Feasibility of a currency union in East Asia using the five-variable structural vector autoregressive model

Economic Analysis and Policy 52 (2016) 45–54

Contents lists available at ScienceDirect

Economic Analysis and Policy journal homepage: www.elsevier.com/locate/eap

Full length article

Feasibility of a currency union in East Asia using the five-variable structural vector autoregressive model Najla Shafighi a , Behrooz Gharleghi b,∗ a

Faculty of Economics and Management, National University of Malaysia (UKM), 43600, Bangi, Selangor, Malaysia

b

Faculty of Business and Management, Asia Pacific University of Technology and Innovation, Bukit Jalil, 57000, Kuala Lumpur, Malaysia

article

info

Article history: Received 18 September 2014 Received in revised form 16 July 2016 Accepted 17 July 2016 Available online 21 July 2016 JEL classification: F31 F36 Keywords: Optimum currency area Structural vector autoregressive Exchange rate East Asian region

abstract Following the closer monetary cooperation among East Asian countries in recent years, this paper empirically investigates the feasibility of forming a currency union in the region by examining the symmetry of underlying shocks for the most recent period (post-crisis 1999–2013) and by testing the level of correlation of the shocks. Using a five-variable structural vector autoregressive model, we identify various types of shocks in ten East Asian economies. An impulse response function and variance decomposition of shocks are used to identify the size, speed of adjustments to the shocks, and the root cause of variability in macro variables. Empirical analysis suggests the capacity of Indonesia, Japan, Hong Kong, Korea, Malaysia and the Philippines to participate in a common currency area. © 2016 Economic Society of Australia, Queensland. Published by Elsevier B.V. All rights reserved.

1. Introduction The emergence of economic integration in a number of global regions, as demonstrated by the establishment of the North American Free Trade Agreement (NAFTA), the European Union (EU), the Mercado Común del Sur (Mercosur), the Economic and Monetary Community of Central Africa (CEMAC), the Organization of Eastern Caribbean States (OECS), the West African Economic and Monetary Union (UEMOA) and the Central American Common Market (CACM), has encouraged tighter economic integration in East Asia. At the same time, flow on effects from the 1997 financial crisis have increased economic inequality in the region while simultaneously generating new economic and political interest in strengthening monetary cooperation (Mishra and Sharma, 2010). One important aspect of monetary integration is its generally recognized macroeconomic benefits in the form of monetary policy which can better manage aggregate demand and promote investment in regional economies (Mundell, 1961). The evaluation of underlying shocks, including supply shocks, demand shocks and monetary shocks, is necessary to assess the feasibility of creating an optimum currency area (Xu, 2006; Soo and Choong, 2010). According to Mundell (1961) and MacKinnon (1963), the demand to peg the bilateral exchange rates of two economies rises with the bilateral intensity of trade, flexibility of factor markets, and symmetry of underlying shocks. However, the correlation of shocks is generally accepted as the main criterion for a country to join a currency union (Huang and Guo, 2006). Mundell (1961) argues that countries with positively correlated economic shocks are suitable candidates for forming a currency union because they tend to use similar policies to adjust imbalances.

∗

Corresponding author. E-mail addresses: [email protected] (N. Shafighi), [email protected] (B. Gharleghi).

http://dx.doi.org/10.1016/j.eap.2016.07.002 0313-5926/© 2016 Economic Society of Australia, Queensland. Published by Elsevier B.V. All rights reserved.

