SHOULD THE GULF COOPERATION COUNCIL (GCC) FORM A CURRENCY UNIOM

Should Asian-African the Gulf Journal Cooperation of Economics Council and (GCC) Econometrics, Form a Currency Vol. 12, Uniom No. 1, 2012: 43-59 43 SHOULD THE GULF COOPERATION COUNCIL (GCC) FORM A CURRENCY UNIOM Sami Alabdulwahab * and Kevin Sylwester ** ABSTRACT Forming a currency union is a major undertaking. As evinced by the European Monetary Union, such decisions could provide numerous benefits but also contain potential dangers. In recent years, the Gulf Cooperation Council has considered forming its own currency union. This study considers whether such a union is appropriate. We construct the same structural vector autoregression for each of the six countries of the GCC. Shocks come from three sources: oil shocks, non-oil supply shocks, and demand shocks. We then examine impulse responses, variance decompositions, and the correlation of the shocks across countries. We generally find support that these shocks are sufficiently similar to justify the formation of a currency union. 1. INTRODUCTION For the past decade, the six countries of the Gulf Cooperation Council (GCC), consisting of Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the United Arab Emirates, has considered forming a currency union. Proposed benefits include the lowering of transaction costs and reduced uncertainty due to the absence of exchange rate fluctuations within the GCC. However, potential costs could also be large. These countries would then rely on a single monetary policy across the GCC. The appropriateness of this policy depends upon whether these countries face similar economic conditions. If not, then monetary policy faces increased dilemmas. What would a GCC central bank do if Saudi Arabia, for example, faced rising unemployment when Kuwait experienced high inflation? This paper follows others that consider whether a group of countries is sufficiently similar in that a single monetary policy is likely to be appropriate. 3 In other words, do the economic shocks that hit these countries move output and the price level in the same direction? Bayoumi and Eichengreen (1994) first used a similar approach for European countries by constructing a Structural Vector Autoregression (SVAR) with the growth rates of output and the CPI as the left-hand side variables. Supply and demand disturbances represent the exogenous shocks. They find that the correlation of these shocks is lower in Europe than in the U.S. Furthermore, responses from supply shocks transpire more slowly in Western Europe than in the U.S. Subsequent researchers have performed similar exercises with other regional candidates * Economics Department at Eastern Illinois University ** Department of Economics, MC 4515, SIU, Southern Illinois University, Carbondale, 62901, E-mail: ksylwest@siu.edu

44 Sami Alabdulwahab and Kevin Sylwester for a currency union. 4 Given the formal discussions by GCC and national officials regarding a currency union, we take the GCC as a relevant case study in determining the appropriateness of a currency union. The GCC originated on May 25 th, 1981. This union initially integrated foreign policy. A second step was taken when the Supreme Council agreed upon a free trade zone in 1983 and then a customs union in 1999. In 2001, the Supreme Council approved the timeline for monetary union, culminating in a common currency in 2010.Obviously, the 2010 objective was not met but nor has talk of a currency union died either. Fasano and Schaechter (2003) predict sizable benefits from such a union, including lower transaction costs and a more efficient allocation of resources. At first, one might believe that examining the similarities of shocks across these countries is trivial since they are all large producers of oil. However, Table 1 presents a different picture since oil and natural gas represent different shares of these economies. With such distinctions, forming a currency union could run into difficulties as asymmetric shocks could lead to different optimal policy responses in different countries. Past work has also considered how suitable is the the GCC to form a common currency area. Table 1 Production of Petroleum, Natural Gas, and Mining as a Percentage of GDP in 2008 Bahrain Kuwait Oman Qatar Saudi Arabia UAE 29.2 59.5 50.8 60.8 57.5 37.0 Zaidi (1990) finds that output in these countries does not respond similarly to unanticipated money growth, implying that shifts in a common monetary policy could have different effects across the region. Dar and Presley (2001) find that GCC countries have low levels of intraregional trade. This lower integration could make forming a currency union more difficult. 5 Jadresic (2002) argues that the success of such integrations depends on the implementation of additional coordination measures such as elimination of domestic and cross-border distortions of trade and foreign investment and also increases in the coordination of policies that enhance macroeconomic stability and integration. Darrat and Al-Shamsi (2005) examine GDP growth, inflation, exchange rates, and money stocks in these countries. They find common long-run movements but do not examine if short run movements are synchronous. Sturm and Siegfried (2005) find structural and monetary convergence but only slow fiscal convergence. In a study most similar to ours, Abu-Bader and Abu-Qarn (2008) analyze common trends and business cycles for GCC countries. They find no evidence of long-run or short-run comovement in output. Using a SVAR, they also find that demand shocks are typically symmetric but that supply shocks are asymmetric. However, they employ the same SVAR that is used in Bayoumi and Eichengreen (1994) for European countries or U.S. regions. We believe that this is an important first step but that other specifications should also be considered given the high reliance these countries have upon oil production. Consider a rise in the price of oil that increases government revenues and so government spending. Abu-Qarn and Abu-Bader (2008) treat such an event as a supply shock. But with an AD-AS model, a positive supply shock should lower the price level. Instead, the rising income and government expenditure resulting from

