Jump to ContentJump to Main Navigation
Prospects for Monetary Cooperation and Integration in East Asia$

Ulrich Volz

Print publication date: 2010

Print ISBN-13: 9780262013994

Published to MIT Press Scholarship Online: August 2013

DOI: 10.7551/mitpress/9780262013994.001.0001

Show Summary Details
Page of

PRINTED FROM MIT PRESS SCHOLARSHIP ONLINE (www.mitpress.universitypressscholarship.com). (c) Copyright The MIT Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in MITSO for personal use (for details see www.mitpress.universitypressscholarship.com/page/privacy-policy). Subscriber: null; date: 16 January 2019

A Reconsideration of Costs and Benefits

A Reconsideration of Costs and Benefits

(p.103) 6 A Reconsideration of Costs and Benefits
Prospects for Monetary Cooperation and Integration in East Asia

Ulrich Volz

The MIT Press

Abstract and Keywords

This chapter discusses the changes that monetary integration might bring about, as well as the costs that it might involve. It presents a critique of optimum currency area (OCA) theory, followed by empirical and theoretical analyses of the dynamic effects and potential benefits of monetary integration in East Asia.

Keywords:   optimum currency area theory, monetary integration, East Asia

So far the analysis was based in the traditional, mostly static, OCA framework. While OCA theory has to offer important insights, some of its assumptions are misleading or even wrong.1 Moreover OCA theory fails to take into account that monetary integration is an endogenous process in which the criteria that have been developed in the OCA literature in order to assess the desirability of monetary integration change as a result of a policy of integration. Acknowledging these weaknesses of OCA theory and taking a dynamic perspective can turn some of the wisdoms of OCA theory on their head and lead to different conclusions regarding the attractiveness of fixed exchange rate regimes or monetary union.

This chapter will therefore go beyond the conventional, static approach and will discuss the changes that monetary integration might bring about, as well as the costs that it might involve. The next section will develop a critique of OCA theory, which will be followed by empirical and theoretical analyses of the dynamic effects and potential benefits of monetary integration in East Asia in the sections after.

6.1 Critique of the Optimum Currency Area Theory

When I began to work on monetary unions a few years ago it was rare for academics outside Europe to be interested in the subject, and highly unusual to find advocates of monetary union outside a few continental Europeans. Now it is almost the majority view.… Currently there seems to be an emerging consensus in favour of monetary union, at least for many small open economies. What are its roots? At a broad level, there are two: (i) the benefits of floating exchange rates have been over-stated; and (ii) the benefits of monetary union have been understated. (Rose 2001: 193)

Recent developments in economic theory and research on monetary integration have pointed to some important theoretical flaws and weaknesses of the standard OCA theory. Taken seriously, these require a fundamental reconsideration of the costs and benefits of monetary integration.2 Three insights are of particular importance. First, it has become clear that, in the face of a vertical long-run Phillips curve, countries cannot use flexible exchange rates to realize a chosen combination of inflation and unemployment, and that (p.104) fine-tuning the real economy through national monetary policy is impossible. Second, modern exchange rate theories tell us that flexible exchange rates are not determined solely or even predominantly by economic fundamentals, which implies that the effectiveness of national exchange rate policy is limited. Third, the Lucas critique has left important marks in the recent literature and challenged the static view of OCA theory. The criticism of OCA theory, together with recent empirical findings, put monetary integration in a much better picture than was previously the case. Building on these insights, the sections after will present East Asia–specific evidence on the prospective costs and benefits of monetary integration.

The “Death of the Phillips Curve”

As noted earlier, Mundell (1961) believed that flexible exchange rates would allow a nation to pursue independent monetary and fiscal policies so as to successfully manipulate aggregate demand to offset private-sector shocks on the supply or demand sides. This understanding of what monetary policy can deliver takes for granted a stable long-term Phillips curve, which is a proposition that governed the theoretical and practical understanding of economic policy-making in the 1960s and 1970s.3 The Phillips curve displays a positive relation between inflation and employment and was interpreted at that time as an option for government authorities to increase employment by pursuing expansionary policy. That is, monetary policy was regarded as an instrument to administer a trade-off between inflation and unemployment. In face of such a stable, inverse relation between inflation and unemployment, it would be possible to use exchange rate policy to realize different combinations of inflation and unemployment. Accordingly, a policy of sustained depreciation of the nominal exchange rate could bring about a sustained weakening of the real exchange rate, a lasting improvement in international competitiveness and the current account, and a permanently lower level of unemployment. The loss of independent monetary and exchange rate policy would therefore mean the loss of one of a country’s most important economic policy tools. Buiter (2000: 245) has termed this common assertion “that monetary policy can be used systematically and effectively to dampen the effect on the real economy of external and/or internal shocks … the fine tuning fallacy.”

This view of a permanent trade-off between inflation and unemployment and a country’s ability to choose an optimum point along its Phillips curve has become untenable. The failure was “to distinguish in a consistent way between short-term nominal rigidities and long-term real rigidities,” a proposition that has led to a “serious overestimation of the power of monetary policy … to influence real economic behaviour” (Buiter 2000: 222).

As was made clear by Phelps (1967) and Friedman (1968), a lowering of unemployment through higher inflation is only possible in the short run and cannot be repeated arbitrarily. The assumed Phillips curve relationship is only valid in the presence of static inflationary expectations or if economic agents live under money illusion, namely if they think in terms of nominal instead of real income. If, instead, price increases are anticipated in wage (p.105) bargaining, there is no—not even a short-run—trade-off between inflation and unemployment. The long-run Phillips curve is therefore a vertical over the “natural rate of unemployment” (Friedman 1968: 8), namely different inflation rates correspond to the same level of unemployment.

If the long-run Phillips curve is vertical and there is no hysteresis in the natural rate of unemployment, temporary real shocks will have only temporary real effects (Buiter 2000). Whether there is a short-term trade-off depends on nominal rigidities, that is, on expectation formation and the behavior of labor unions and employers. Nominal exchange rate depreciations through expansionary policy therefore have at most a temporary effect on the real exchange rate and on international competitiveness. The transitory real effects will become smaller and shorter-lived if nominal exchange rate depreciations become a frequently and predictably used instrument to boost international competitive advantage.

The “death of the Phillips curve,” as Niskanen (2002) has called it, implies that monetary and exchange rate policy will have no significant and permanent impact on the path of an economy’s capacity output.4 This is not to deny that monetary policy can have powerful real effects, but these are only transitory. At the long sight, monetary policy can best contribute to macroeconomic performance and stability by anchoring medium and long-term inflationary expectations and thereby eliminating an important source of uncertainty affecting household and business decisions. To have stabilization ambitions for monetary and exchange rate policy much beyond this, “is likely to lead to greater volatility and instability in the real economy” (Buiter 2000: 246).5

An Understanding of the Exchange Rate as an Asset Price

The second fundamental shortcoming of OCA theory is its failure to allow properly for the implications of international capital mobility and the associated disruptive potential of exchange rate flexibility. This has led to an overemphasis on the stabilizing potential of market-determined nominal exchange rates, and a failure to recognize its destabilizing potential (Buiter 2000).

Starting point of the whole OCA framework is the assumption that the members of a currency area lose the exchange rate as an important adjustment instrument. This is based on a view of the exchange rate originating in the flow model of exchange rate determination, according to which the exchange rate provides a speedy and effective adjustment for external disequilibria. Exchange rates are thought to be influenced primarily by changes in economic fundamentals relating to the current account. Exchange rate theory has long parted from such simplistic assumptions.6 Modern exchange rate theories describe the exchange rate as an asset market variable, which is not only influenced by transactions in the current account but also—or maybe even dominantly—by financial flows and expectations.7

The portfolio balance approach to exchange rate determination emphasizes changes in the relative supplies of assets denominated in different currencies as a fundamental cause (p.106) of exchange rate movements.8 A central assumption of the portfolio balance model is the imperfect substitutability between domestic and foreign assets. This marks a departure from the proposition of identical substitution elasticities of currencies (an implicit assumption of PPP and interest rate parity theories), which rests on the reduction of money to its role as unit of account and means of transaction in the present (Hauskrecht 2001). But if the functions of money as means of payment for transactions in the future and its function as store of value are considered, this assumption becomes untenable.9 Currencies differ in their quality as store of value and have different nonpecuniary rates of return. The different valuations of currencies depend, among other factors, on the historical development of nominal exchange rates, a country’s international net asset position, the size of the currency area and the investment opportunities therein, and expectations about the currency’s future value.