46

N. Shafighi, B. Gharleghi / Economic Analysis and Policy 52 (2016) 45–54

The main objective of the current paper is to assess the attractiveness and feasibility of creating a currency union in East Asia on the basis of the optimum currency area (OCA) theoretical framework. First, we focus on the symmetry of various structural economic shocks across economies. This contrasts with the literature which has largely adopted a straightforward approach mainly limited to the examination of cross-country macroeconomic variables, such as real output, the consumer price index (CPI) and interest rates (e.g., Ito, 1994; Takuchi, 1994; Goto and Hamada, 1994; Kwan, 1998; Kandil and Trabelsi, 2012; Hsu, 2010). Second, we develop a five-variable structural vector autoregressive (SVAR) model instead of the two-variable models developed in the earlier studies (and which consisted of only one supply shock and one demand shock—e.g., Bayoumi and Eichengreen, 1993; Lee and Koh, 2012), or the three-variable (e.g., Chow and Kim, 2003; Zhang et al., 2004) and four-variable models (e.g., Huang and Guo, 2006). To assess the extent of the symmetry of shocks in East Asian exportoriented countries, our model incorporates more than the typical two-variable model (Ling, 2001) integrating external supply, external monetary, domestic supply, demand, and monetary shocks into the SVAR model. Third, using the latest data and five-variable SVAR model can contribute to an improved understanding of the underlying forces that determine economic movements across East Asian countries after the 1997 financial crisis. Five ASEAN members (Indonesia, Malaysia, Philippines, Singapore and Thailand) and five East Asian countries (China, Hong Kong, Korea, Taiwan and Japan) are selected for this study. The remainder of this paper is organized as follows. Section 2 reviews some empirical studies on OCAs. Section 3 describes the data and presents the econometric methodology of the SVAR. Section 4 provides the empirical results and findings. Finally, Section 5 concludes with relevant policy implications. 2. Previous studies Bayoumi and Eichengreen (1993) published one of the first empirical papers that deal with macroeconomic disturbance. They apply a variant of the VAR model introduced by Blanchard and Quah (1989) to the members of the European Community to measure the nature of economic disturbance within groups of countries. Their SVAR model is premised in turn on the aggregate demand–aggregate supply (AD–AS) model in which a supply shock can influence output and price level in both the long-run and the short run, whereas a demand shock has no effect on output in the long-run (Bayoumi and Eichengreen, 1993). In a related study on monetary integration in East Asia, Bayoumi and Eichengreen (1994) apply a similar technique to differentiate demand and supply shocks and to estimate the respective correlations of these shocks. They find symmetry in supply shocks among Indonesia, Malaysia, Hong Kong and Singapore, and also between Korea and Japan. Therefore, it may be assumed that these two groups of countries are more likely to form an OCA than other countries in the region. Lee and Koh (2012) empirically assess the desirability of East Asian economies (ASEAN with three other East Asian countries) forming a monetary union using a two-variable SVAR model to measure the macroeconomic disturbances (supply shocks and demand shocks), and to identify potential candidates in forming an OCA. Their findings show that East Asian countries exhibit less symmetry in underlying shocks but faster adjustment to such shocks – especially after the financial crisis – and which therefore increases the likelihood of monetary integration among ASEAN countries. Zhang et al. (2004) employ a three-variable SVAR model consisting of supply, demand, and monetary shocks to test the symmetry of these shocks among East Asian countries. In doing so they analyze the feasibility of forming an OCA. They conclude from their results that an OCA is not feasible in the East Asian region. Chow and Kim (2003) investigate the feasibility of a common currency peg in East Asia based on a three-variable SVAR methodology, which differentiates global supply, regional supply, and domestic supply shocks. They find that domestic outputs of these countries are influenced more by country-specific shocks than by regional shocks. They further suggest that East Asian countries are structurally different from one another and, thus, are subject to asymmetric shocks. Therefore, based on the OCA theory, a common currency peg in East Asia will would be costly and difficult to sustain. Huang and Guo (2006) employ a four-variable SVAR model, which includes external global supply shocks, domestic supply and demand and monetary shocks in the assessment of the feasibility of creating an OCA in East Asia. This model accounts for the effect of external global supply shocks. Their findings confirm those of Bayoumi and Eichengreen (1994) who argue that this region is not ready to form an OCA, although Korea, Hong Kong, Indonesia, Malaysia, Singapore and Thailand are better suited to the creation of a currency union. In seeking to provide a more comprehensive and reliable model this paper adds two external global shocks (supply and monetary) and three domestic shocks (supply, demand, and monetary) to the literature by introducing the five-variable SVAR model. 3. Methodology In this section, we extend the previous works by improving the methodology used to evaluate the symmetry of shocks in East Asian countries. Underlying shocks can be global or country-specific. Therefore instead of a four-variable model (four shocks), we consider a model with two external global shocks (supply and monetary) and three domestic shocks (supply, demand, and monetary). Given that East Asian countries typically adopt an export-oriented strategy, incorporating two external shocks into the model provides more information and a more rigorous methodology on which to decide whether or not to adopt a common currency.

N. Shafighi, B. Gharleghi / Economic Analysis and Policy 52 (2016) 45–54

47

External global monetary shocks are a significant source of information in estimating how these economies respond to changes in their real exchange rate. This assessment is particularly useful in examining and adopting an optimal exchange rate policy. High levels of sensitivity of prices in real exchange rate changes demonstrate the viability of the real exchange rate stabilization policy. The possibility of forming a monetary union can thus be strengthened if the correlations in monetary shocks are positive. We employ a five-variable SVAR model to examine the symmetry of shocks according to the OCA literature. All variables – global supply (y∗ ), global monetary (p∗ ), domestic monetary (p), and domestic demand (e) – are  supply (y), domestic  expressed in natural logarithms and therefore, xt = y∗t , p∗t , yt , pt , et . Let

′  1xt = 1y∗t , 1p∗t , 1yt , 1pt , 1et ,

(1)

and ∗

∗

εt = (εts , εtm , εts , εtm , εtd )′ ,

(2) s∗ t

m∗ t

where ∆ represents the first-difference operator, and ε , ε , εts , εtm , and εtd denote global supply, global monetary, domestic supply, domestic monetary and domestic demand shocks, respectively. The structural model can then be written as xt = A0 εt + A1 εt −1 + A2 εt −2 + · · · + Ap yt −p =

∞ 

Ai εt −i .

(3)

i=0

In matrix form, it is expressed as xt = A(L)εt .