Should the Gulf Cooperation Council (GCC) Form a Currency Uniom 45 the oil price increase has often been inflationary in these countries and so oil shocks need not affect countries similarly as do other supply shocks. Moreover, if output growth and inflation are responding to oil price changes differently in these countries, then that would also imply that a common monetary policy might not be optimal. Therefore, to better apply the Bayoumi and Eichengreen (1994) methodology to the GCC, we consider a three-by-three SVAR with oil prices as the third variable. Shocks to the economy then stem from oil price movements, shifts in demand, and shifts in the non-oil supply of output. The rest of the paper is organized as follows. Section 2 presents the SVAR and the empirical methodology. Section 3 describes the data and presents diagnostic checks to ensure that their inclusion in the SVAR is appropriate. Results are given in section 4. Finally, section 5 provides concluding discussion. 2. METHODOLOGY We employ two methodologies. The first is from Alesina et al. (2003) who focus on price and output co-movements. They consider relative prices between two countries, i and j: P it / P jt where P denotes the price level. They then estimate the following equation: P P P ln b b ln b ln tij P P P (1) it i, t1 i, t2 0 1 2 jt j, t1 j, t 2 Equation (1) represents relative prices at time t between two countries conditional on these relative prices in the previous two periods. The most important component in equation (1) is the error term. The error term in this equation measures the relative price that is not predicated by its two lagged values. Their second step is to calculate the co-movement of the relative prices root-mean squared error by using the following equation: VP ij 1 T 2 ˆ tij T 3 (2) t1 Equation (2) implicitly captures the predicted relative price co-movement between the two countries. When the predicted relative price co-movement between the two countries becomes greater,the VP ij value will be lower. Following Alesina et al. (2003), we will also use the same procedure for output comovements. First, we will start with the ratio Y it / Y jt. Second, we will estimate the following equation that represents relative income today conditional on the previous two periods: Y Y Y ln c c ln c ln tjj Y Y Y (3) it i, t1 i, t2 0 1 2 jt j, t 1 j, t 2 After estimating (3), we use the residual v tkk to calculate the relative output co-movements by using the following equation:

46 Sami Alabdulwahab and Kevin Sylwester VY ij 1 T 2 vˆ tij T 3 (4) t1 As the VY ij value becomes lower relative to the output,predicted output co-movements between the two countries become larger. An advantage of the above methodology is that no identifying restrictions are required to analyze output and price co-movements across countries. However, a disadvantage is that one cannot examine similarities across the underlying shocks but only commonalities among outcomes. This is a problem if countries enact different policies to produce similar outcomes. For example, suppose a positive demand shock hits country A, whereas a negative demand shock occurs in country B. If country A enacts a contractionary monetary or fiscal policy, whereas policy makers in country B launch expansionary policies, then output and inflation in the two countries might not greatly move in either country. The above methodology would then show these countries to have similar output and price movements even though these commonalities are results of different policy mixes. Furthermore, these similarities of outcomes might not occur in a common currency union where only a common monetary policy can be chosen. To determine whether these countries face similar underlying shocks, we consider a second methodology, namely structural VAR s. Bayoumi and Eichengreen (1994) use structural vector autoregressions (SVAR) to analyze the shocks during and after the Bretton Wood s agreement. They use a two dimensional SVAR by introducing two shocks, demand and supply shocks. They restrict the model by imposing the assumption that in the long run there is no effect of demand shocks upon output (vertical long-run supply curve). A positive demand shock would then raise price and output in the short run but only prices would increase in the long run. This restriction will allow the SVAR to identify the structural demand and supply shocks. In this study, we will introduce oil shocks to the model since the oil sector is a large portion of the macro structure of these countries. Much of the literature exploring similarity of shocks in potential common currency areas follows Bayoumi and Eichengreen (1994) in that only supply and demand shocks are examined. However, oil price shocks are also important for GCC countries, albeit to varying degrees. In one sense, though, such shocks could be categorized as aggregate supply shocks. But on the other hand, adjusting to oil price shocks could entail using different policy instruments than the usual fiscal and monetary policy tools. Therefore, examining oil price shocks separately from other types of supply shocks can not only better determine why countries are hit with different shocks, but can also hold implications for how OPEC s policy of influencing oil prices can help create a viable currency union. For example, if the effects of oil price shocks are of different magnitudes in GCC countries, then stabilizing oil price fluctuations could be important for the success of a GCC common currency area. Thus, we will explicitly account for oil price shocks in the analysis by extending the two-by-two SVAR from Bayoumi and Eichengreen (1994) to a three-by-three SVAR with percentage changes in real oil prices as a third variable.

Should the Gulf Cooperation Council (GCC) Form a Currency Uniom 47 Identifying the three-by-three model requires nine restrictions. The first three are that the three structural shocks are orthogonal. The second three consist of normalizations for the variance of the structural shocks. These first six are standard in the literature, and analogous assumptions are made in Bayoumi and Eichengreen (1994). The final three are also common and place restrictions on the long-run effects of shocks. Like Bayoumi and Eichengreen (1994), we assume that demand shocks have no long-run effect on output; that is, the long-run aggregate supply curve is vertical. However, two additional assumptions are required when real oil prices are included in the model. We assume that neither demand shocks nor supply shocks have long-run effects on real oil prices. It is certainly true that shocks in these countries could contemporaneously influence their production of oil and so influence the current global price of oil. For example, an increase in demand could lead to increases in oil production to generate revenues for this added spending. On the other hand, negative supply shocks could also lead to more oil production to compensate for less income from other sectors of the economy. Given the importance of GCC countries for global oil supplies, such actions could have non-negligible effects on oil prices. Nevertheless, we assume that the long-run effect of such shocks on world oil prices is zero. For one, non-oil shocks in these countries are unlikely to have a large impact on the longrun demand for oil which is mostly driven by global income growth and long-run technological change. Second, domestic (that is, within the GCC) demand and non-oil supply shocks are also unlikely to affect long-run oil production capabilities in other countries and so would not affect oil production in these other countries. Finally, such domestic shocks would only influence long-run oil supplies from these GCC countries if they influence long-run production capabilities and if future production is constrained by the production potentials determined by future infrastructure 6. That is, even if today s shocks affect investments into the petroleum sector, they will only influence future oil production if future governments reach capacity constraints as to how much oil they can extract and sell. Given that these cases rely on a specific set of conditions that we do not consider general or prevalent, we complete the identification of the model by assuming that demand and non-oil supply shocks do not affect long-run real global oil prices. Formally, the model is: OPt a11 ( L) a12 ( L) a13 ( L) Oilt Y t a21 ( L) a22 ( L) a23 ( L) st Ai t i A( L) t (5) i0 P a31 ( L) a32( L) a33 ( L) t dt OP represents the change in oil prices, Y represents the change in GDP, and P represents the change in the price level. a ij (L) are polynomials, and A are matrices in the lag operator L. Furthermore, a ij (L) drive the path of the effect of shock i on variable j. Oilt, st, and dt represent oil shocks, supply shocks, and demand shocks. The estimation starts with a reduced form of the VAR as follows: xt B( L) xt 1 t (6)

