil effects, this paper also uses three non-linear specifications of oil prices: (1) separate oil price variables for price increases and decreases, as in Mork (1989); (2) net specification (Hamilton, 1996 and 2003), where the relevant oil price variable is defined as the net amount by which these prices in quarter t exceed the maximum value reached in the previous four quarters; and (3) scaled oil price increase and decrease, proposed by Lee et al., (1995). In Mork (1989), the asymmetric specification distinguishes between the positive rate of change in the oil price, +, and its negative rate of change, –, which are defined as follows: roilt+ = max (0, (rolit – roilt – 1)), (2) – roilt = min (0, (rolit – roilt – 1)). (3) Hamilton (1996) proposed a different non-linear transformation, using an explanatory variable called net oil price increase (NOPI). This is defined as the amount by which oil prices in quarter t, roilt , exceed the maximum value over the previous four quarters, and zero otherwise. That is: NOPIt = max {0, roilt – max {roilt – 1 – roilt – 2, roilt – 3, roilt – 4}}. (4) Hamilton’s (1996) definition is also asymmetric in the specific sense that it captures oil price increase-type shocks while neglecting the impact of oil price declines. This is inspired by earlier evidence that oil price decreases played a smaller role in the US business cycle. Lee et al., (1995) argued that oil price shock is likely to have greater impact on real GNP in an environment where oil prices are stable than in one where oil price movements are frequent and erratic. The AR (4)-GARCH (1, 1) model is calculated as below: ∆ roilt = β0 + β1 ∆ roilt – 1 + β2 ∆ roilt – 2 + β3 ∆ roilt – 3 + β4 ∆ roilt – 4 + εt

(5)

εt | It – 1 ~ N (0, ht)

(6)

OIL PRICE SHOCKS AND MACROECONOMIC ACTIVITIES IN MALAYSIA

129

ht = γ0 + γ1 e 2t – 1 + γ2 ht – 1

(7)

SOPI t = max (0, eˆt / hˆt )

(8)

SOPDt = min (0, eˆt / hˆt )

(9)

where SOPI is scaled oil price increase, while SOPD denotes scaled oil price decreases. The scaled model builds on the asymmetric model, where it employs a transformation of the oil price that standardises the estimated residuals of the autoregressive model by its time-varying (conditional) variability. This transformation seems very plausible in light of the pattern of oil price changes over time, with most changes being rather small and punctuated by occasional sizeable shocks (Jiménez-Rodríguez and Sánchez, 2004). 4.3 The Vector Autoregression (VAR) Model A number of the studies cited have made use of VAR models. This technique treats all variables in the system as endogenous and regresses each current (non-lagged) variable in the model on all the variables in the model lagged a certain number of times. The VAR technique is appropriate in this case because of its ability to characterise the dynamic structure of the model as well as its ability to avoid imposing excessive identifying restrictions associated with various economic theories. The VAR model may be viewed as a system of reduced form equations in which each of the endogenous variables is regressed on its own lagged values and the lagged values of all other variables in the system. This paper employs the following unrestricted VAR model of order p (VAR (p)): Yt = c + Σ Ai Yt – 1 + εt

(10)

where Yt is a (n × 1) vector of endogenous variables, c is the intercept vector of the VAR, Ai is the i th matrix of autoregressive coefficients, and εt is the generalisation of a white noise process. A separate VAR model is estimated for each measure of oil price shocks. Each VAR model includes RGDP, REER, GOVT, EXPORT, and CPI. Yt = [Oil price variables, GOVT, RGDP, CPI, REER, EXPORT]

(11)

The VAR system can be transformed into its moving average (MA) representation in order to analyse the system’s response to a real oil price shock, that is: ∞

Yt = ε + ∑ i = 0 γ i µt − i

(12)

where γ0 is the identity matrix, ε is the mean of the process. The MA representation is used to obtain the forecast error variance decomposition (FEVD) and the impulse response function (IRF). 5. ESTIMATION RESULTS 5.1 Test of Stationarity Ohanian (1988), among other researchers, cautions against interpreting results derived from VAR models estimated with potentially integrated regressors. Because of this, the ADF test for first-order