(4)

A is a 5 × 5 matrix that defines impulse responses of endogenous variables to structural shocks. It is assumed that shocks are serially uncorrelated with a covariance matrix normalized to the identity matrix. The model implies that economic variables are subject to five shocks; therefore, we decompose world real GDP, the world inflation rate, domestic real GDP, the domestic inflation rate, and the real exchange rate given they are combinations of the five types of shocks. The equations can then be written as; ∗

1y∗t = A11 (L)εts ,

(5)

s∗ t s∗

m∗ t m∗

s∗ t s∗

m∗ t m∗

1pt = A21 (L)ε + A22 (L)ε

,

1yt = A31 (L)εt + A32 (L)εt

+ A33 (L)ε ,

∗

(6) s t

(7)

1pt = A41 (L)ε + A42 (L)ε

+ A43 (L)ε + A44 (L)ε ,

1et = A51 (L)εt + A52 (L)εt

+ A53 (L)ε + A54 (L)ε + A55 (L)ε .

s t

m t

s t

(8)

m t

d t

(9)

World output is considered to evolve exogenously as presented in Eq. (5). World inflation is affected by world output only; therefore, domestic variables have no effect on global variables (Bernanke, 1986; Sims, 1986; Blanchard and Quah, 1989; Chow and Kim, 2003; Huang and Guo, 2006) as shown in Eq. (6). Domestic variables are affected by both the global and domestic shocks. The effect of domestic shocks on domestic variables leads to several assumptions listed below. (i) Monetary shocks and demand shocks are postulated to have no long-run effect on real output. This would imply the following restrictions: ∞ 

A34i = 0

and

i =0

∞ 

A35i = 0.

(10)

i=0

(ii) Demand shocks have no long-run effect on inflation. This is true especially when we consider the sticky price monetary model (Dornbusch, 1976; Meese and Rogoff, 1983) and the relative price monetary model (Chinn, 1998) of exchange rate determination—both of which state that price levels affect the exchange rate. That is, ∞ 

A45i = 0.

(11)

i =0

The given restrictions are sufficient to identify the above matrix. Therefore, we can rewrite the system as

A (L) 11 ∗ 1pt  A21 (L)    1yt  = A31 (L) 1p  A (L) t 41 1et A51 (L) 1y∗  t

0 A22 (L) A32 (L) A42 (L) A52 (L)

0 0 A33 (L) A43 (L) A53 (L)

0 0 0 A44 (L) A54 (L)

 ε s∗  t ∗  εtm    s  .   εt   m  εt  A55 (L) εd 0 0 0 0

t

(12)

48

N. Shafighi, B. Gharleghi / Economic Analysis and Policy 52 (2016) 45–54

The reduced-form VAR model for estimation is given by

1xt = B(L)1xt −1 + ut ,

(13)

where ui is a vector-reduced form disturbance. A moving average representation of Eq. (13) is expressed as

1xt = C (L)ut ,

(14)

where C (L) = (1 − B(L)L) and the real matrix C (L) is, by construction, C0 = I. By comparing Eqs. (3) and (14), we obtain the relationship between structural and reduced form disturbances as being ut = A0 εt . Hence, estimates of A0 are necessary in recovering the time series of structural shocks εt . Given that structural shocks are mutually orthogonal, and each shock has a unit variance, a relationship can be found between covariance matrices, and is expressed as −1

C (1)



C (1)′ = A(1)A(1)′ ,

(15)

where −Eut ut − EA0 εt εt A0 − A0 A0 . If H denotes the lower triangular Choleski decomposition of C (1) C (1) , then A(1)—H represents the long-run restrictions implying that A(1) is also lower triangular. Consequently, A0 = C (1)−1 A(1) = C (1)−1 H. ∗ ∗ Given an estimate of A0 , a time series of structural shocks, εt = (εts , εtm , εts , εtm , εtd )′ , can be recovered.



′

′ ′

′



′

4. Empirical results All data are expressed in logarithmic form and are drawn from the International Financial Statistics published by the International Monetary Fund. The period spans from 1999 Q1 to 2013 Q4 for Indonesia, Malaysia, Philippines, Singapore, Thailand, China, Japan, Korea, Hong Kong and Taiwan. Real GDP is real output expressed in US dollars, the interest rate (IR) is proxied by central bank policy interest rates; the CPI is a measure for change in prices and the exchange rate (EXR) is proxied by the real effective exchange rate. Stationary properties of time series under consideration have been examined, and all variables are found to be integrated of order one I(1) based on the Phillips–Perron and KPSS tests. Banerjee et al. (1992), and Vogelsang and Perron (1998) used unit root tests which endogenously determined break dates from the data (while considering all possible break dates) and the results indicating integration of the order of one. Therefore, the first-difference of all variables is used to ensure stationarity of variables. For the SVAR estimation, lag length is uniformly chosen to be based on the SBC given most systems show a lag length of one (The unit root results are available upon request from authors). 4.1. Correlations of structural shocks The degree of shock symmetry among countries under consideration of forming a currency union is calculated by correlating identified disturbances. For this reason, the correlation of five structural shocks is estimated using the SVAR model in ten East Asian economies from 1999 Q1 to 2013 Q4. Generally, the assumption is that if the correlation of shocks is positive, the shocks are considered symmetric, whereas if the correlation is negative or insignificant, then the shocks are considered asymmetric. In addition, we apply the (Kendall and Gibbons, 1990) correlation statistic to check whether the correlations are statistically significant. Conditions for the creation of a common currency area can be considered favorable in the East Asian region if results show symmetry and significance. 4.1.1. Correlations of external supply shocks Correlations of external supply shocks among East Asian economies are reported in Table 1. From Table 1, correlations of external supply shocks are evidently positive and significant among the chosen economies. The most likely reason for this is the adoption of an export-oriented strategy in these countries in recent years. Therefore, external supply shocks quickly transmit into the region. A high correlation of shocks from an external source thus generates greater advantage for countries to form a monetary union. 4.1.2. Correlation of external monetary shocks Table 2 displays the correlation of external monetary shocks. Similar to the results for external supply, correlations of external monetary shocks are positive and highly significant. The result can be attributed to the fact that these countries are part of the USD bloc (Moosavi and Azali, 2014; Lim, 2005). That is, any changes in level of prices in the United States are quickly transmitted through the region. 4.1.3. Correlation of supply disturbances Table 3 presents correlations of supply shocks among surveyed countries. We observe that supply shocks are strongly correlated only among China, Hong Kong, Taiwan and the four ASEAN countries, namely, Malaysia, Philippines, Singapore and Thailand. This correlation may exist due to the financial crisis that severely affected these economies. However, while correlations exists among the rest of countries they are statistically insignificant and no correlation exists among the countries with negative signs.