48 Sami Alabdulwahab and Kevin Sylwester The v t is a vector reduced form disturbance where the Moving Average (MA) is represented by the following equation: x t = C(L) v t (7) By introducing equations (5) and (7), we can see the relationship between them can be captured by the following equation: v t = A 0 t (8) To impose our identifying restrictions, we assume that supply and demand shocks have no long-run effect on oil prices: a 12 (L) = 0 and a 13 (L) = 0. We also assume from the AD-AS model that demand shocks have no effect on output in the long-run: a 23 (L) = 0. 3. DATA The data consists of Crude Oil Prices, (real) GDP, and the CPI from 1963 to 2005 7. GDP comes from the World Economic Outlook of the IMF and the CPI comes from the IMF s International Financial Statistics. We omit 1990 and 1991 from Kuwait due to the Iraqi invasion. All variables in levels are expressed as natural logarithms. Figures 1 and 2 present the growth rate of GDP and the CPI for all six countries. From figure 2, there is an upward trend in most of the GCC countries CPIs. Of note in figure 1 is the greater similarity of growth rates since the mid 1990 s, possibly due to the growing degree of economic policy coordination among GCC members. Figure 1: GDP Growth

Should the Gulf Cooperation Council (GCC) Form a Currency Uniom 49 Figure 2: The CPI of all GCC Countries 3.1. Descriptive Statistics Table 2 provides descriptive statistics. Table 2 Descriptive Statistics Country S. Arabia Kuwait UAE Oman Qatar Bahrain GDP Growth Rate Mean 0.084 0.043 0.072 0.088 0.049 0.052 Median 0.062 0.025 0.080 0.080 0.046 0.039 Std. Dev. 0.183 0.234 0.166 0.121 0.127 0.080 Inflation Rate Mean 0.032 0.040 0.097 0.032 0.085 0.038 Median 0.010 0.042 0.054 0.037 0.056 0.021 Std. Dev. 0.071 0.057 0.076 0.039 0.072 0.061 Percentage Change in Oil Prices (same for all countries) Mean 0.032 Median -0.022 Std. Dev. 0.276

50 Sami Alabdulwahab and Kevin Sylwester 3.2. Unit Root Tests Implementation of a SVAR requires the data to be stationary. To check for stationarity, we use Augmented Dickey Fuller (ADF). We use Akaike Information Criterion (AIC) to specify the lag length for the series. For each series, we test for unit roots in the three series in levels as well as in first differences. Table 3 presents results. All series exhibit a unit root when expressed as levels. When taking first differences, the null of no unit root is rejected at conventional levels for all series except Bahrain s GDP and Oman s CPI. However, failing to reject the null does not mean that the null is true. Moreover, to keep specifications similar across countries we consider the first differences of GDP, CPI, and oil prices. Table 3 Unit Root tests Country S. Arabia Kuwait UAE Oman Qatar Bahrain GDP in Levels ADF Statistic 2.094-2.263 2.385 2.516 2.369 1.361 # Lags 1 1 1 1 1 1 GDP in first differences ADF Statistic -5.006-5.975-4.413-4.594-3.971-1.139 # Lags 1 1 1 1 1 4 CPI in levels ADF Statistic 1.002 4.149 0.077 0.387 0.808 1.336 # Lags 3 1 1 1 1 1 CPI in first differences ADF Statistic -2.685-4.215-1.622-1.462-1.758-2.127 # Lags 1 1 1 1 1 1 Oil Prices in Levels ADF Statistic 0.3983 # Lags 1 Oil Prices in first differences ADF Statistic -6.537 # Lags 1 The ADF critical values for 42 observations are -1.611 (10 percent), -1.949 (5 percent) and -2.622 (1 per cent). Before proceeding with the results, a final overview of the data suggests high correlations between countries with regards to both GDP growth rates and inflation rates. These correlations are reported in tables 4 and 5. The smallest growth correlation is 0.285 between UAE and Oman. All growth correlations between Saudi Arabia and the other GCC members exceed 0.5. To put these numbers in context, the smallest growth correlation between the Mid-East and Rocky Mountain regions of the U.S. found by Bayoumi and Eichengreen (1994) is 0.27. For Europe, the growth correlation between Ireland and Germany is only 0.09. On the other hand, inflation correlations are not as high for GCC countries, spanning 0.38 to 0.81 between Saudi Arabia and other GCC members although they still compare favorably with pre-emu Europe. Inflation correlations between Germany and other EMU countries spanned -0.07 to 0.69 as