130

THE JOURNAL OF WORLD ECONOMIC REVIEW

unit roots is used to determine the proper transformation for each variable. Table 1 provides the unit root regression results in levels and first differences of the five macroeconomic variables entered in the model and the corresponding critical value of 10%, 5%, or 1% to reject the null hypothesis of the presence of a unit root. The ADF statistics in the table suggest that all five macroeconomic variables are integrated of order one I(1), whereas the first differences are integrated of order zero I(0). These non-stationary variables were transformed by taking their first differences in order to exhibit stationarity. As for the oil price variables, they are by definition stationary at levels or I (0) because of the construction of the variables themselves. Each oil price measure is derived from taking the first difference (or the change) between the present value and the past value of the price of oil according to the oil price variable definition. Besides, the graphical plots of each oil price measures reveal the no trend pattern which satisfies the stationarity at levels criterion. Thus, given the different orders of integration for the variables involved in the analysis, the paper needs to carry out an unrestricted VAR exercise for these variables with four lags, as found to be optimal by Hannan–Quinn information criterion (HQIC) (refer Table 2). Table 1 Unit Root Tests Level GOVT (log) RGDP (log) CPI (log) REER (log) EXPORT (log)

First difference

t-Statistics

Probability

t-Statistics

– 2.57684 – 2.29774 – 1.95557 – 2.60845 – 2.34454

0.2921 0.4288 0.6143 0.2781 0.4047

– 4.31009 – 4.10499 – 6.9044 – 5.75454 – 7.57397

Probability 0.00583 *** 0.0102 ** 0.0000 *** 0.0001 *** 0.000 ***

Note: One/two/three asterisks respectively denote rejection of the null hypothesis of the presence of unit root at a 10%/5%/1% critical level. Table 2 Optimal Lag Length Using the Hannan-Quinn Information Criterion VAR order p

Log linear real oil price

Hamilton net oil price increase

Mork oil price increase

Mork oil price decrease

Lee et al. SOPI

0 1 2 3 4* 5

-10.1378 -20.138 -20.405 -20.3071 -20.7585* -20.5883

– 12.1764 – 21.257 – 21.4501 – 21.2929 – 21.7105* – 20.9996

– 11.6963 – 21.1562 – 21.1626 – 21.0661 – 21.5434* – 20.8298

– 11.9479 – 22.419 – 22.7997 – 22.2804 – 23.0827* – 22.7192

– 8.2822 – 18.2802 – 18.0344 – 17.7178 – 18.4522* – 18.3578

Lee et al. SOPD – 8.27712 – 18.3259 – 18.5315 – 18.0207 – 18.7177* – 18.6848

* Optimal lag length. Significant coefficients in bold.

5.2 Impulse Response Function: Theory The next step is to estimate the VAR. The estimates along with their standard deviation values are available from the authors upon request. Since the estimates of individual coefficients in VAR do not have a straightforward interpretation, this paper will focus the VAR results in terms

OIL PRICE SHOCKS AND MACROECONOMIC ACTIVITIES IN MALAYSIA

131

of IRFs and variance decomposition. IRFs trace the effect of one standard deviation shock to one of the innovations on current and future values of each of the endogenous variables in the system. If the innovations εt are contemporaneously uncorrelated, the interpretation of the impulse response is straightforward. The i th innovation εt is simply a shock to the i th endogenous variable yt. According to Runkle (1987), reporting IRFs without standard error bars is equivalent to reporting regression coefficients without t-statistics. Accordingly, this paper estimates 90% confidence intervals for the IRFs. The middle lines in the IRF figures represent the IRF, while the bars represent confidence intervals. In this regard, when the horizontal line falls into the confidence interval, the null hypothesis – that there is no effect of oil price shocks on other macroeconomic variables – cannot be rejected. This paper classifies the variables from the most exogenous to the less exogenous. The following ordering of variables is applied for linear and three non-linear oil price specifications: real oil price variables, real government expenditure (GOVT), real GDP, inflation (CPI), REER, and EXPORT. The baseline ordering assumes that goods market prices adjust slowly to disturbances in the oil market. The ordering also assumes that GOVT contemporaneously affects RGDP following an oil price shock. This is plausible if GOVT is dependent upon the anticipated revenue that the government receives from oil exports. Other plausible orderings could also be constructed using the Cholesky decomposition. These alternative orderings may yield different outcomes and interpretations of the IRFs. In the baseline ordering, the real oil price variables are ranked as largely exogenous variables. This is because Malaysia’s oil production and exports account for less than 1% of total world oil output; hence Malaysia is incapable of influencing global oil prices. Furthermore, demand for crude oil is largely determined by global economic growth, speculator operations in oil markets, and the policies of key oil consumers on strategic petroleum reserves. The second variable in the ordering is government expenditure (GOVT). Oil revenue contributes up to 40% of Malaysian government income and is used for paying governmental employees, subsidies, and so on. Thus, it is a plausible assumption that changes in oil price will have an immediate effect on government expenditures. The latter is then allowed to feed into changes in real GDP (RGDP). RGDP is also affected instantly by the level of government demand. The government, as a main recipient of oil rents, attempts to distribute some of these through various types of expenditures. Positive development in oil prices, which results in higher levels of government expenditures and raises income per capita, pushes the effective demand upward. In the short run, the limited capacity of domestic supply and the rising cost of doing business may push general consumer prices upward, fuelling inflation. Increasing inflation due to oil price increase appreciates the REER. This measures the relative price of non-tradable goods to tradable ones and gauges the competitiveness of an economy. The REER is a weighted real exchange rate index, with the weights assigned to trading partners of the local economy. If domestic prices increase while prices abroad remain unchanged, this would increase the relative prices of tradables, and the competitiveness of the economy will fall. Finally, a shock in REER contemporaneously affects real exports in Malaysia. Any significant developments in exchange rate markets may also affect the competitiveness of Malaysian products and foreign trade.