N. Shafighi, B. Gharleghi / Economic Analysis and Policy 52 (2016) 45–54

49

Table 1 Correlations of external supply shocks.

CH HK ID JP KR MY PH SG TH TW

CH

HK

ID

JP

KR

MY

PH

SG

TH

TW

1.00 0.68* 0.48* 0.59* 0.56* 0.64* 0.68* 0.61* 0.65* 0.54*

1.00 0.61* 0.61* 0.55* 0.66* 0.81* 0.75* 0.78* 0.61*

1.00 0.55* 0.66* 0.53* 0.61* 0.64* 0.64* 0.51*

1.00 0.67* 0.55* 0.66* 0.57* 0.62* 0.51*

1.00 0.58* 0.60* 0.57* 0.61* 0.59*

1.00 0.72* 0.75* 0.76* 0.60*

1.00 0.79* 0.83* 0.61*

1.00 0.79* 0.57*

1.00 0.64*

1.00

Note: * Denotes significance at 5%.

Table 2 Correlations of external monetary shocks.

CH HK ID JP KR MY PH SG TH TW

CH

HK

ID

JP

KR

MY

PH

SG

TH

TW

1.00 0.79* 0.60* 0.58* 0.57* 0.70* 0.66* 0.76* 0.53* 0.63*

1.00 0.70* 0.65* 0.65* 0.73* 0.68* 0.74* 0.52* 0.68*

1.00 0.70* 0.77* 0.62* 0.67* 0.64* 0.57* 0.58*

1.00 0.82* 0.58* 0.57* 0.61* 0.52* 0.61*

1.00 0.60* 0.58* 0.61* 0.54* 0.65*

1.00 0.74* 0.77* 0.57* 0.69*

1.00 0.75* 0.61* 0.57*

1.00 0.59* 0.68*

1.00 0.47*

1.00

Note: * Denotes significance at 5%.

Table 3 Correlations of supply shocks.

CH HK ID JP KR MY PH SG TH TW

CH

HK

ID

JP

KR

MY

PH

SG

TH

TW

1.00 0.41* −0.33 0.18* −0.05 0.30* 0.43* 0.18* 0.30* 0.49*

1.00 −0.20 0.01 0.10 0.36* 0.19* 0.18* 0.21* 0.50*

1.00 −0.01 0.06 0.10 −0.18 0.09 −0.22 −0.19

1.00 −0.14 0.02 0.02 0.01 −0.09 0.28*

1.00 0.04 −0.18 0.30* 0.00 0.01

1.00 0.26* 0.34* 0.02 0.41*

1.00 0.08 0.41* 0.23*

1.00 0.08 0.23*

1.00 0.03

1.00

Note: * Denotes significance at 5%.

4.1.4. Correlations of monetary disturbances Table 4 shows the correlation of monetary disturbances. Results reveal a significant correlation among Singapore, Philippines, Malaysia, Japan and Thailand. Another significant group consists of Korea, Malaysia and the Philippines, followed by yet another significant group consisting of China, Hong Kong, Korea and Indonesia. Correlations exists among ASEAN members but are statistically insignificant. Unlike Huang and Guo (2006) who identified a leading role for Japan in the region’s capital market, we identify in this study a correlation among Japan, Thailand and Taiwan only. This can be partially explained by the Japanese automobile companies’ increasing investment in these two countries. 4.1.5. Correlations of demand disturbances From Table 5, the correlations of demand disturbances (considered a transitory shock) are significant in all countries except for China and Hong Kong. This means that a significant correlation exists between the five ASEAN members and Korea and Taiwan. In addition, a significant correlation also exists among Japan, Malaysia, Singapore, Thailand and Taiwan. This reveals tight economic relationships among ASEAN members. On the other hand, China and Hong Kong show no correlation with other economies. This can be attributed to two factors. For China, the country remains a major source of imports for

50

N. Shafighi, B. Gharleghi / Economic Analysis and Policy 52 (2016) 45–54

Table 4 Correlations of monetary shocks.