Should the Gulf Cooperation Council (GCC) Form a Currency Uniom 51 reported in Bayoumi and Eichengreen (1994). Of course, inflation correlations between U.S. regions are higher given the common currency. Therefore, these raw correlations suggest that GCC countries are no less favorable for currency union than were European countries before 1992. However, one reason for these results could be that GCC countries enact different stabilizing policies because they are hit with different shocks. The next section will consider this possibility. Table 4 Correlation Table for GCC Countries GDPs Growth and Oil Prices Growth Country Bahrain Kuwait Oman Qatar Saudi UAE Oil Bahrain 1 0.501 0.455 0.513 0.701 0.608 0.579 Kuwait 0.501 1 0.270 0.389 0.687 0.337 0.894 Oman 0.455 0.270 1 0.600 0.509 0.285 0.460 Qatar 0.513 0.389 0.600 1 0.683 0.738 0.573 Saudi 0.701 0.687 0.509 0.683 1 0.637 0.728 UAE 0.608 0.337 0.285 0.738 0.637 1 0.379 Oil 0.579 0.894 0.460 0.573 0.728 0.379 1 Table 5 Correlation Table for GCC Countries CPIs Growth and Oil Prices Growth Country Bahrain Kuwait Oman Qatar Saudi UAE Oil Bahrain 1 0.150 0.464 0.554 0.809 0.562 0.250 Kuwait 0.150 1 0.322 0.494 0.380 0.469 0.240 Oman 0.464 0.322 1 0.720 0.389 0.764 0.130 Qatar 0.554 0.494 0.720 1 0.557 0.969 0.192 Saudi 0.809 0.380 0.389 0.557 1 0.555 0.213 UAE 0.562 0.469 0.764 0.969 0.555 1 0.162 Oil 0.250 0.240 0.130 0.192 0.213 0.162 1 Tables 6 and 7 present VP values for price and output co-movements among countries. The values in table 6 are similar in magnitude to those found by Alesina et al. (2003) regarding several European countries and the averages with the other members of the Euro 12 from 1960 to 1997. Given that averaging provides less weight on country specific aberrations, the results in table 6 comparing individual GCC countries to one another provides evidence that these GCC countries are no more dissimilar than European countries were pre-euro, at least in regards to relative prices. However, table 7 tells a different story regarding output co-movements. These values for VY are an order of magnitude higher than their European counterparts in Alesina et al. (2003), suggesting that these countries have been hit by different shocks. These shocks appear to have had bigger effects on output than upon prices, perhaps because these countries already maintain some degree of monetary coordination among them. Moreover, the fixed exchange rate policies of these countries could also mute the impacts of shocks upon prices.