132

THE JOURNAL OF WORLD ECONOMIC REVIEW

5.3 IRF: Results Figure 1 shows IRFs for one standard innovation shock in real oil price growth (∆ ROIL) (in linear definition) for 1990:1–2007:4. Generally, most of the variables show an increase during the first few quarters, with the exception of REER. The IRFs generated from the VAR model using the linear specification of crude oil prices shows that a positive oil price shock leads to an increase in GOVT, persisting for two quarters, after which government expenditures decreases. However, this effect is not statistically significant. The shocks in real oil price increase RGDP and EXPORT in the first and second quarters after the initial shock. However, such increase does not last long (i.e. RGDP and real exports decline after the second quarter).

Figure 1: IRFs for Linear Oil Price Specification ( ROIL) Shock to  ROIL

The impulse response of the REER suggests that, in the long run, this variable reacts to a symmetric shock in real oil price by appreciating. The significant real effective depreciation (i.e., decrease in REER value) in the short run between first and second quarters could be because rising oil prices lead directly to higher inflation for the major trading partners of Malaysia via higher import prices, while domestic prices of energy products are subsidised in Malaysia and are below global market prices. However, in the medium and long run, the second round effects of higher inflation in Malaysia through increasing nominal wages appear to surpass the inflation experienced by trading partners, leading to a real appreciation of the Malaysian ringgit. However, this is not statistically different from zero. Figure 2 to Figure 4 illustrate the IRFs based on one standard deviation shock to the three non-linear oil price increase measures for 1990:1 to 2007:4. Overall, the results are comparable. The response of GOVT to a positive oil price shock is not statistically significant for most quarters, although the shock causes GOVT to decline significantly between the third and fourth quarters and between the sixth and seventh quarters. As an oil-exporting country, Malaysia’s RGDP is expected to improve at least in the short run consequent upon positive oil price shocks. Results from Figure 2 to Figure 4 show that RGDP increases for two quarters following a one standard deviation shock to Hamilton’s NOPI, SOPI, and Mork’s Price Increase specifications. This confirms the stimulus effect of positive oil price shocks on domestic output. However, the

OIL PRICE SHOCKS AND MACROECONOMIC ACTIVITIES IN MALAYSIA

Figure 2: IRFs for Non-Linear Oil Price Specification (Hamilton’s NOPI) Shock to Hamilton’s NOPI

Figure 3: IRFs for Non-Linear Oil Price Specification (SOPI) Shock to Scaled Oil Price Increase (SOPI)

Figure 4: IRFs for Non-Linear Oil Price Specification (Mork’s +) Shock to Mork’s Oil Price Increase