CH HK ID JP KR MY PH SG TH TW

CH

HK

ID

JP

KR

MY

PH

SG

TH

TW

1.00 0.18* 0.27* −0.04 0.18* −0.08 0.09 0.09 0.02 −0.05

1.00 0.06 0.30* 0.13 0.06 0.21* 0.08 0.10 0.04

1.00 −0.20 0.10 0.21* 0.08 0.10 −0.15 −0.16

1.00 0.02 0.14 0.07 0.10 0.16** 0.20*

1.00 0.19* 0.24* 0.12 0.13 −0.12

1.00 0.19* 0.03 0.20* 0.06

1.00 −0.01 0.20* 0.12

1.00 0.16** 0.10

1.00 0.21*

1.00

Note: * Denotes significance at 5%. ** Denotes significance at 10%. Table 5 Correlations of demand shocks.

CH HK ID JP KR MY PH SG TH TW

CH

HK

ID

JP

KR

MY

PH

SG

TH

TW

1.00 −0.07 0.09 −0.03 0.09 0.05 0.12 0.09 0.11 −0.00

1.00 0.03 0.09 0.01 −0.05 −0.01 0.03 0.08 0.07

1.00 0.08 0.28* 0.27* 0.26* 0.43* 0.28* 0.20*

1.00 0.04 0.31* 0.02 0.28* 0.25* 0.24*

1.00 0.18* 0.29* 0.33* 0.34* 0.13

1.00 0.14 0.46* 0.33* 0.30*

1.00 0.15 0.31* 0.07

1.00 0.40* 0.30*

1.00 0.25*

1.00

Note: * Denotes significance at 5%.

ASEAN members; therefore, an increase in China’s price levels which are driven by its demand shocks can have only a minor impact on the demand of East Asian countries. For Hong Kong, the reason for the insignificance is because Hong Kong is a financial center that mainly specializes in financial services, making it difficult for demand shocks to be transmitted by intra-regional trade activities. 4.2. Size of disturbances The size of disturbances affecting each country is worthy of further investigation. This is because larger shocks result in higher instability of endogenous variables that hinder the feasibility of currency union. The impulse responses for external supply and monetary shocks are assumed to be only minor based on this analyses. Thus, we focus on the remaining three domestic shocks. To compute the size of supply disturbances, given supply shocks have a permanent effect on output, we use the average absolute value of the long-run (20-quarter horizon) effect of a unit shock on real GDP changes. To calculate the size of monetary and demand disturbances, we use the average absolute value of the short-run (2-quarter horizon) effect of a unit shock on CPI and real exchange rate changes, respectively, because the effect of these two shocks are assumed to be transitory. The smaller the size of shocks, the more feasible is the formation of monetary union. These computations are in accordance with those reported by Zhang et al. (2004), Huang and Guo (2006), and Bayoumi and Eichengreen (1993). Table 6 shows the estimated size of domestic supply, monetary and demand shocks. The sizes of the shocks are small compared with those reported by Huang and Guo (2006) who use data from 1970 to 2002. This suggests that the formation of a currency union becomes more feasible with recent economic developments in the region. The speed of adjustment is computed as the response share after the third year of the 20-quarter long-run effect: the faster the adjustment to disturbances, the lower the cost of forming a monetary union. According to Table 6, the speed of adjustment in East Asia is high—that is, these economies adjust rapidly to disturbances. The speed of adjustment has increased in recent years compared with that reported by Zhang et al. (2004) who use data from 1980 to 2000. They identified the size of supply, monetary, and demand shocks to be 0.022, 0.013 and 0.041, respectively, with corresponding speeds of adjustment of 0.995, 0.968, and 0.977. Our findings can also be compared with those of Lee and Koh (2012) who use data from 1970 to 2008. They find the size of supply and demand shocks to be 0.029 and 0.101 respectively, and the corresponding speeds of adjustment to be 0.950 and 1.10. Therefore, the speed of adjustment has increased during the recent years, indicating the feasibility of the currency union formation in the region. That is, based on the OCA literature, countries become likely candidates for a monetary union if their shocks are correlated and small, and if these economies adjust quickly to disturbances (Lee and Koh, 2012; Bayoumi and Eichengreen, 1994; Huang and Guo, 2006).

N. Shafighi, B. Gharleghi / Economic Analysis and Policy 52 (2016) 45–54

51

Fig. 1. Responses of real exchange rates to the structural shocks.

4.3. Impulse response analysis Impulse response functions are applied to assess the effects of structural shocks in real exchange rates. Exchange rates become a less compelling tool if the responses are similar across economies, thus decreasing the cost of currency union formation. Fig. 1 shows the dynamic effect of one structural shock (external supply, external monetary, supply and monetary) on the real exchange rate over a 20-quarter period for each economy. From Fig. 1, even though the path and magnitude of the response varies, it can be concluded that real the rate has an immediate negative response to the external supply shocks in almost all economies except for China and Hong Kong. With regards to the effect of external monetary shocks on the real exchange rate, China, Indonesia, Korea, Malaysia, Singapore and Taiwan tend to associate with an immediate negative response that converges towards equilibrium as the forecast horizon increases. In the case of Hong Kong, Japan, Philippines and Thailand, the real exchange rate associates with a positive response that converges towards the equilibrium in the long-run. Generally, supply shocks provoke a positive response to the real exchange rate in all economies except for Japan and Philippines. The ambiguity of the impact of supply shocks on real exchange rates in different economies is consistent with the theoretical macroeconomic model (Buiter, 1995). Finally, in the case of monetary shocks, real exchange rate responds