With the exchange rate driven predominantly by asset market developments that are not only based on economic fundamentals and rational behavior, the exchange rate cannot be regarded as a reliable instrument for adjustment to real economic shocks—and hence it cannot be lost as a policy instrument. Quite the opposite, it might itself become a source of instability and distort production, employment, investment, and consumption (Schelkle 2001a).

There are multiple examples of (irrational) exuberance or pessimism in financial markets, such as speculative bubbles, collective mood swings, herd behavior, bandwagon effects, noise trading, panic trading, or trading by agents caught in liquidity shortage in other financial markets that can lead to excessive volatility and persistent misalignments in the foreign exchange markets (Buiter 2000). Other destabilizing phenomena are overshooting exchange rates or self-fulfilling prophecies and speculative attacks that can lead to currency crises. Hence, at times, exchange rate movements may not only not stabilize but need to be stabilized in order to avoid disturbances in the real economy. So use of other economic policy instruments may be required to stabilize the exchange rate, reducing instead of increasing domestic policy autonomy. For economies that are strongly affected by such phenomena, in the extreme case, abandoning the exchange rate could therefore mean a benefit and a gain in degrees of freedom of economic policy-making. As Schelkle (2001b: 27) maintains: “taking the exchange rate as an asset price seriously turns the basic message of the OCA approach on its head.”

The potential advantages of exchange rate flexibility as a short-term adjustment mechanism or shock absorber therefore need to be weighted against the potential destabilizing role of the exchange rate as a source of shocks and instability. Arguably the globalization of capital and financial markets has increasingly impaired the effectiveness of monetary and exchange rate policy as stabilization instruments.10 In the view of Cooper (1999: 112), “flexible exchange rates will gradually evolve from being mainly a useful shock absorber for real shocks into being mainly a disturbing transmitter of financial shocks, increasingly troublesome for productive economic activity. Thus a cost–benefit calculation for flexible (p.107) versus fixed rates will gradually alter the balance against flexibility, even for large countries.” Rose (2001: 194) also doubts the value of the exchange rate as a useful economic policy instrument:

Floating exchange rates are said to provide insulation, and to be an additional tool of monetary policy. In practice, they just as often introduce shocks that have to be offset through other tools of economic policy. Rather than being part of the solution, they are frequently part of the problem. That’s why so many countries seem to have a “fear of floating” in the memorable phrase of Calvo and Reinhart. […] eliminating exchange rate volatility seems almost to be a free lunch, […]. As a result, thinking about the exchange rate as an extra tool for macro-management is starting to seem unworldly. There are exceptions of course […]. But those cases are…exceptions.

The Lucas Critique

The Lucas critique and the time-inconsistency literature pose another challenge to the quintessentially static approach of OCA theory.11 Lucas’s (1976) critique of the then usual way of (econometrically) modeling supposedly structural macroeconomic relations as functions of past policies made clear that such relations cannot be used in simulations designed to predict the effect of a different policy regime. This is because shifts in economic policy often produce a completely different outcome as agents adapt their expectations to the new policy stance.12 The behavior of economic agents and henceforth also economic structures will respond to economic policy changes. In other words, “[t]he structure of an economy is endogenous to the economic policies applied to it” (Schelkle 2001b: 18).13

OCA theory has generally treated the above-discussed criteria as exogenous variables. Yet it is obvious that if economic structures change with economic policies, then the OCA criteria become endogenous instead (Schelkle 2001a). If the economic environment changes in the process of monetary integration, then the convergence of certain parameters need not necessarily be a precondition for successful monetary integration, but rather change could be the desired result. As noted earlier, going back to Fleming (1971), especially the convergence of inflation rates is regarded as precondition for joining a currency area since this is a long-term condition for balanced national accounts within that area. The time-inconsistency argument, however, reverses the order between what can be regarded as precondition and what can be seen as the desired outcome of monetary integration. As an application of the Lucas critique, inflation convergence can be considered as an intended result of monetary integration between countries with different historical inflation histories (Gandolfo 1992; Tavlas 1993b).

Recent contributions have particularly discussed how a stabilization of exchange rates can induce changes in the labor and factor markets or changes in the monetary policy regime that suppress inflationary wage and price setting and that bring about a convergence of inflation rates.14 The time inconsistency literature was able to show that monetary credibility of a country with high inflation rises through the entry into a currency area because policy makers get their “hands tied.” The strategy of tying one’s hands (Giavazzi and Pagano 1988) (p.108) aims at the import of stability, since an exchange rate target (as discussed earlier) requires the subordination of national economic policies and currency devaluations can no longer be used to compensate for inflationary price and wage policies. The assumption is that national governments would not be able to see through such a policy course on their own.

Schelkle (2001a) maintains that especially structural inflation might only be come by with a policy of monetary integration.15 Indeed structural inflation might have been a main reason for Italy’s membership in the EMS, even though it meant a complete departure from past economic policies (see Giavazzi and Pagano 1988).16 The convergence of interest rates was also one of the Maastricht criteria, and the convergence that was achieved illustrates the structural change in national economic policies that was made possible by the political decision to join EMU. In this light, the monetary convergence that was achieved can be regarded as an endogenous result of the integration process, not as a condition that had to be fulfilled ex ante (Tavlas 1993; Schelkle 2001b).

The endogenous change of OCA parameters during the process of monetary integration, however, need not necessarily lead to an optimization of the currency area. The conclusion whether integration will lead to an optimalization or suboptimalization can be ambivalent for single criteria (Schelkle 2001a). As will be discussed later with respect to the effects of real and financial integration on the convergence or divergence of output fluctuations, various transmission channels might exist that work into opposite directions.

Summing up, the Lucas critique makes clear that it is misleading to require certain conditions, which need to be viewed as the desired outcome of monetary integration, to be fulfilled ex ante. The Lucas critique, however, does not invalidate the OCA analysis on East Asia conducted in section 5.3. But it is crucial to be aware that the structures and characteristics of the East Asian economies analyzed there might be subject to change. To fully understand the costs and benefits of monetary integration, it is thus important to overcome the static OCA perspective and also analyze the changes that will be induced by a policy of monetary integration. But the focus should not only be on allocative effects—the development context of monetary integration and the potential accumulative effects also need to be brought to attention. In particular, it is important to consider if and how a strategy of monetary integration can help overcome structures that put a constraint to economic development (Roy and Betz 2000). In this respect the analysis of monetary integration, particularly between developing and emerging countries as in East Asia, also needs to address how it can help create favorable macroeconomic conditions that are conducive to financial market development and to overall economic growth and development.

Building on these insights, the following sections will go beyond the conventional arguments and attempt at analyzing the effects and potential benefits of monetary integration in East Asia. Four different aspects of monetary integration will be analyzed. First, section 6.2 turns to a discussion of the relationship between monetary integration and trade and estimate the trade-creating effects of monetary integration in East Asia. Second comes (p.109) an empirical investigation of the effects of real and financial integration on output fluctuations in East Asia in section 6.3. This is basically a practical application of the Lucas critique and an attempt to analyze the endogeneity of one of the most important OCA criteria, namely the symmetry of output fluctuations among countries of a currency area. Third, section 6.4 draws attention to the potential role of regional monetary integration in developing regional capital markets and in overcoming the problems of asset and liability dol-larization. Fourth, and finally, section 6.5 extends the discussion of the alleged costs of monetary integration and shows that, instead of losing economic policy autonomy, monetary integration might even be a way of (re)gaining some degrees of monetary independence in East Asia.