133

134

THE JOURNAL OF WORLD ECONOMIC REVIEW

increase in output does not last long. RGDP declines continuously after the second quarter to the tenth quarter. This finding is consistent with those of Hamilton (1983) and Mork (1989), who found decreases in RGDP (or GNP) after an oil price shock. Furthermore, findings from positive oil price shock models are in line with those of Abeysinghe (2001), who concluded that, although the direct impact of high oil prices on Malaysia is positive, it cannot escape the contractionary effect on growth coming through its trading partners, and in the long run Malaysia would also lose out. Inflation (CPI) reacts positively to oil price increases in both the short run and the long run. The inflationary effects of positive oil price shocks on the Malaysian economy can be explained through the demand side and supply side. Increasing oil revenues lead to higher levels of government expenditures. Given the dominant role of the government in the domestic economy, current and capital expenditures of the government will rise as oil revenues increase. In addition, because of increased net foreign reserves of the Central Bank, the money supply will increase. The increased money supply and government expenditures will shift the aggregate demand curve upward. On the supply side, oil price increases raise production costs as domestic fuel price rises. Increased cost of doing business and lower profit margins would erode producers’ profits and may cause them to cut back on output, shifting the supply curve upward. Real export response to a shock on positive changes in real oil prices is positive and lasts until the end of the period. Real exports increase significantly for three quarters after the initial shock, and peak between the 2nd and 3rd quarters when measured using Hamilton’s NOPI, SOPI, and Mork’s Price Increase specifications. The medium- and long-term decreasing trend may be due to weak consumer demand from Malaysia’s major trading partners, namely the US, Japan, Singapore, and China, consequent upon the positive oil price shocks. The adverse impacts of oil price increase on these countries, especially the US, raise inflation and reduce the real disposable income of consumers, thus limiting their demand for imports. In an oil-exporting economy, an oil price boom is expected to lead to a real appreciation of exchange rates and a decline in nonoil exports in the long term. This is often taken as the main symptom of Dutch disease. This trend (real appreciation of REER) however is not observed for the positive oil shock measures, as the real appreciation of REER is not statistically significant. For Scaled Oil Price Decrease and Mork’s Oil Price Decrease specifications, Fig. 5 to Fig. 6 show that the responses of CPI, real exports, and RGDP to a one standard deviation negative shock are the opposites of those obtained from the increase specifications. Inflation declines continuously for five quarters, while real exports react negatively to negative oil price shock both in the short and in the long term. Real exports decline in the first quarter after the initial shock but it moves back to its pre-shock level at the 10th quarter. Oil price decrease also has a prolonged negative impact on RGDP between the first and second quarters. The impact on RGDP is significantly negative for the first two quarters. However, the response on RGDP is not statistically significant over the long term. This is in line with the findings of Mork (1989), whereby the impacts of negative oil price shocks on US output growth were not statistically significant. For REER, negative oil price shocks lead to significant real appreciation between the first and third quarters. This is expected because falling oil prices may reduce the inflation rates of trading partners by a larger proportion than in Malaysia, as many commodity prices in Malaysia, such as fuel, are state controlled. This then causes Malaysia’s REER to appreciate.

OIL PRICE SHOCKS AND MACROECONOMIC ACTIVITIES IN MALAYSIA

135

Figure 5: IRFs for Non-Linear Oil Price Specification (SOPD) Shock to Scaled Oil Price Decrease (SOPD)

Figure 6: IRFs for Non-Linear Oil Price Specification (Mork’s -) Shock to Mork’s Oil Price Decrease

The response of real government expenditure to negative oil price changes is not significantly different from zero. At first glance, this might seem counterintuitive. However, the Malaysian policy of reducing fuel subsidies during periods of low oil prices and using the oil revenues in part to finance capital expenditures and the payment of external debts has been effective in managing oil wealth. When comparing the results between the positive (increase) and negative (decrease) non-linear IRF simulations, the paper in general finds no significant difference between these two approaches. In other words, there is no evidence of an asymmetric relation between oil price and macroeconomic activities in Malaysia. In most cases, the IRF results of the negative non-linear specifications are just the opposites of the results obtained from the positive non-linear measures. This is in line with the linear approach in which oil prices are assumed to have symmetrical impacts on real activity. However, the paper does find some differences on the response of government expenditures (GOVT) to positive and negative oil price shocks. Positive oil price shocks (as observed from SOPI and Mork ) causes GOVT to increase for two quarters