52

N. Shafighi, B. Gharleghi / Economic Analysis and Policy 52 (2016) 45–54

Fig. 1. (continued) Table 6 Size and speed of adjustment. Country

Supply shocks

Monetary shocks

Demand shocks

Size

Speed

Size

Speed

Size

Speed

CN HK ID JP KR MY PH SG TH TW

0.011 0.003 0.005 0.003 0.005 0.003 0.003 0.003 0.006 0.004

0.829 1.012 0.739 2.064 1.395 0.913 0.981 1.417 1.194 1.587

0.006 0.004 0.012 0.003 0.004 0.003 0.005 0.005 0.004 0.004

1.00 0.923 0.799 0.070 1.107 0.706 1.118 0.349 3.100 0.566

0.005 0.001 0.043 0.036 0.027 0.019 0.025 0.014 0.021 0.019

0.691 1.961 6.514 1.808 2.341 1.493 2.538 3.848 1.578 2.139

Average

0.004

1.213

0.005

0.973

0.021

2.491

Note: CN: China, HK: Hong Kong, ID: Indonesia, JP: Japan, KR: Korea, MY: Malaysia, PH: Philippine, SG: Singapore, and TH: Thailand, and TW: Taiwan.

positively in Japan, Korea, Malaysia, Philippines, Singapore and Thailand, but negatively in China, Hong Kong, Indonesia and Taiwan. To determine the common exchange rate policy in the region, the shape and magnitude of responses to structural shocks by exchange rate must converge across economies. Therefore, currency union in East Asia can be formed only if the response functions are similar. Fig. 1 suggests that overall patterns of responses are similar only in Indonesia, Japan, Hong Kong, Korea, Malaysia and the Philippines. As a result, the cost of relinquishing their control over the exchange rate should be relatively small. 4.4. Variance decomposition (VDC) VDC is applied to detect the contribution of each shock from the five variables. At the 20-quarter time span, the percentage change variation of the forecast error in exchange rates, prices and real output is decomposed for each shock. Table 7 shows the VDC results for real output, prices and real exchange rates.

N. Shafighi, B. Gharleghi / Economic Analysis and Policy 52 (2016) 45–54

53

Table 7 VDC of the changes in real output, price and exchange rate. Country

External supply shocks

External monetary shocks

Supply shocks

Monetary shocks

Demand shocks

Real output

CN HK ID JP KR MY PH SG TH TW

2.97/19.55 5.20/59.44 4.04/24.56 0.12/1.25 15.07/7.44 30.40/45.40 9.81/36.61 22.95/34.51 2.08/51.92 2.93/9.07

19.61/21.01 8.83/6.26 25.63/31.81 15.64/35.56 19.11/10.90 0.02/14.25 10.04/9.93 3.10/15.79 3.18/2.20 14.56/13.69

77.40/49.60 85.95/31.37 70.31/32.77 84.22/41.47 65.81/60.59 69.56/9.97 80.14/47.20 73.94/43.19 94.73/42.74 82.50/64.19

0.00/6.54 0.00/1.31 0.00/9.63 0.00/8.62 0.00/19.30 0.00/6.26 0.00/4.82 0.00/3.93 0.00/0.05 0.00/12.42

0.00/3.27 0.00/1.60 0.00/1.21 0.00/13.80 0.00/1.75 0.00/24.09 0.00/1.42 0.00/2.55 0.00/3.07 0.00/0.59

Average

9.55/28.97

11.97/16.14

78.45/42.30

0.00/7.28

0.00/5.33

Prices

CN HK ID JP KR MY PH SG TH TW

8.37/47.57 0.99/65.89 1.19/50.43 7.61/8.55 1.96/0.25 0.41/32.85 3.55/68.62 0.00/7.94 10.61/51.80 0.00/22.22

5.65/1.05 11.48/1.19 0.00/4.56 23.82/21.06 34.58/34.27 33.91/25.21 38.02/13.21 1.88/25.69 56.05/8.51 0.82/26.80

0.21/7.39 8.94/6.06 0.58/16.45 1.89/6.59 2.64/32.89 0.01/1.88 0.43/0.07 0.62/46.22 0.01/24.53 30.17/5.59

85.75/40.80 78.57/24.89 98.21/22.53 66.66/54.84 60.80/30.77 65.66/6.98 57.98/17.14 97.48/8.37 33.32/4.60 68.99/12.56

0.00/3.17 0.00/1.94 0.00/6.00 0.00/8.91 0.00/1.80 0.00/33.05 0.00/0.93 0.00/11.75 0.00/10.53 0.00/32.81

Average

3.46/35.61

20.62/16.15

4.55/14.76

70.44/22.34

0.00/11.09

CN HK ID JP KR MY PH SG TH TW Average

0.99/38.32 0.07/1.98 7.60/7.69 1.40/2.80 25.28/18.04 4.97/25.25 3.18/10.83 4.77/9.30 2.69/14.34 15.27/17.64 6.62/14.61