6.2 Monetary Integration and Trade

As was discussed in section 5.3, over the past few decades, the share of intraregional trade in total trade has reached high levels in East Asia, and the region is set to increase these trading links even more. In addition to trade linkages, intraregional interdependence in FDI has also increased dramatically (Kawai and Urata 2004). Multinational corporations—often from Japan and Korea, but also from Hong Kong, Singapore, the United States, the European Union, and since recently also China—have developed extensive regional production networks in East Asia.17

For a region as economically intertwined as East Asia, exchange rate policy spillover effects from one country to another are of great importance. There are numerous problems faced by countries that are close trading partners but follow different exchange rate regimes.18 First, exchange rate upheavals could lead to reduced exports from the country that loses competitiveness to its partners. This could evoke increased protectionism and even a scaling back or elimination of trade agreements. The country that loses competitiveness as a result of a real exchange rate appreciation vis-à-vis its trading partners may employ anti-dumping or other administrative measures (if tariffs are precluded through trade agreements) to protect domestic firms. The protectionist response may trigger a trade war as well as a round of beggar-thy-neighbor devaluations. Second, investments may be relocated because of severe exchange rate disagreements. Regional trade agreements may spark fierce competition for the location of investment, and swings in the bilateral exchange rates may have important consequences for the location of new investments and might even shift the location of existing investments. Third, and finally, a change of the exchange regime in one of the partner countries may cause an exchange rate crisis in the other. Exchange rate depreciation in one country may reduce the credibility of the partner’s commitment to a fixed parity and generate speculative attacks on its currency. A country may thus be forced to abandon its preferred exchange rate policy due to the exchange rate disagreement. All this, in turn, could have repercussions for the country that sparked the crisis (see chapter 3).

(p.110) Hence the deeper trade relations become and the more comprehensive trade and investment agreements are the more important the question of macroeconomic policy coordination becomes. Increasing intraregional trade turns formerly second-order effects into issues that have to be addressed with first-order preference. In highly integrated areas policy coordination to achieve real exchange rate consistency is therefore essential.

Exchange rate cooperation becomes even more important if countries also compete against one another in third markets, as is the case in East Asia. Hence McKinnon (2005: 5) describes mutual exchange rate stability in East Asia as the “quintessential public good.” Because of neighborhood effects, he reasons, a nation’s decision to fix or float should not be made independently of other nations’ decisions.

While these arguments favor exchange rate coordination over flexible exchange rates, there is no implicit logic for monetary unification. Eichengreen (1998) points out that the need for monetary integration depends on the degree of trade integration that is to be achieved. A free trade area or customs union can be sustained despite the existence of separate national currencies that fluctuate against one another.19 But deeper integration, including the free movement of goods, services, capital, and people, implies an even greater degree of openness of domestic markets and more intense cross-border competition, making exchange rate changes more disruptive. The European Union has always given great importance to this issue. Indeed, the EU reasoning for monetary union can be described as “one market, one money” (Emerson et al. 1992). If East Asian leaders want to foster deeper integration akin to Europe’s single market, they will need to contemplate far-reaching monetary integration like in Europe.

More recently the potential trade-creating effect of a common currency union has gained wide attention in the literature, reinforcing the trade argument for monetary integration. Rose (2000) turned the EU logic upside down and established the case for “one money, one market.”20

While it is impossible to accurately predict the trade-creating effect of a potential monetary union in East Asia, we can estimate the effect that the similarity of currency regimes had on bilateral trade in the region. To do this, we apply the following gravity model to the trade data from thirteen East Asian countries over the period 1980 to 2003:21

T ijt = β 0 + β 1 F ijt + β 2 D ijt + β 3 I 1ijt + ε ijt ,

where Tijt is an indicator for the level of bilateral trade integration between countries i and j at time t, Fijt is an indicator for the intensity of bilateral FDI, and Dijt is an indicator for the similarity of currency regimes of the two countries in question. I1ijt is a vector comprising the standard variables of the gravity model, and εijt is a well-behaved error term.

The gravity model has become standard in the international trade literature, and hence one can draw on a rich body that employs the gravity model.22 We include the standard variables of the gravity model, namely the product of the natural logarithms of the populations of the two countries under investigation, a variable describing their combined GDP (p.111) computed as the product of the natural logarithms of each country’s GDP, the natural logarithm of the distance between the two countries, and a dummy variable indicating whether the countries share a common land border.23

As in Frankel and Rose (1998), bilateral trade integration is measured by computing

T ijt = x ijt + m ijt X it + M it + X jt + M jt ,

where xijt are the exports from country i to j during period t, mijt are the imports to country i from j, and Xit, Mit, Xjt, Mjt are the total exports/imports of country i/j respectively. A higher value of Tijt thus represents a greater trade intensity between countries i and j. As is common in the literature, we take the natural logarithm of the ratio. The data are from the IMF’s Direction of Trade Statistics (DTS). Data for Taiwan are from the Taiwan Statistical Data Book 2004.

Similarly FDI intensity is measured as

F ijt = fd i ijt +fd i jit GD P i +GD P j ,

where fdiijt denotes FDI to country i from country j and fdijit denotes FDI to country j from i. The bilateral FDI flows are divided by the joint GDP of the two countries. As with the trade measure, Fijt takes a higher value the more integrated the countries are. The bilateral FDI data are taken from the UNCTAD Foreign Direct Investment Statistics and CEIC Asia Database. The GDP data are from IFS and the Taiwan Statistical Data Book 2004.

To measure the similarity of the currency regimes of the two countries in question, hypothetical currency baskets are estimated in order to compute

D ij =1| w it w jt |,

where wit is the weight of the US dollar in a hypothetical currency basket of country i. To estimate the weights of the currencies included in the hypothetical currency baskets against which East Asian countries manage their currencies, we follow Frankel and Wei (1994) and regress each East Asian currency e on a constant c, the US dollar, the euro, and the Japanese yen:

Δln ( e CHF ) t =c+ b 1 Δln ( USD CHF ) t + b 2 Δln ( EUR CHF ) t + b 3 Δln ( JPY CHF ) t + ε t .

The Swiss franc, which can be assumed to be uncorrelated with the three basket currencies as well as with the East Asian currencies, is used as numéraire in order to minimize multicolinearity problems. The β coefficients are the weights of the basket. Δ stands for the first-difference operator and’t for time. All variables are in natural logarithms. b is used to compute the above indicator (Dij); that is, b1 = wit. If both countries give the same (p.112) weight to the dollar in their (hypothetical) currency basket, Dij is one; if one of the countries chooses a hard fix to the dollar (b1 = 1) while the other one chooses a zero dollar weight in its currency basket (b1 = 0), Dij is zero.24

The results presented in table 6.1 show a strong trade-creating effect of similar currency regimes in East Asia. The coefficients for the similarity of the currency regime are large and significant. The trade-creating effect gets smaller if China is excluded from the regression, which can be explained by a “Japan effect”: Japan is a driving force in intraregional trade but is the only country in the region that does not adhere to the informal “East Asian dollar standard” (McKinnon 2001, 2005). If Japan is omitted from the regression, the currency effect becomes even larger and highly significant. That is, exchange rate stability is beneficial for trade within East Asia.25

The coefficients of the other variables are in line with theoretical expectations. Strong FDI links, a shared land border, and larger GDP all increase bilateral trade whereas distance and population have the inverse effect.26 Given the previous findings in the literature that monetary union has an even larger effect on trade than a hard peg, it is also sensible to expect a significant trade-creating effect of monetary unification in East Asia. In his seminal study, Rose (2000) estimates that two countries that use the same currency trade three times as much than they would with separate currencies.27 While a reduction of

Table 6.1 Determinants of trade in East Asia

Full sample

Without China

Without Japan

Without China and Japan










Similarity of currency regime


















Common border




































Number of observations





R squared





Note: *** denotes statistical significance at the 1 percent level, ** at the 5 percent level, and * at the 10 percent level. Standard errors are in parentheses.