136

THE JOURNAL OF WORLD ECONOMIC REVIEW

but subsequently decline until the 4th quarter before returning to its pre-shock level in the 5 th quarter. This may reflect the typical spending habit of the Malaysian government during period of high oil prices where its expenditure increases (in tandem with the increase in oil revenue) but declines when economic growth is low as evident from the decreasing RGDP from 3 rd quarter onwards (see Fig. 2 to Fig. 4 for evidence) . In contrast, the response of GOVT to oil price decreases does not exhibit any significant cyclical trends for the 10-quarter period. One possible reason could be the prudent oil wealth management by the Government during periods of low oil prices in financing public expenditures. The findings also show that positive oil price shocks may not necessarily lead to persistent positive impacts on GDP growth for an oil-exporting country such as Malaysia, in contrast to the findings of Berument et al., (2010) for Middle East countries. In other words, not all oil-exporting countries benefit from the hike in oil price, unless it is their major income earner. For countries such as Iraq, Kuwait, and the United Arab Emirates, oil accounts for more than 75% of GDP. For Malaysia, oil contributes only 6% to export earnings and accounts for less than 9% of GDP. The relatively small share of oil to the economy makes Malaysia no less vulnerable than some oil-importing countries to oil price shocks. Moreover, the Malaysian economy has been heavily dependent on continued growth in the US, China, and Japan as top export destinations and key sources of foreign investment. While Malaysia has profited from increased world energy prices, its strong dependency on manufacturing exports makes it vulnerable to fluctuations in global demand and the performance of its trading partners. The slowdown of the US economy in recent years and high crude oil prices are expected to impact export growth adversely, especially the manufacturing sector, resulting in lower forecast GDP growth as simulated by the IRF analysis. 5.4 Robustness Checks The IRFs obtained from the baseline ordering is based on the assumption that the Malaysian government manages its spending (GOVT) in anticipation of the revenue it receives from oil exports. Thus, allowing shocks to GOVT to contemporaneously affect RGDP. However, if the impacts of oil price shocks are allowed to transmit directly into price levels (CPI) or that the government spending is dependent upon what it actually earned from oil exports, different outcome from IRFs may be obtained. Thus, for robustness checks, the paper performed IRFs with two alternative orderings for symmetric and asymmetric definition of oil prices shocks 3. The first alternative ordering assumes prices are flexible – [Oil Price variable, CPI, REER, EXPORT, GOVT and RGDP]. This ordering allows the price level (CPI) to adjust contemporaneously to shocks in oil prices. The second alternative ordering follows the price rigidity assumption – [Oil Price variable, EXPORT, GOVT CPI, REER, and RGDP]. This ordering considers the contributions of oil export revenue to the GOVT. The IRFs obtained from the two alternative orderings are comparable to the IRFs from the baseline ordering. In spite of the alternative orderings based on the price flexibility and price rigidity, responses of the five macroeconomic variables to an oil price shock however remains unchanged. 5.5 VAR Results: Variance Decomposition Analysis Variance decomposition represents the VAR system dynamics by providing information about the relative importance of each random innovation to the variables in the model. It shows how

OIL PRICE SHOCKS AND MACROECONOMIC ACTIVITIES IN MALAYSIA

137

many of the unanticipated changes or variations of the variables in the model can be explained by various shocks. This sub-section presents variance decomposition analysis according to both linear and non-linear oil price specifications. Table 3 shows the variance decomposition of the VAR model using linear and non-linear oil price specifications. The columns in Table 3 are labelled alphabetically from A to E, each column represents the impact of oil price shocks using a different oil price specification on the other variables in the VAR system. The rows are numbered from 1 to 20. Each row represents the percentage of the change in each variable that is the result of an oil price shock which is shown 1, 4, 8 and 10 quarters in the future. For example, column A and row 5 is read as A5 and corresponds to the value 47%. For linear specification (∆ ROIL), oil price shocks contribute a relatively large share to the variation of RGDP and EXPORT. The largest effect of an oil shock on a variable’s variation is through RGDP, accounting for approximately 47% (A5) in the first period and 50% (A6) in the fourth period. Likewise, crude oil prices account for 25% to 52% (A17, A18) of real exports volatility between the first and fourth quarters. Meanwhile, crude oil prices are marginal sources of variation in government expenditure (GOVT). For inflation (CPI), the contribution of oil price shocks is about 28% (A10) in the fourth quarter, rising to 40% (A11) in the eighth quarter. Shocks to oil price account for up to 23% (A14) of shocks in REER in the second quarter, decreasing marginally to 17% (A15) in the eighth quarter. For non-linear oil price measures, both oil price increases and decreases affect the volatility of macroeconomic variables in the model to varying degrees. By and large, the contribution of positive oil price changes to each variable’s variation is greater than negative oil price changes, especially for inflation, GOVT and EXPORT. For inflation fluctuations, positive oil price shocks have a stronger short- and long-run role compared to negative oil price shocks. While negative oil price shocks (MORK”) account for just 6% (E10) and 14% (E11) of variances in inflation in the fourth and eighth quarters, respectively, positive oil price shocks (MORK+) explain about 30-45% (C10, C11) of inflation fluctuations for the same period. This again confirms the high inflationary pressures observed during periods of high oil prices. The other important aspect of asymmetric oil price shocks is in their effects on GOVT fluctuation. While negative oil price shocks (MORK” and SOPD) play almost no role in variations in this variable, positive oil price shocks (MORK+ and SOPI) have the largest effect in both the short and the long term. Positive oil price shocks explain about 6-7% (C1, D1) of fluctuations in GOVT for the first quarter after the shock, increasing to about 23% (C3, D3) in the eighth quarter after the shock. Again, positive compared to negative oil price shocks have larger explanatory effects on EXPORT fluctuations. This role increases gradually from 12% (C17) (when measured using MORK+) in the first quarter after the shock to about 49% (C20) at the end of the period. In contrast, the MORK” specification accounts for about 20% (E17 to E20) of variations in EXPORT over the ten-quarter period. Similarly, the contribution of Mork’s positive of oil price shocks to REER averaged around 18% (C9 to C12) during the ten-quarter period compared to just 9% (E9 to E12) when using Mork’s negative oil price shocks. For fluctuations of RGDP, Mork’s negative oil price shocks have a stronger short-run role compared to Mork’s positive oil price shocks. While Mork’s positive oil price shocks account for 27% (C5) of variances of RGDP in the first quarter, Mork’s negative oil price shocks explain about 46% (E5) of RGDP fluctuations for the same quarter.