52 Sami Alabdulwahab and Kevin Sylwester Table 6 Price Co-Movement among GCC Countries Measured as Root Mean Square Error Country Saudi Kuwait UAE Oman Qatar Bahrain Saudi 1 Kuwait 0.0633 1 UAE 0.0393 0.0569 1 Oman 0.0439 0.0560 0.0305 1 Qatar 0.0416 0.0557 0.0195 0.0319 1 Bahrain 0.0414 0.0746 0.0348 0.0449 0.0382 1 *The lowerthe value the higherthe co-movement Table 7 Out-Put Co-Movement among GCC Countries Measured as Root Mean Square Error Country Saudi Kuwait UAE Oman Qatar Bahrain Saudi 1 Kuwait 0.1988 1 UAE 0.1619 0.2434 1 Oman 0.1728 0.2225 0.1929 1 Qatar 0.1517 0.2056 0.1420 0.1089 1 Bahrain 0.1587 0.2159 0.1384 0.1167 0.1084 1 *The lower the value the higherthe co-movement 4. SVAR RESULTS Table 8 presents correlations across the three shocks. Saudi Arabia is used as the basis for comparison. Correlations across oil shocks are quite high and range from 0.57 to 0.88. 8 Correlations across supply shocks and demand shocks are less strong. In fact, most correlations across supply shocks with Saudi Arabia s are negative. To put these numbers in better context, consider the Bayoumi and Eichengreen (1992) results for Europe. Their reported correlations are those with Germany. The range of supply shock correlations spanned -0.06 (Ireland) to 0.61 (Belgium). The range of demand shocks spanned -0.08 (Ireland) to 0.39 (Denmark). This range is comparable with our results for GCC countries regarding demand shocks but our results with supply shocks are far lower. Of course, correlations across oil price shocks are quite high. Table 8 Correlations of Shocks Across Countries Shocks Oil Price Supply Demand Saudi Arabia 1.00 1.00 1.00 Bahrain 0.88-0.19 0.33 Kuwait 0.62-0.04 0.14 Oman 0.67-0.19 0.27 Qatar 0.74-0.0004 0.06 United Arab Emirates 0.58 0.27-0.17

Should the Gulf Cooperation Council (GCC) Form a Currency Uniom 53 A second type of comparison is to see how the endogenous variables respond over time to these shocks. Figures 3 through 8 present impulse response functions of the endogenous variables to each of the three underlying shocks. From Figures 3, 5, and 7 output responds similarly in all three countries to each of the underlying shocks. From each type of shock, output increases but then stabilizes by the fourth year. Shocks to the price level (Figures 4, 6, and 8), however, show greater diversity. Demand shocks in figure 8have similar effects although some evidence arises of a slight nonmonotonicity for Saudi Arabia. Supply shocks in figure 6 lower prices for Qatar and Bahrain but appear to raise them (in contrast to the AD-AS model) for the other four countries. The price level responds similarly to oil price shocks in all countries except Kuwait. In the latter, some evidence arises that the price level slightly falls instead of rises. Of course, the lingering effects of the Iraqi invasion might have caused this different behavior. Of these cases, the only wide disparity across countries occurs for the effects of supply shocks upon the price level. Given these findings, the next question is to examine how important these respective shocks are for output and price level movements. Table 9 provides variance decompositions for output and prices in the short and long run. Oil price shocks are the main determinants of output fluctuations for all countries except Oman. They are the main drivers of price fluctuations for all countries but Kuwait and Oman. Of note is the apparent unimportance of supply shocks for price movements in Bahrain and Qatar. As seen above, impulse responses of the CPI from supply shocks differ for these two countries than for the four others. The variance decomposition Figure 3: The Effect of Oil Shock on GDP

54 Sami Alabdulwahab and Kevin Sylwester Figure 4: The Effect of Oil Shock on CPI Figure 5: The Effect of Supply Shock on GDP

Should the Gulf Cooperation Council (GCC) Form a Currency Uniom 55 Figure 6: The Effect of Supply Shock on CPI Figure 7: The Effect of Demand Shock on CPI