(p.113) exchange rate volatility also has a positive effect on trade, Rose assumes the trade creating effect of a monetary union to be much larger than of a currency area with fixed rates but separate currencies.28 Rose and van Wincoop (2001) even speak of national currencies as significant barriers to international trade.

While there has been a lot of controversy about the actual magnitude of the trade-creating effect of monetary union, a broad agreement has emerged that a currency union has a significant positive effect on trade.29 Even though a doubling or tripling of trade as a result of monetary unification deems unreasonably high, these estimates become more plausible when compared with the home market bias in international trade. Rose and Engel (2002) find that members of a currency union are more integrated than countries with separate currencies, but they also find that trade integration is much lower than within countries. Like Rose and Engel, several other studies have revealed that intra-national trade (i.e., trade between regions within a nation) is much larger than between countries. McCallum (1995) shows that trade between Canadian provinces in 1988 was 20 times higher on average than between Canadian provinces and US states. This is remarkable as there are no language, geographic, or tariff barriers to trade. The studies by Helliwell (1998) on trade between Canada and the United States and by Nitsch (2000) on European trade estimate the “home bias” to be around a factor of ten. One possible explanation for this phenomenon is the use of a common currency (Frankel and Rose 2002).

An increase in trade is likely to affect economic growth. To estimate the implications of a common currency on trade and income, Frankel and Rose (2002) proceed in two steps.30 First, they estimate the effect of a common currency on trade, which (as in Rose 2000) leads to a tripling of trade, without any indication of trade diversion. In the second stage, they estimate that an increase in trade by one percent leads to an increase in per capita income by at least a third of a percent over a period of twenty years. They then combine these two results to estimate the effect of a common currency on income. According to their results, EMU membership of a country like Poland, which conducts half of its total trade with the euro area, would lead to an increase of per capita income of 20 percent.31

Indeed, a variety of factors influence international trade, and the choice of exchange rate regime is only one of them. A common language, cultural similarities, a common history, colonial linkages, geographic proximity and shared borders, membership in a free trade area, legal systems, physical infrastructure, and others factors might play a role in explaining trade relations.32 The discussion on the trade-creating effect of monetary integration should therefore not be misunderstood as denouncing these influences. For sure, the development of physical infrastructure such as roads, bridges, telecommunication, and the like can have an important stimulus on trade between neighboring countries. Particularly in less developed regions, such as the Mekong delta that connects Lao, Cambodia, and Vietnam, such measures will probably have a much greater effect than monetary integration. Nevertheless, the influence of monetary integration, and especially monetary (p.114) unification, seems to be substantial. Frankel and Rose (2002) even believe it to be as big as the trade-creating effect of a FTA.

6.3 Effects of Real and Financial Integration on East Asian Business Cycles

As discussed in section 5.2, a strong output correlation between potential members of a common currency area is a standard criterion of OCA theory. Traditionally analysis of output cycles was limited to an assessment of the historical record as in section 5.3. More recently, however, the endogenous nature of business cycle correlations has become an important topic of research (basically as an application of the Lucas critique), raising doubts about the explanatory power of a simple ex ante analysis of business cycle correlations when assessing the suitability of countries to form a currency area (Frankel and Rose 1998).

Real and financial integration are both likely to be affected by monetary integration, which may also result in a change in national business cycles (see figure 6.1). The direction of these effects, however, is not clear a priori. Over the past years, particularly the impact of an increase in trade integration on business cycles has received great attention following Frankel and Rose’s (1998) seminal paper on the endogeneity of the OCA criteria. Nevertheless, the effects of real and financial integration on output fluctuations have not yet been investigated in a systematic way in the East Asian context, despite its great importance for the discussions about East Asian monetary cooperation and integration.

The analysis in this section aims to fill this gap by building on recent contributions to the literature by Imbs (2004, 2006). A system of simultaneous equations is estimated to analyze whether and how trade and financial linkages influence business cycle

A Reconsideration of Costs and Benefits

Figure 6.1 Effects of monetary integration on business cycles

(p.115) synchronization in East Asia, both directly and indirectly. We will first review the theoretical literature and previous empirical findings and then explain the model used for our estimations.

Theoretical Literature and Previous Empirical Findings

From a theoretical angle, the effects of trade and financial integration on business cycles are ambiguous. On one hand, increased intra-industry trade deriving from economies of scale and product differentiation should lead to more synchronized business cycles, a view taken by the European Commission (Emerson et al. 1992) in support of European monetary integration. On the other hand, if classic Ricardian or Heckscher–Ohlin trade prevails, increased trade could result in more specialization of industries, and therefore make countries or regions more vulnerable to idiosyncratic, industry-specific shocks (Kenen 1969; Krugman 1993).

Similarly financial integration could have a divergent effect on business cycles by fostering industrial specialization through the provision of risk insurance across countries/regions and through the reallocation of capital according to countries’ comparative advantage. As financial integration is also likely to have an effect on trade integration, through the provision of trade credit, it could amplify the convergent or divergent effect of trade on output fluctuations (see figure 6.2). Moreover financial linkages could positively affect output co-movement by generating demand-side effects, namely if consumers from different countries have financial investments in each other’s countries, they will all enjoy (suffer) a boom (bust) in any of the countries (see section 5.2).

As theory remains ambiguous, the proof of the pudding has to be found in empirical analysis. Following Frankel and Rose (1998), a burgeoning literature on the impact of an increase in trade integration on business cycles has developed. In their paper

A Reconsideration of Costs and Benefits

Figure 6.2 Effects of financial integration on business cycles

(p.116) Frankel and Rose use a panel of bilateral trade and business cycle data spanning twenty industrialized countries over thirty years and find that closer international trade links result in more closely correlated business cycles across countries. Studies by Flandreau and Maurel (2005), Gruben, Koo, and Millis (2003), and others have confirmed this positive relationship. Imbs (2004) also finds the overall effect of trade on business cycle synchronization to be strong, and resulting primarily from intra-industry trade, as predicted by theory.

The empirical results on the link between financial integration and business cycles are less clear to date. One strand of literature finds a divergent effect of financial integration on business cycles. Heathcote and Perri (2004) and Kalemli-Ozcan, Sorensen, and Yosha (2001, 2003), for instance, find that income insurance through financial integration fosters industrial specialization and thus leads to more asymmetric business cycles. On the contrary, Claessens, Dornbusch, and Park (2001) and Imbs (2004, 2006) find that financial integration leads to a synchronization of business cycles.

The studies mentioned so far have all focused on industrial countries. However, a few studies have scrutinized East Asian business cycles as well. Kim, Kose, and Plummer (2003) analyze the business cycles of seven East Asian countries between 1960 and 1996 and compare them with the business cycles of G-7 countries. They find signs of an East Asian business cycle, namely high correlations between individual countries and a constructed East Asian cycle, but they do not analyze the factors that influence the cycles. McKinnon and Schnabl (2003) focus on fluctuations of the yen–dollar exchange rate as a major force affecting the business cycles of the smaller East Asian countries. Moneta and Rüffer (2006) estimate a dynamic common factor model for output growth of ten East Asian countries and find that all countries except China and Japan share a significant common component factor and that a number of external factors, including the yen/dollar rate and the price of oil, play an important role in synchronizing output cycles. Socorro Gochoco-Bautista (2008) finds that the global electronics cycle is a common regional factor that drives industrial output growth within the region.33

A few papers have analyzed the effect of trade integration on business cycles in East Asia using simple ordinary least squares (OLS) regressions. Shin and Wang (2003) analyze the business cycles of twelve Asian countries between 1976 and 1997 and find hints that intra-industry trade in East Asia contributes to a convergence of cycles. Using a sample of ten East Asian countries for 1981 to 1995, Choe (2001) finds a positive effect of trade on business cycle correlations. Crosby and Voss (2002) regress business cycles on trade intensity and some macroeconomic variables for thirteen countries for 1980 to 1999 without finding any significant impact of trade on output fluctuations.