138

THE JOURNAL OF WORLD ECONOMIC REVIEW

SOURCES OF SHOCKS

ROW RESPONSE VARIABLES

Table 3 Variance Decomposition COLUMN

∆ ROIL

NOPI

MORK+

SOPI

MORK-

SOPD

1 2 3 4

Real Govt 1/QTR 4/QTR 8/QTR 10QTR

1% 10% 14% 14%

0% 4% 15% 19%

6% 18% 23% 25%

7% 14% 23% 26%

0% 1% 1% 1%

1% 5% 7% 7%

5 6 7 8

Real GDP 1/QTR 4/QTR 8/QTR 10QTR

47% 50% 28% 22%

37% 37% 24% 21%

27% 38% 24% 21%

20% 33% 24% 23%

46% 40% 22% 18%

34% 30% 16% 13%

9 10 11 12

CPI 1/QTR 4/QTR 8/QTR 10QTR

2% 28% 40% 45%

0% 14% 27% 33%

2% 30% 45% 48%

1% 20% 31% 36%

0% 6% 14% 18%

0% 12% 13% 14%

13 14 15 16

REER 1/QTR 4/QTR 8/QTR 10QTR

9% 23% 17% 15%

7% 15% 14% 12%

2% 27% 23% 20%

3% 15% 22% 18%

13% 12% 9% 10%

11% 17% 12% 10%

17 18 19 20

EXPORT 1/QTR 4/QTR 8/QTR 10QTR

25% 52% 56% 55%

14% 36% 44% 43%

12% 45% 51% 49%

7% 29% 36% 36%

16% 20% 22% 22%

15% 25% 23% 23%

A

B

C

D

E

F

The variance decomposition also suggests that positive oil price shocks account for around one-third of variability in real output. The contribution of oil innovations to output movements in Malaysia is relatively lower when compared to major oil-producing countries such as Iran and Saudi Arabia. For example, Mehrara and Oskoui (2007) found that oil price shocks explain about 54% of output fluctuations in Iran in the first year, increasing to more than 73% after two years. In Saudi Arabia, oil innovations explain about 65% of output movements in the short and long run. The relatively lower influence of oil price changes on Malaysia’s RGDP can be attributed to the reasonably successful experience of Malaysia to diversify away from resource-based production, including oil. Since oil accounts for only 6% of export earnings, the instability in the world economy brought about by fluctuations in oil prices have relatively less influence on the Malaysian economy. When comparing the results obtained from the IRF, evidence of symmetric relationship between oil prices and macroeconomic variables (as observed from the variance decomposition) cannot be supported. This is because, in most cases, the contributions of positive oil price