7.73/1.14 19.22/15.38 4.48/8.89 5.39/18.88 11.44/38.50 12.14/7.13 0.00/15.19 16.11/35.47 1.61/1.29 4.90/4.70 8.30/14.65

0.67/5.19 2.25/1.32 0.51/12.67 7.57/5.05 8.31/12.53 0.00/3.81 0.48/0.07 0.40/40.93 12.63/39.00 0.09/8.63 3.29/8.92

17.87/41.20 1.78/2.04 0.22/14.35 0.33/11.03 3.30/6.94 3.69/7.17 5.56/3.09 0.11/6.21 0.01/0.05 0.00/0.83 3. 28/9.29

72.17/14.12 76.65/79.28 87.17/56.38 85.30/62.22 51.65/23.97 79.17/56.61 90.76/70.79 78.60/18.07 83.03/45.29 79.71/68.18 78.42/49.50

Exchange rates

Note: the values indicate the percentage change of the forecast error variance in the real output, prices and exchange rates that is due to the shock at the 1-/20 quarters. CN: China, HK: Hong Kong, ID: Indonesia, JP: Japan, KR: Korea, MY: Malaysia, PH: Philippine, SG: Singapore and TH: Thailand and TW: Taiwan.

Supply shocks account for the major variation in the average real output for all East Asian economies. Compared with Zhang et al. (2004) who investigated the variance decomposition of East Asian economies during the pre-crisis period from 1980 to 1997, we find that the financial crisis has reduced the effect of supply disturbances on real output. Supply shocks accounted for more than 85% and only 78% of variability in real output during the pre- and post-crisis periods, respectively. In the case of price levels, monetary shocks are the predominant disturbances in its variations followed by external monetary shocks. These countries are export-oriented with USA as the major trade partner, Therefore, any changes in the US price level immediately affects the price levels of these economies. Finally, fluctuations in the real exchange rates are predominantly created, on average, by innovations in demand shocks for all economies. Specifically, variations in the exchange rate are generated by monetary shocks for China, external monetary shocks for Hong Kong and external supply shocks for Korea. These findings in exchange rate variations have important policy implications for exchange rate regimes in China, Korea and Hong Kong. 5. Conclusion In this paper, we examine the feasibility of creating a currency union in East Asia. This is due to this region being observed to be moving unalterably closer towards monetary integration as a result of the Chiang Mai Initiative that expanded bilateral currency swap arrangements within the region. A five-variable SVAR model is developed to measure various types of shocks based on the OCA theory. According to this theory countries are plausible candidates for forming a monetary union if they meet the following criteria: (i) their disturbance correlations are small in size, and (ii) they respond similarly to the disturbances and adjust rapidly to the shocks. Overall, our empirical results do not display strong evidence to support the formation of an optimum currency area in the East Asian region. However, some sub-regions are better candidates for a currency arrangement. Results show that Indonesia, Malaysia, Philippines, Japan, Hong Kong and Korea are ready to form a currency union because these economies exhibit the following characteristics: (i) significant and positive correlations of underlying disturbances, (ii) disturbances are of small size, (iii) similar impulse response functions of real exchange rates to the shocks and (iv) rapid adjustment to the shocks. We conclude that some sub-regions are potential candidates for forming