The only study so far, to our knowledge, that has also looked at the effects of financial integration on output co-movements in East Asian is by Shin and Sohn (2006), who consider a sample of nine East Asian countries for the period 1971 to 1999/2003. While these authors find that greater trade enhances co-movements of output, they cannot detect any (p.117) significant effect of financial integration, a result they attribute to the presently low level of financial integration in the region.34

Methodology and Data

Most model specifications of the above-mentioned studies (including Frankel and Rose 1998) take a similar form to

ρ ijt = α 0 + α 1 T ijt + α 2 F ijt + α 3 S ijt +α C ijt + ε ijt ,

where ρijt denotes bilateral GDP correlations between countries i’ and j, Tijt stands for bilateral trade integration, Fijt for bilateral financial integration, Sijt for specialization, Cijt for some other control variables, and εijt for an error term.

Such single equation models only estimate reduced effects of trade, financial integration, specialization, and so forth, but do not distinguish between direct and indirect effects. The existence of indirect effects in different directions may cause the direct effects to appear small (as they may cancel each other out) in single equation regressions. Furthermore endogeneity is an important concern with such regressions if there are several variables on the right-hand side of the equation; a reason why some researchers have chosen to employ instrumental variables.

To tackle the endogeneity problem and to account for the various indirect effects, such as the effects of financial integration on business cycles through trade and specialization, we choose a simultaneous equations approach, following Imbs (2004, 2006). The specification used here takes the following form:

ρ ijt = α 0 + α 1 T ijt + α 2 F ijt + α 3 S ijt + α 4 D ijt + ε ijt 1 ,
T ijt = β 0 + β 1 F ijt + β 2 D ijt + β 3 I 1ijt + ε ijt 2 ,
F ijt = γ 0 + γ 1 T ijt + γ 2 I 2ijt + ε ijt 3 ,
S ijt = δ 0 + δ 1 T ijt + δ 2 F ijt + δ 3 I 3ijt + ε ijt 4 ,

where Dijt is the similarity of currency regimes of the two countries in question, I1ijt the standard variables of gravity model (distance, common language and border, etc.), I2ijt the institutional variables describing the quality of governance, and I3ijt the stage of development of respective countries.

Panel data on thirteen East Asian countries over a time period ranging from 1990 to 2003 are used. The countries are Cambodia, China, Hong Kong, Indonesia, Japan, Korea, Lao, Malaysia, the Philippines, Singapore, Taiwan, Thailand, and Vietnam.35 The sample is divided into two subsample periods to capture the change over the years: 1990 to 1996 and 1999 to 2003. The crisis years of 1997 and 1998 are excluded.

For all variables, bilateral data are used, as this is argued to be more accurate than aggregate data as used in most studies. While aggregate data are easier to collect, their (p.118) use is likely to spoil the empirical results because they describe the general openness of an economy with respect to trade, finance, and the like, rather than integration between two economies. The different variables are constructed as follows.

Bilateral GDP Correlations

GDP correlations are measured as

ρ ijt =corr ( Y i , Y j ) t

where Yt and Yj are the measures of economic activity of the respective countries, for which we use GDP at constant local prices. Before taking correlations, the data are detrended using first-order differencing.36 All data are annual and are from WDI and the Taiwan Statistical Data Book 2004.

Bilateral Trade Integration

For bilateral trade integration we can employ the same trade measure already used in the analysis of the trade-creating effects of monetary integration in section 6.2. That is, bilateral trade integration is measured by

T ijt = x ijt + m ijt X it + M it + X jt + M jt ,

where xijt are the exports from country i to j during period t, mijt are the imports to country i from j, and Xit, Mit, Xjt, Mjt are the total exports/imports of country i/j respectively. The data are from the same sources as in section 6.2.

Bilateral Financial Integration

A more difficult task, which is indeed crucial to this analysis, is to measure financial integration. There are various ways of defining and measuring financial integration. The direct measurement of international financial markets integration has proven, so far, to be an extremely arduous task plagued by ambiguity. This ambiguity results because there is no single mechanism to measure the degree of integration that is free of conceptual and technical difficulties and that is exempted from interpretation problems. The acknowledgment of these difficulties justifies why the literature has developed without pointing to a single measure but rather a set of measures. Financial integration clearly has to be seen as a multi-facetted phenomenon which cannot be captured by any one measure.

The choice of measures that could be used for measuring financial integration in East Asia, however, is restricted due to the characteristics of the financial markets of the region and to the related issue of data availability. As discussed in section 5.3, security markets are weakly developed in the majority of East Asian countries, precluding the use of any measure that makes use of bond or equity indicators. Interest rate parity conditions could be used (as done by De Brouwer 1999), but money markets are also hardly developed in the poorer countries in question.37 This leaves the banking sector and FDI.38 While cross-border (p.119) banking activities would make a potentially useful indicator, there are no data available, effectively reducing the choice to FDI as a measure of financial integration.39 Choosing FDI, however, also makes sense given the structure of financial linkages of the region, in which it plays a pre-eminent role. To measure financial integration, we thus employ the same FDI measure from section 6.2:

F ijt = fd i ijt +fd i jit GD P i +GD P j ,

where fdiijt denotes FDI to country i from country j and fdijtt, denotes FDI to country j from i. The bilateral FDI flows are divided by the joint GDP of both countries. Again, the data are from the same sources cited in section 6.2.


To measure specialization, we follow Imbs (2004) in using sectoral real value added to construct

S ij = | s ni s nj | ,

where sni is the share of industry n in GDP of country i. The sectoral shares s are computed using one-digit manufacturing value added data from the UN Statistical Yearbook.40 The more similar the sectoral shares of each industry between two countries, the closer this indicator is to zero. If the value added of all sectors is identical for both countries, the indicator takes the value zero.

Similarity of Currency Regimes of the Two Countries in Question

Shin and Wang (2003) suggest employing a number of variables to accommodate for macro policies as possible determinants of business cycle correlations. They use bilateral correlations of government budget deficits as a proxy for fiscal policy coordination and correlations of M2 growth rates as a measure for monetary policy coordination. We argue that such measures of policy coordination are not appropriate for East Asia given that there is actually no direct macro policy coordination between these countries. Instead, we propose to use a variable for the similarity of currency regimes. As mentioned before, East Asian countries—except for Japan—basically adhere to the informal East Asian dollar standard; namely they pursue (soft) pegs to the US dollar. The external exchange rate constraint can be expected to have an impact on macro policies and thus on business cycles (despite the widespread existence of capital controls in the region). The currency regimes are interpreted as exogenously determined, in a sense that the current situation constitutes a noncooperative Nash equilibrium and that no single country conforming to the East Asian dollar standard will by itself move away from this regime. Hence the currency regime variable is employed as an exogenous variable in system (1) through (4). As in section 6.2 we use a first measure of currency regime (p.120) similarity based on the Frankel and Wei estimations and a second one based on bilateral exchange rate volatility.

Instrumental Variables

I1ijt, I2ijt, and I3ijt are instrumental variables that contain the vectors of the exogenous determinants of the endogenous variables described above. To instrument trade, we can draw on a rich foundation of literature that employs the gravity model of international trade. I1ijt thus includes the standard variables of the gravity model. These are the product of the natural logarithm of the population of both countries, a variable describing combined GDP computed as the product of the natural logarithm of the GDP of both countries, the natural logarithm of the distance between the two countries, and a dummy variable indicating whether the countries share a common land border. Data sources are the same as in section 6.2.

I2ijt captures institutional variables for financial integration. The variables describe the quality of governance, following Imbs (2006), who first suggested instrumenting financial integration with institutional variables, drawing on the finding of La Porta et al. (1998) that legal institutions are an important determinant of financial development. Especially when considering the specific definition of financial integration used in this paper, namely as bilateral FDI, such instrumental variables are appealing, as one would generally expect FDI to increase with improving governance. The governance indicators used here were developed by the World Bank and include “voice and accountability,” “political stability,” “government effectiveness,” “regulatory policy,” “rule of law,” and “control of corruption.”41 The “raw” indicators are measured in units ranging from -2.5 to 2.5, with higher values corresponding to better governance outcomes. For the bilateral indicators used here, these indicators are first rescaled by simply adding 2.5 to each value so that the new range is 0 to 5. Then the sum of each pair is taken. The bilateral indicators hence range from 0 to 10, with a higher value indicating better governance in both countries.