OIL PRICE SHOCKS AND MACROECONOMIC ACTIVITIES IN MALAYSIA

139

shocks on the macroeconomic variables are higher than the negative oil price shocks. In other words, there is a difference between the effects of positive and negative oil price shocks on economic activities, pointing to an asymmetric relationship between oil prices and the macroeconomic variables. Due to the mixed evidence obtained from the analyses, the study cannot conclude whether the impacts of oil price shocks on the Malaysian economy is symmetric or asymmetric. The mixed results could partly be attributed to the relatively short sample period used in the study. The estimation results could be improved if the dataset could be extended up to the early 1980s, i.e. when the first two global oil shocks occurred. However, such dataset is currently unavailable for use. 6. CONCLUSION This paper has studied the effects of oil price shocks on the real economic activity of Malaysia. It has focused on the relationship between oil prices and GDP growth, analysed in terms of VAR using four specifications, namely a linear model and three leading non-linear specifications proposed in the literature. IRFs and variance decomposition were estimated to assess how oil price shocks move through major channels of the Malaysian economy and the contribution of shocks to the variability of the variables in the system. Five macroeconomic variables were taken into consideration: Real Government Expenditure (GOVT), Real Effective Exchange Rate (REER), Real Gross Domestic Product (RGDP), Real Export (EXPORT), and Inflation (CPI), together with four real oil price specifications. Prior to estimation, a unit root test and cointegration test were conducted to determine the structure of data to be used for each VAR model. While the paper found the variables to be integrated of order one, it also found evidence of cointegration for each of the VAR models employed. Therefore, the VAR models were estimated with data at levels instead of first differenced. The IRF obtained from the linear oil price specification indicated that oil price movements lead to declines in RGDP in the long term after experiencing growth in the short term. However, only marginal impacts are seen in government expenditure (GOVT). Analysis from the non-linear oil price specifications produced comparable results. Hamilton’s (1996) NOPI, Lee et al.’s (1995) SOPI, and Mork’s (1989) positive oil price shock measure increase aggregate output in the short term but negatively affect output growth in the long term. For non-linear oil price decrease specifications, real output responds negatively in the short term before recovering to its pre-shock level in the long term. The insignificant difference between the effect of oil price increases and decreases (as observed from IRF) suggest a symmetric relationship between aggregate economic activity and oil prices. Notwithstanding this, the variance decomposition estimated from the non-linear VAR model shows that oil price increases contribute much higher to the variability of RGDP, real exports and inflation than that of oil price decreases. This on the contrary points to an asymmetric relationship between oil prices and macroeconomic variables for the Malaysian economy. The mixed evidence found for this study may be attributed to the relatively short duration of dataset available in the study. Results from the VAR estimation reveal that increases in oil prices have a larger impact on the Malaysian economy compared to oil price decreases. The linear and non-linear oil price specifications show that RGDP responds negatively to oil price increase in the long term despite

140

THE JOURNAL OF WORLD ECONOMIC REVIEW

experiencing growth in the short term. Though Malaysia produces and exports oil, the contribution to the economy is relatively small. Therefore, the economic stimulus provided by higher oil-export earnings in Malaysia would be more than outweighed by the depressive effect of higher prices on economic activity in importing countries. In other words, the windfall revenue from the oil sector may not be sufficient to cushion the economy from slowdown experienced by neighbouring and major trading partners. This finding is consistent with those of Abeysinghe (2001) and Mehrara and Oskoui (2007), who reached similar conclusions about the long-run effects of oil price shocks on oil-exporting countries. Given these results, forthcoming fiscal policies must not assume that future effects of upcoming oil shocks will be the same as in the past. Nevertheless, analysing economic policy reactions amidst these shocks will show how effective the chosen monetary or fiscal policies are in minimising adverse effects. The recent hike in world crude oil prices has prompted the Malaysian government to encourage the use of alternative energy sources such as natural-gas vehicles (NGV). Since the implementation of the 8th Malaysia Plan (2001-2005), renewable energy has been identified as the fifth fuel alongside coal, gas, oil, and hydropower. This effort has minimized the dependence of Malaysia’s power-generation sector on oil and gas as more coal-fired power plants are approved. Its Small Renewable Energy Programme (SREP) also encourages the generation of electricity under 10 MW. Regarding its fuel subsidy policy, the government plans to reduce the gap between domestic retail and world prices gradually. This involves reducing the subsidies for consumers progressively, as has been done since June 2008. By increasing prices gradually, consumers should be able to adjust their expenditure, hence minimising the adverse effects of future oil price shocks. Acknowlwdgement The author would like to thank Dr. Kwok Tong Soo for valuable comments on earlier drafts.

Notes 1.

The manufacturing sector has consistently contributed between 29% and 30% to Malaysian GDP since 2002 and accounts for 75% of gross export earnings (Malaysia Economic Report, 2008/2009).

2.

PETRONAS (Petroliam Nasional Berhad) is Malaysia’s nationally owned petroleum company.

3.

The results for other shocks are available upon request from the authors.

References Abeysinghe T., (2001), Estimation of Direct and Indirect Impact of Oil Price on Growth, Economics Letters, 73(2), 147-153. Bernanke B. S., Gertler M., and Watson M., (1997), Systematic Monetary Policy and the Effects of Oil Price Shocks, Brookings Papers on Economic Activity, 1, 157. Berument M. H., Ceylan N. B., and Dogan N., (2005), The Impact of Oil Price Shocks on the Economic Growth of Selected MENA Countries, The Energy Journal, 31. Burbridge J., and Harrison A., (1984), Testing for the Effects of Oil-Price Rises Using Vector Autoregressions, International Economic Review, 25(2).