54

N. Shafighi, B. Gharleghi / Economic Analysis and Policy 52 (2016) 45–54

an OCA, which are comparable with the findings of Huang and Guo (2006), Lee and Koh (2012), Nguyen (2010), and Zhang et al. (2004). What are the policy implications from this study? Structural shocks in East Asia are, on average, symmetric with small size. Furthermore, adjustment to shocks in East Asia is faster than in previous periods covered by other researchers. The implication is that, given the disparities, the formation of a currency union has become more feasible for the region during recent years especially after the global financial crisis. Exclusion of China from a monetary union raises a concern. China is a large and growing economy with massive international reserves that can contribute to the stability of the currency union. Hence, its exclusion may result in a major disadvantage to the region. On the basis that monetary union is a desired objective for this region, greater efforts need to be exerted to enhance the business cycle synchronization among China and ASEAN countries. Based on the study by Lee and Azali (2010), a rise in trade integration in the region leads to more synchronized business cycles. Recent implementation of the free trade agreement between China and ASEAN (ACFTA) will therefore increase the integration in trade and lead to greater correlation of business cycles among these countries making them better suited to the formation of a monetary union. Results from the VDC analysis show that for all East Asian countries, supply shocks are the predominant shocks which reflect the variability of average real output. In contrast, monetary shocks are the predominant shocks reflecting the variability of average price levels for all East Asian economies except Thailand. Fluctuations in real exchange rates are predominantly caused by demand shocks in all East Asian countries. Monetary shocks, external monetary shocks, and external supply shocks are the predominant causes of exchange rate variabilities for China, Hong Kong and Korea, respectively. Thus, important policy adjustments are needed in exchange rate regimes in these countries. Acknowledgments We would like to thank Professor Clevo Wilson, the Editor-in-Chief, and the anonymous referees for their constructive comments. References Banerjee, A., Lumsdaine, R.L., Stock, J.H., 1992. Recursive and sequential tests of the unit-root and trend-break hypotheses: theory and international evidence. J. Bus. Econom. Statist. 10 (3), 271–287. Bayoumi, T., Eichengreen, B., 1993. Shocking aspects of European monetary integration. In: Torres, F., Giavazzi, F. (Eds.), Adjustment and Growth in the European Monetary Union. Cambridge University Press, Cambridge, pp. 193–229. Bayoumi, T., Eichengreen, B., 1994. One money or many? Analyzing the prospects for monetary unification in various parts of the world. In: Princeton Studies in International Finance. In: International Finance Section, vol. 16. Princeton University. Bernanke, B., 1986. Alternative explanations of money income correlation. Carnegie–Rochester Conf. Ser. Public Policy 25, 49–100. Blanchard, O., Quah, D., 1989. The dynamic effects of aggregate demand and supply disturbances. Amer. Econ. Rev. 79, 655–673. Buiter, W., 1995. Macroeconomic Policy during a Transition to Monetary Union, Center for Economic Performance Discussion Paper no. 261. Chinn, M.D., 1998. Before the fall: were East Asian currencies overvalued? NBER Working Paper No. 6491. Chow, H.K., Kim, Y., 2003. A common currency peg in East Asia? Perspectives from Western Europe. J. Macroecon. 25, 331–350. Dornbusch, R., 1976. The theory of flexible exchange rate regimes and macroeconomic policy. Scand. J. Econ. 78 (2), 255–275. Goto, J., Hamada, K., 1994. Economic preconditions for Asian regional integration. In: Ito, T., Krueger, A.O. (Eds.), Macroeconomic Linkages. University of Chicago Press, Chicago, pp. 359–388. Hsu, H.F., 2010. Is a common currency area feasible for East Asia? A multivariate structural vector autoregression approach. J. Asian Econ. 24 (4), 391–411. Huang, Y., Guo, F., 2006. Is currency union a feasible option in East Asia? A multivariate structural VAR approach. Res. Int. Bus. Finance 20, 77–94. Ito, T., 1994. On the possibility of a Yen Bloc. In: Glick, R., Hutchison, M. (Eds.), Exchange Rate Policy and Interdependence: Perspectives from the Pacific Basin. Cambridge University Press, Cambridge, pp. 317–343. Kandil, M., Trabelsi, M., 2012. Is the announced monetary union in GCC countries feasible? A multivariate structural VAR approach. Middle East Dev. J. 4 (1), 1–18. Kendall, M., Gibbons, J.D., 1990. Rank Correlation Methods, fifth ed. Oxford University Press, New York. Kwan, C.H., 1998. The theory of optimum currency areas and the possibility of forming a Yen Bloc in Asia. J. Asian Econ. 9 (4), 555–580. Lee, Grace H.Y., Azali, M., 2010. The endogeneity of the optimum currency area criteria in East Asia. Econ. Model. 27 (1), 165–170. Lee, Grace H.Y., Koh, Sharon G.M., 2012. The prospects of a monetary union in East Asia. Econ. Modell. 29, 96–102. Lim, L.K., 2005. A dollar or yen currency union in East Asia. Math. Comput. Simul. 68, 509–518. Ling, H.Y.P., 2001. Optimum currency areas in East Asia. ASEAN Econ. Bull. 18 (2), 206–217. MacKinnon, R.I., 1963. Optimum currency areas. Amer. Econ. Rev. 53, 717–725. Meese, R.A., Rogoff, K., 1983. Empirical exchange rate models of the seventies, do they fit out of sample? J. Int. Econ. 14, 3–24. Mishra, R.K., Sharma, C., 2010. Real exchange rate behavior and optimum currency area in East Asia: Evidence from generalized purchasing power parity. Int. Rev. Financ. Anal. 19, 205–213. Moosavi, R.M., Azali, M., 2014. Monetary integration and optimum currency area in ASEAN+3: What we need for a new framework? Int. J. Econ. Financ. Issues 4 (2), 277–285. Mundell, R.A., 1961. A theory of optimum currency areas. Amer. Econ. Rev. 51, 657–665. Nguyen, T., 2010. An analysis of East Asian currency area: Bayesian dynamic factor model approach. Int. Rev. Appl. Econ. 24 (1), 103–117. Sims, Christopher A., 1986. Are forecasting model usable for policy analysis. Fed. Reserve Bank Minneapolis Q. Rev. 10 (1), 3–16. Soo, S.C., Choong, C.K., 2010. An emergence of a common currency area in the selected East Asian economies: A revisit. Singap. Econ. Rev. 55 (02), 353–376. Takuchi, H., 1994. On the internationalization of the japanese yen. In: Ito, T., Krueger, A.O. (Eds.), Macroeconomic Linkages. University of Chicago Press, Chicago, pp. 335–355. Vogelsang, T.J., Perron, P., 1998. Additional tests for a unit root allowing for a break in the trend function at an unknown time. Internat. Econom. Rev. 39, 1073–1100. Xu, X., 2006. A currency union for Hong Kong and Mainland China? J. Int. Money Finance 25 (6), 894–911. Zhang, Z., Sato, K., McAleer, M., 2004. Asian monetary integration: a structural VAR approach. Math. Comput. Simulation 64, 447–458.