Specialization is instrumented by I3ijt which describes the stage of development of respective countries. More specifically, we employ the natural logarithm of the pairwise difference in GDP per capita, drawing on the finding by Imbs and Wacziarg (2003) that the development of sectoral specialization is related to the development of income per capita.

Instrumental Variable Equation Results

First we replicate Frankel and Rose’s original regression

ρ ijt =α+β T ijt + ε ijt

for East Asia, where trade is instrumented with the gravity variables. As in Frankel and Rose, who find a strongly positive and statistically significant effect of greater trade intensity on the correlation of economic activity, the results for East Asia presented in table 6.2 (p.121)

Table 6.2 Effect of trade intensity on income correlation (instrumental variable estimates of β)

Number of observations


Whole sample




Without China




Without Japan




Without China and Japan




Note: Standard errors in parentheses;

(*) significant at 10 percent level.

indicate a positive effect of trade integration on output correlations. Also, when the regression is run without China, Japan, and China and Japan, respectively, the effect remains positive, even though it becomes smaller and loses significance when China is omitted. Frankel and Rose note themselves that equation (5) may contain an omitted variable bias, but we may interpret these results as a first hint that increasing trade integration in East Asia has a positive effect on business cycle correlations.42

System Equation Results

Table 6.3 depicts the system equation results.43 As before, we run the regression for the whole sample as well as without China, Japan, and China and Japan, respectively. First, the system equation is estimated as described in equations (1) through (4). The results are presented in columns 2, 4, 6, and 8. To control for a potential world business cycle, a new variable is added to equation (1), namely the natural logarithm of the average US growth (USG) rate over the respective periods. The regression results including this variable are presented in columns 3, 5, 7, and 9. It is noteworthy that the inclusion of the world business cycle variable does not seem to affect the estimation results in a substantial way.

The system equation results substantiate the instrumental variable findings, also indicating a positive direct effect of trade integration on output correlations. However, the results are only statistically significant when China is omitted (column 4). By excluding Japan and including the world business cycle variable, the coefficient even becomes negative. The results for the effect of FDI integration on output fluctuations are more robust: in all regressions we see a positive, partially highly significant effect of FDI integration on output co-movements.

Looking at the whole sample (columns 2 and 3), both trade and financial integration stimulate each other as suggested by positive coefficients for β1 and γ1 in equations (2) and (3). Especially, in view of the close relationship between FDI and intra-industry (p.122)

Table 6.3 System equation results

Whole sample

Without China

Without Japan

Without China and Japan

Number of observations














GDP correlation Trade







































0.0082* *






































Trade FDI



















































Common border

















Product of GDP

















Product of population

















FDI Trade

















Voice and accountability

















Political stability

















Government effectiveness

















Regulatory quality

















Rule of law

















Control of corruption

















Specialization Trade


































Difference in GDP per capita

















Note: Standard errors are in parentheses; * significant at 10 percent, ** significant at 5 percent, and *** significant at 1 percent. No constants reported.

(p.123) (p.124) trade in East Asia (Fukao, Ishido, and Ito 2003), this is a very plausible result. Declining transportation costs and decreasing trade barriers have brought about an internationalization of production processes, often referred to as production fragmentation or vertical specialization. In East Asia, this has led to the development of extensive regional production networks, as discussed earlier. Over the past decade, China has grown into a manufacturing hub for East Asia, mainly serving as the last segment in the international production chain. It is therefore sensible that β1 becomes negative when China is omitted from the regression. This might be a manifestation of the special role that China occupies in the regional trade-FDI network.

The results for equation (4) are fairly robust with large coefficients. They suggest that more trade and FDI integration lead to less specialization. This is also consistent with the direction of the estimates for trade and financial integration in equation (1). Indeed, if trade induces less specialization (the European Commission argument) as suggested by the estimates for δ1, then one should expect a positive effect of trade on output synchronization, namely a positive α1, which we find. The same holds for FDI integration.

Less clear-cut and somewhat puzzling are the estimates for specialization in equation (1). Theory predicts that specialization leads to less synchronous output (the Krugman argument), so we should expect a negative coefficient for α3. In table 6.3, however, most estimates for α3 show a positive sign. But it should be noted that the α3 coefficients are very close to zero, so that the specialization effect is very small regardless of whether it is positive or negative. The result might also be an indication that the Krugman argument simply does not apply in the case of East Asia, where trade and FDI are very intertwined, as discussed above. Arguably, specialization that is induced by FDI might actually increase synchronization, because of the nature of the regional production networks. This is certainly an issue that needs further investigation in future research.

As regards exchange rate volatility, the effect on output correlation is negative for the whole sample, but the results are not consistent in the other cases. The relationship between exchange rate volatility and trade, however, is straightforward and confirms the gravity model estimations from the last section. In equation (2) we find a very large and significant impact of exchange rate volatility on trade.44 That is, more volatility significantly reduces trade in East Asia. This finding is important as it provides a strong argument for regional exchange rate coordination, especially considering that intraregional trade accounts for more than half of total trade for the majority of East Asian countries.

The overall effect of trade integration on output fluctuations constitutes the direct effect (α1) and the indirect effects through the impact of trade on FDI (γ1α2) and specialization (δ1 α3). Accordingly, the overall effect of financial integration on business cycles constitutes the direct effect (α2), as well as the indirect effects through trade (β1α1) and specialization (δ1 α3). As noted before, the direct effects are positive in both cases (α1 〉 0, α2 〉 0).45 Because trade and FDI stimulate each other, their indirect effects on output synchronization are positive as well (γ1α2 〉 0, β1α1 〉 0).46

(p.125) The indirect effects resulting from the impact of trade and FDI integration on specialization are less robust, but most estimates for α3 show a positive sign, which would imply a negative indirect effect, as both trade and FDI have negative effects on specialization in equation (4). If the overall effects of trade integration on business cycles is calculated for the baseline scenario in column 2, we find the overall effects on output correlations to be positive for both trade integration (α1 + γ1α2 + δ1α3 = 0.0241) and FDI integration (α2 + β1α1 + δ2α3 = 0.0217). Calculating the overall effects of trade and FDI integration for the other regression results in columns 3 through 9 also yields positive effects.


Summing up, the system equation results indicate that both trade and FDI integration have a positive direct impact on output fluctuations in East Asia. Moreover the overall effects are positive as well, as trade integration tends to stimulate FDI integration, and vice versa. The estimates for specialization effects are less robust, but in most cases we find a positive—albeit small—effect on output correlations. Furthermore the results suggest that FDI and trade integration reduce specialization, which might be a result of production fragmentation and vertical specialization that has taken place in East Asia over the past decade.

If correct, the results suggest that further economic integration will make output cycles more synchronous in East Asia. In other words, as East Asia becomes more integrated through trade and FDI investment, its output fluctuations will also become more similar. This, in turn, implies that a common macro and exchange rate policy will increasingly suit each individual East Asian country as the need for tailor-made national policies diminishes. Moreover the estimates reconfirm our previous findings that the similarity of exchange rate regimes has a significant positive effect on trade integration in East Asia, underscoring the argument for intraregional exchange rate stabilization. However, a word of caution is warranted: some effects are not always statistically significant, and the quality of the data used for the estimations is not as good as one would wish.

6.4 Monetary Integration, Investment Conditions, and the Development of Regional Capital Markets

In the classic as well as neoclassic traditions, money is regarded as a veil over the real economy.47 That is to say, money is merely a means of payment on spot markets and a store of value. The neutrality of money implies that the central bank cannot affect the real economy, and money therefore also has no impact on the development of a country’s economy. But if money is regarded as a means of deferred payment in an uncertain world with enterprises that finance investment through capital markets, the denomination of debt becomes crucial for production, employment, and growth (Nitsch 2006).