56 Sami Alabdulwahab and Kevin Sylwester results, however, suggest that these supply shocks are not the main determinants of price level movements in these countries. Table 9 Variance Decompositions Country Variable Period Oil shock Supply shock Demand shock 1 48.93 48.40 2.65 GDP 3 50.61 47.20 2.17 6 55.98 42.65 1.35 Bahrain Long-run 71.53 28.46 0.00 1 42.50 1.50 56.00 CPI 3 61.09 3.05 35.86 6 69.00 2.61 28.39 Long-run 76.13 2.31 21.56 1 64.29 31.79 3.92 GDP 3 70.46 27.46 2.08 6 73.26 25.53 1.21 Kuwait Long-run 77.22 22.78 0.00 1 14.05 37.23 48.72 CPI 3 11.70 37.97 50.33 6 10.70 38.22 51.08 Long-run 9.35 38.55 52.10 1 5.60 82.31 12.09 GDP 3 20.76 71.89 7.35 6 30.00 66.00 4.00 Oman Long-run 42.31 57.69 0.00 1 27.50 7.15 65.35 CPI 3 33.89 9.57 56.54 6 37.70 9.37 52.93 Long-run 41.21 9.03 49.76 Qatar 1 52.81 43.97 3.22 GDP 3 54.36 42.72 2.92 6 56.90 40.76 2.34 Long-run 71.97 28.03 0.00 1 47.16 14.49 38.35 CPI 3 52.43 8.95 38.62 6 55.28 6.55 38.17 Long-run 58.81 3.94 37.25 1 78.04 21.23 0.73 GDP 3 77.67 21.45 0.88 6 76.28 23.17 0.55 Long-run 73.95 26.05 0.00 contd. table 9

Should the Gulf Cooperation Council (GCC) Form a Currency Uniom 57 S. Arabia UAE 1 31.65 5.48 62 CPI 3 52.08 1.12 46.80 6 60.24 1.61 38.15 Long-run 64.65 2.00 33.35 1 68.53 0.51 30.96 GDP 3 84.64 1.22 14.14 6 91.99 2.56 5.45 Long-run 73.66 26.05 0.00 1 45.35 25.73 28.92 CPI 3 58.15 27.37 14.48 6 56.21 34.68 9.11 Long-run 11.56 78.51 9.93 To summarize the above results, we first note that correlations across shocks are high compared to what Bayoumi and Eichengreen found for pre-emu Europe and even compare similarly to correlations across U.S. regions. Second, the impulse response functions are generally similar across countries with the lone exception of how supply shocks affect price levels. However, price level movements are least affected by supply shocks. Instead, demand shocks and oil shocks play greater roles although not identically across countries. Therefore, we find great similarity of output movements across countries but more variation regarding price level movements. 5. CONCLUSION This study examined whether GCC countries should form a currency union, focusing on the similarity of output and price movements across these countries. Our results generally provide support for a currency union although the lack of correlation regarding supply shocks is one exception. Output growth and inflation co-movements are strong. Moreover, the shocks, especially oil shocks, hitting these economies are generally correlated, at least relative to pre- EMU shocks hitting European countries. Although these countries differ in how important oil is for these economies, oil shocks generally dominate according to the variance decompositions. Only for Omani output and Kuwaiti and Omani inflation do other shocks appear more important. Therefore, given the similarity of shocks across these countries, a common monetary policy appears appropriate. Our findings contrast those of Abu-Qarn and Abu-Bader (2008) who find less support for a currency union. One reason for our different results could be the different specification, allowing oil prices to explicitly enter the model. As stated, we believe that such a specification should at least be considered given the different structure of GCC economies relative to those of Europe or the U.S. Another reason for our different results could be that we use a longer stretch of data, going back to 1963, ten years before the first oil price shock. Of course, other concerns and issues also arise with currency unions. How much will transaction costs fall? Can these countries follow similar fiscal policies? Will a common currency allow some countries to overborrow? These other factors should certainly be taken into account