OIL PRICE SHOCKS AND MACROECONOMIC ACTIVITIES IN MALAYSIA

141

CIMB, (2008), Special Reports on Fuel Price Increase-Subsidy Restructuring, Reality Bites, Published by the Edge Communications Sdn Bhd, Malaysia. Cunado J., and Perez de Gracia F., (2005), Oil Prices, Economic Activity and Inflation: Evidence for Some Asian Countries, The Quarterly Review of Economics and Finance, 45(1), 65-83. Darby M. R., (1982), The Price of Oil and World Inflation and Recession, American Economic Review, 72(4). Dornbusch R., S. Fischer, and R. Startz, (2001), Macroeconomics, (8th Edition), New York, McGraw Hill. Eltony M. N., and Al-Awadi M., (2001), Oil Price Fluctuations and Their Impact on the Macroeconomic Variables of Kuwait: A Case Study Using a VAR Model, International Journal of Energy Research, 25, 939-959. Engle R. F., and Yoo B. S., (1991), Forecasting and Testing in Cointegrated Systems, In Engle R. F., & Granger C. W. J., (Eds.), Long-Run Economic Relationships: Readings in Cointegration, Oxford University Press, USA. Farzanegan M. R., and Markwardt G., (2009), The Effects of oil Price Shocks on the Iranian Economy, Energy Economics, 31, 134-151. Hamilton J. D., (1983), Oil and the Macroeconomy Since World War II, Journal of Political Economy, 91(2), 248. Hamilton J. D., (1994), Time Series Analysis, New Jersey, Princeton University Press. Hamilton J. D., (1996), This is what Happened to the Oil Price-Macroeconomy Relationship, Journal of Monetary Economics, 38(2), 220. Hamilton J. D., (2003), What is an Oil Shock? Journal of Econometrics, 113, 363-398. Hunt B., Isard P., and Laxton D., (2001), The Macroeconomic Effects of Higher Oil Prices, IMF Working Paper. Hunt B., (2006), Oil Price Shocks and the U.S. Stagflation of the 1970s: Some Insights from GEM, Energy Journal, 27(4), 61-80. Jaafar A. H., Chamhuri S., and Al-Amin A. Q., (2008), Impacts of External Price Shocks on Malaysian Macro Economy-An Applied General Equilibrium Analysis. Jiménez-Rodríguez R., and Sánchez M., (2004), Oil Price Shocks and Real GDP Growth: Empirical Evidence for Some OECD Countries, European Central Bank Working Paper Series, (362). Jimenez-Rodriguez R., (2009), Oil Price Shocks and Real GDP Growth: Testing for Non-Linearity, Energy Journal, 30, 1-23. Lee K., and Ni S., (1995), Oil Shocks and the Macroeconomy: The Role of Price Variability, Energy Journal, 16(4). Lee K., and Ni S., (2002), On the Dynamic Effects of Oil Price Shocks: A Study Using Industry Level Data, Journal of Monetary Economics, 49(4), 823-852. Lutkepohl H., (2007), New Introduction to Multiple Time Series Analysis, Berlin, Springer. Masih R., and Masih A. M. M., (1996), Macroeconomic Activity Dynamics and Granger Causality: New Evidence from a Small Developing Economy Based on a Vector Error-Correction Modelling Analysis, Economic Modelling, 13(3), 407-426.

142

THE JOURNAL OF WORLD ECONOMIC REVIEW

Mehrara M., and Oskoui K. N., (2007), The Sources of Macroeconomic Fluctuations in Oil Exporting Countries: A Comparative Study, Economic Modelling, 24(3), 365-379. Mork K. A., (1989), Oil and the Macroeconomy when Prices Go Up and Down: An Extension of Hamilton’s Results, Journal of Political Economy, 97(3). Mork K. A., Olsen O., and Mysen H. T., (1994), Macroeconomic Responses to Oil Price Increases and Decreases in Seven OECD Countries, Energy Journal, 15, 19-35. Ohanian L. E., (1988), The Spurious Effects of Unit Roots on Vector Autoregressions: A Monte Carlo Study, Journal of Econometrics, 39, 251-266. Runkle D. E., (2002), Vector Autoregressions and Reality, Journal of Business & Economic Statistics, 128-133. Sims C. A., (1980), Macroeconomics and Reality, Econometrica (Pre-1986), 48(1).