(p.126) Acknowledging the central role of money in economic development, an analysis of monetary integration between emerging and developing economies, such as those in East Asia, has to address different questions than does an analysis of monetary integration between developed countries, such as those in the euro area. The former has to take into account the specific problems developing countries face.48 Two of the features that characterize any less developed country are a weak currency status (Schelkle 2000) and a lack of financial maturity (Bordo and Flandreau 2003).49 As will be argued below, developing countries’ lack of an international currency puts a number of constraints on their development prospects and increases financial fragility. This section therefore investigates whether, and to what extent, monetary integration could be used as a strategy to overcome these impediments in order to create favorable macroeconomic conditions, that is, to promote accumulation and to reduce the risk associated with asset and liability dollarization.

The international monetary system is characterized by currency competition and a hierarchy of currencies (figure 6.3). Cohen (1998: 114) likens currency competition to a “vast, three-dimensional pyramid: narrow at the top, where a few popular currencies dominate; increasingly broad below, reflecting varying degrees of competitive inferiority.”50

Different currencies exhibit different qualities in their function as storage of wealth. In an open economy the central bank has to defend the national currency against competition with other currencies in the portfolios of economic agents. Because of devaluation expectations, an agent holding a weaker currency will demand a higher premium, that is, a higher real interest rate. The international hierarchy of currencies hence sets off a cumulative process by which investments in hard currency countries can be financed most cheaply of all, while market forces cause higher interest rates to be demanded in weaker currency countries, making credits dearer and thus retarding growth and development (Nitsch 1999b). High real interest rates are a problem for some East Asian countries, but not all.

A Reconsideration of Costs and Benefits

Figure 6.3 The currency pyramid

(p.127) While least developed countries such as Cambodia (11.2 percent) and Lao (24.2 percent) suffer from relatively high real interest rates, others such as China (2.4 percent), Malaysia (2.3 percent), the Philippines (4.3 percent), and Thailand (2.2 percent) have managed to establish relatively low levels of real interest despite their weak currency status.51

Low productive power domestically corresponds to low monetary power internationally. This is demonstrated by the fact that developing countries’ currencies are not capable of entering into international contracts (Riese 2004). Weak currency countries are confronted with what Eichengreen and Hausmann (1999) have termed “original sin.”52 They describe original sin as a situation in which the domestic currency cannot be used to borrow long term, even domestically. As a result financial fragility is unavoidable because all domestic investment will have either a currency mismatch (projects that generate domestic currency will be financed with an international currency) or a maturity mismatch (long-term projects will be financed with short-term loans). Both currency and maturity mismatches increase the danger of financial crises. Indeed original sin is widely regarded as one of the causes of the Asian crisis (see the analysis in chapter 3).

Moreover, in a phenomenon McKinnon (2005) has named “conflicted virtue,” weak currency countries are unable to lend in their own currencies, forcing creditor countries with weak currencies to cumulate currency mismatches. As most East Asian countries have turned into creditor countries after the Asian crisis, the massive buildup of foreign exchange assets has created a conflicted virtue problem for them. While Brunei, Japan, and Singapore have had current account surpluses for partly more than two decades and China has had more modest current account surpluses since 1994, even the five former crisis economies (Indonesia, Korea, Malaysia, the Philippines, and Thailand), which had current account deficits before 1997, have now accumulated large stocks of liquid dollar assets in both private and public portfolios (see annexes 7 and 9). With mounting dollar claims, non-US holders of dollar assets have to worry more about domestic currency runs, which would cause a domestic currency appreciation and hence a decline of their net wealth. Countries are therefore inclined to avoid large-scale appreciation of their currencies, which might invoke protests from deficit countries about unfair competition through an undervalued currency.53

A weak currency status therefore brings with it multiple problems, from retarding financial and real development and increasing the risk of financial crisis to triggering potential trade conflict. A hardening of the currency, or a “pyramid-climbing,” is therefore a crucial precondition for creating favorable investment conditions and enabling sustainable economic development and for overcoming the problems of original sin and conflicted virtue. To analyze if and how monetary integration could contribute to a hardening of currencies, it is important to understand what makes a currency an international currency.

As Hyman Minsky observed, anyone can create money by issuing IOUs,54 but the difficult part is to get them to be generally accepted (Kregel 2006). The easiest way to have (p.128) one’s IOUs accepted is to generate liabilities that can only be extinguished through possession of these IOUs. A government can domestically enforce the acceptance of its currency through the fiscal system (and up to a certain degree through its legal system). It can create a tax liability on its citizens that can only be redeemed by rendering the government’s IOU in the form of money issued by the government. But this does not work internationally, as a government can only tax its own citizens who are subject to government regulations but cannot force nonresidents to hold claims. The only way to make the currency internationally accepted is by building an expectation that these liabilities will act as perfect substitutes for other countries’ monetary authorities’ liabilities.

A number of conditions can be identified that contribute to building such expectations. First, the confidence in a currency’s future value is dependent on the political stability of the country of origin (Cohen 2000). This is the quintessential precondition for establishing a track record of relatively low inflation and low inflation variability. Second, countries need to develop sound and credible fiscal institutions.55 In conjunction with noninflationary income policies, an austere fiscal framework lays the groundwork for a noninflationary monetary environment with low nominal as well as real interest rates.

Third, countries need to establish credible monetary regimes. Unpredictable monetary policy makes agents unsure about the future real value of their assets issued in domestic currency and may lead them to denominate them in international currency (Jeanne 2005). Establishing a strong, (de facto) independent central bank with strong inflation aversion and a clear monetary policy objective is an important way to keep down inflationary expectations and to reduce this uncertainty.

Fourth, not running into international debt but instead striving for a surplus in the trade and current account favors national autonomy and employment and helps to create expectations of an appreciation of the national currency (Nitsch 2006). From a long-term development perspective, it is not the short-term stabilization of the exchange rate that is of central importance but rather the durable enhancement of a currency’s quality (Fritz 2002). The quality of a nation’s currency is undermined when a currency regime is chosen that achieves price and exchange rate stabilization at the cost of an increase in the country’s foreign debt. Instead, countries need to develop the ability to generate foreign reserves by generating export surpluses. Such a strategy requires a tendency toward an undervaluation of the currency (Riese 2004). Keeping domestic money scarce represents the necessary condition for securing the undervaluation of the currency because—due to its deflationary effects internally—it allows a country to achieve a price advantage over foreign countries, providing the precondition for the stimulation of exports. Successful examples of this development strategy are Western Germany in the 1950s and Japan in the 1960s and 1970s. The East Asian tiger economies have very successfully followed this strategy more recently.

But developing sound fiscal and monetary institutions and generating export surpluses might not be enough. The literature on the determinants of key currency status points to (p.129) another factor, namely the size of the economy. Matsuyama, Kiyotaki, and Matsui (1993) explain the international use of currencies and, succinctly, the determinants of key currency status as a function of relative country size and the degree of international economic integration. Because of network externalities and transaction costs, the global portfolio is concentrated in very few currencies. In some ways money is comparable to language, whose usefulness is also dependent on the number of people with whom one can communicate; similarly a currency’s utility rises with the number of other market participants using the same currency (Dowd and Greenaway 1993). A currency’s attractiveness increases with its transactional liquidity, which in turn is dependent on the existence of well-developed and broad financial markets that offer a wide range of short and long-term investment opportunities as well as fully operating secondary markets (Cohen 2000). Eichengreen, Hausmann, and Panizza (2005) point out that larger countries offer significant diversification possibilities, while smaller countries add fewer diversification benefits relative to the additional costs they imply.

As a result the global portfolio is concentrated in a small number of currencies (those at the top of the international currency pyramid) for reasons partly beyond the control of even the countries that follow sound domestic policies. Developing key currency status is hence a very difficult and maybe even impossible endeavor for small economies.56 There is also empirical support for this view. Eichengreen et al. (2005) show larger economies to have less of a problem with original sin than smaller ones. Using three different measures of size (log of total GDP, log of total domestic credit, and log of total trade), their estimates suggest that all measures of size are robustly correlated to original sin.