58 Sami Alabdulwahab and Kevin Sylwester and so our findings are not by themselves deterministic. Nevertheless, they do suggest that these countries are sufficiently similar (at least relative to other regions) so that a currency union is an appropriate move. Notes 1. See Bayoumi and Eichengreen (1994) and Alesina, Barro and Tenreyro (2002). 2. See: Zhang and Sato (2008), Ward and Jayaraman (2006), Buiget and Valev (2005), Chaplygin et al. (2006), Mikek (2009), and Bergman (1999) for various examples. 3. Hebous (2006) argues that such low intra-region trade integration could lower the benefits of forming a currency union. 4. Such as by affecting investment in various types of oil-producing infrastructure. 5. The crude oil price is taken as the price of the OPEC reference basket. 6. Theoretically, these oil price shocks should be perfectly correlated across countries since they represent the same change to the oil price. One reason their estimates could differ is that each estimate captures not the true shock but instead the force of the shock upon the country under examination which could differ across countries. References Abu-Qarn, A. and Abu-Bader, S. (2006), On the Optimality of a GCC Monetary Union: Structural VAR, Common Trends and Common Cycles Evidence. World Economy, 31, 612-630. Alberto, A., Barro, R. J.and Tenreyro, T. (2003), Optimal Currency Areasin: NBER Macroeconomics Annual 2002, 301-356. Bayoumi, T. and Eichengreen, B. (1994), Shocking Aspects of European Monetary Integration.In: Adjustment and growth in The European Monetary Union.(Ed.) F. Torres, and F. Giavazzi, Cambridge University Press, New York. Bergman, M. U. (1999), Do Monetary Unions Make Economic Sense? Evidence from the Scandinavian Currency Union, 1873-1913, Scandinavian Journal of Economics, 101, 363-377. Buigut, S. K. and Valev, N. T. (2005), Is the Proposed East African Monetary Union an Optimal Currency Area? A Structural Vector Autoregression Analysis, World Development, 33, 2113-2133. Buiter, W. H. (1999), The EMU and the NAMU: What is the case for North American monetary union? Canadian Public Policy, 25, 285-305. Chaplygin, V., Hughes-Hallett, A. and Richter, C. (2006), Monetary Integration in the Ex-Soviet Union: a Union of Four? Economics of Transition, 14, 47-68. Darrat, A. F. and Al-Shamsi, F. S. (2005), On the Path of Integration in the Gulf Region Applied Economics, 37, 1055-1062. Fasano, U. and Schaechter, A. (2003), Monetary Union among the Member Countries of the Gulf cooperation Council.Washington D.C.: IMF. Funke, M. (1997), The Nature of Shocks in Europe and in Germany.Economica, 64, 461-469. Hebous, S. (2006), On the Monetary Union of the Gulf States,Kiel Advanced Studies Working Paper No. 431, Kiel Institute for the World Economy. Jadresic, E. (2002), On a Currency for the GCC Countries.Washington, D.C: International Monetary Fund, Washington, D.C.

Should the Gulf Cooperation Council (GCC) Form a Currency Uniom 59 Laabas, B., &Limam, I. (2002), Are GCC Countries Ready for Currency Union? Kuwait: Arab Planning Institute, Kuwait. McKinnon, R. I. (1963), Optimum Currency Areas.American Economic Review, 53, 717-724. Mikek, P. (2009), Shocks to New and Old Europe: How Symmetric? Journal of Common Market Studies, 47, 811-830. Mundell, R. A. (1961), A Theory of Optimum Currency Areas. American Economic Review, 51, 657-665. Roland, V. (1990), Currency Competition and European Monetary Integration.Economic Journal, 100, 936-946. Sturm, M. and Siegfried, N. (2005), Regional Monetary Integration in the Member States of the Gulf Cooperation Council, European Central Bank Occasional Paper 31. Ward, B. and Jayaraman, T. K. (2006), A Single Currency for Pacific Island Countries? An SVAR Analysis, International Economics, 59, 83-111. Wyplosz, C. (1997), EMU: Why and How It Might Happen. Journal of Economic Perspectives, 11, 3 22. Zaidi, I. (1990), Monetary Coordination among the Gulf Cooperation Council Countries.World Development, 18, 759-768. Zhang, Z. and Sato, K. (2008), Whither a Currency Union in Greater China, Open Economies Review, 19, 355-370.