In the face of these constraints, regional monetary integration could be employed by developing nations as a strategy for overcoming their weak currency status and the inability to enter into international contracts with domestic currency. While a hardening of the national currency could, in principle, also be achieved by each country alone through austere monetary and fiscal policies and through the generation of export surpluses, it might not suffice to develop an international currency. Monetary integration, however, could address both issues at the same time. One the one hand, it would place an external constraint on countries participating in the monetary integration process, facilitating the domestic policy adjustment necessary for a hardening of the currency. In the European context Giavazzi and Pagano (1988) have termed this the “advantage of tying one’s hands” (see section 6.1). Monetary integration could be used as a disciplinary device for inflation-prone countries by forcing policy makers to pursue more restrictive fiscal and monetary policies than they would otherwise. In return, countries would enjoy potential credibility gains. This strategy worked reasonably well in Europe in the run-up to monetary union.57

At the same time monetary integration (monetary union in particular, but also the creation of a basket currency) would address the problems of original sin and conflicted (p.130) virtue by creating a larger economic entity with vast investment opportunities that would be hard to ignore by international asset managers and that, if backed by austere monetary and fiscal policies, would increase chances of entering the club of international currencies. The underlying logic is that the whole would be equal to more than the sum of its parts. To be sure, size alone is not enough: Russia, Argentina, and Brazil are examples of large emerging economies that, despite their size, face problems of original sin. But, as explained before, good domestic policies are necessary as well, and over most of the past decades these three countries have not been prime examples for prudent economic policy-making.

Bordo, Meissner, and Redish (2005), in an analysis of how five former British colonies (the United States, Canada, Australia, New Zealand, and South Africa) overcame original sin, point to another interesting factor: the role of shocks such as wars and massive economic disruptions. For instance, the onset of World War I essentially closed the London capital market and led Canada, Australia, New Zealand, and South Africa to suspend the gold convertibility of their domestic currencies and raise funds domestically; that is, the disruption of the war basically forced these countries to create domestic bond markets.58 An interesting parallel can be drawn with East Asia, where the Asian crisis had been such a major shock. As one reaction to the crisis, the region has begun trying to develop regional bond markets (see chapter 3). ABF I and II are first attempts to bundle bond issues by East Asian countries in order to make it more attractive for international investors to include them in their portfolios.

This might be a step in the right direction for East Asia, and pursuing a similar cooperative path in monetary policy could help overcome the problems of original sin and conflicted virtue that are associated with a weak currency status. Individually, most East Asian nations will have little prospect of escaping this trap. For instance, it is hard to imagine how a small developing economy like Vietnam will be ever able to develop its national currency, the dong, into an internationally accepted currency.59 Yet, united, the region is far too important to be ignored by currency traders. As a possible first move in this direction, in December 2005 the ADB suggested the launch of a currency unit comprised of a basket of regional currencies akin to the ECU. This virtual basket currency could give observers a taste of the kind of standing a regional East Asian currency could achieve in international financial markets. The benefits of such a “parallel currency approach” to monetary integration will be discussed in chapter 11.

6.5 Monetary Integration and the Recovery of Monetary Policy Influence

When you got nothing, you got nothing to lose.

—Bob Dylan in “Like a Rolling Stone,” from the album Highway 61 Revisited (1965)

(p.131) In this last section of the chapter, the potential for regional monetary integration to be a way for East Asian nations to (re)gain some degrees of monetary influence in the region is highlighted. Monetary cooperation and integration are not necessarily equivalent to a loss of autonomy, as a country can only lose what it has. This might seem paradoxical at first, as it is usually claimed that monetary integration means a loss of monetary independence (see section 5.1). This equivalence, however, is true only if the participating countries enjoyed monetary autonomy before. In this context it is important to distinguish between sovereignty and policy autonomy: a country might well have the formal right to pursue a certain policy but find itself unable to do so because of policy choices of other nations (DeMartino and Grabel 2003).60

In the discourse preceding the creation of EMU, opponents of monetary unification argued that giving up one’s own currency would mean losing a forceful policy instrument. However, the discussion in Europe had been somewhat unreal. As Goodhard (1995: 458) pointed out, through their EMS membership EMS countries (except Germany) had already abandoned discretionary monetary policy long before monetary union was finalized, meaning that there would be “virtually no economic cost in doing so formally and completely by moving to a full monetary union.”61 According to De Grauwe (1994: 3) it was “no exaggeration to state that Germany dominates the monetary conditions in Europe.”

For a formal test of this hypothesis, the following simple Taylor rule is estimated:

r ct = b 1 + b 2 I P ct-2 + b 3 π ct-2 + b 4 r Bt-2 +ε,

where rct is the central bank interest rate of country c at time t, IPct-2 is country c’s lagged industrial production, and πct-2 is the lagged inflation rate. rBt-2 is the lagged Bundesbank discount rate.62 The results for the period ranging from August 1971 to December 1998, namely from the breakdown of the Bretton Woods system to the launch of the euro, are presented in table 6.4.

The estimates show that the monetary policy decisions of all European countries analyzed here, with the exceptions of Greece and Norway (both of which were not members of the EMS),63 were driven to a large extent by the interest rate policy of the Bundesbank.64 While domestic inflation and industrial production also seem to have played a role in the conduct of many of these countries’ monetary policies, the Bundesbank discount rate had a far greater impact for all countries. US interest rate policy, in contrast, and as one would expect, was not influenced by the Bundesbank.

One can push Goodhard’s argument further and argue that the European countries (except Germany) that entered monetary union not only did not lose monetary autonomy but actually (re)gained some degrees of monetary policy influence through EMU membership. Indeed it is arguable that Germany was the only country that actually lost monetary autonomy. Virtually all other EMU member countries have obtained a voice in monetary policy decisions. Instead of following the Bundesbank’s policy stance, all EMU member countries are now represented through their national central bank governors in the ECB’s (p.132)

Table 6.4 Determinants of monetary policy decisions, August 1971 to December 1998

Bundesbank discount rate


Industrial production


R squared





























































































































United Kingdom









United States









Source: Calculation with data from IFS and Global Financial Data.

Note: Newey West standard errors in parentheses. *** denotes statistical significance at the 1 percent level, ** at the 5 percent level, and * at the 10 percent level.

(a.) Luxembourg and Belgium had a monetary union, the Belgium–Luxembourg Economic Union, since 1922.

(p.133) Governing Council, the ECB’s monetary policy-making body. Until Slovakia joined EMU as the sixteenth member on January 1, 2009, the central bank governors of each euro area member, along with the six members of the ECB’s Executive Board, had the right to vote at each council meeting, providing smaller countries such as Austria and Luxembourg with the same de jure influence on monetary policy as the larger members France, Germany, Italy, and Spain.

With Slovakia’s entry the voting rules in the ECB Governing Council have been changed to a rotating system to ensure efficient decision-making in the face of an increase of the Council’s current size of twenty-two members to maybe thirty or more in the process of euro area expansion. The new voting scheme gives more weight to the larger euro area economies, but maintains a bias in favor of the smaller countries.65 In any case, regardless of the distribution of voting rights, the ECB statutes stipulate that national central bank governors sit on the Governing Council in a personal and independent capacity, not as representatives of their own countries.

Among the central banks of the old EU member countries that did not join EMU, the monetary authorities of Denmark and Sweden have basically been forced to shadow ECB policy, without exerting any influence themselves on the policy they have to follow.66

A different but also Dylanesque situation applies for most East Asian countries when it comes to the conduct of their monetary policy. As mentioned earlier, the region’s exchange rate policy can be described by the East Asian dollar standard, a situation in which all East Asian countries with the exemption of Japan operate more or less tight pegs to the US dollar. The pegs were temporarily lifted during and after the Asian crisis (except in the cases of China, Hong Kong, and Myanmar), but over the last few years there has been a resurrection of the dollar standard (McKinnon 2001; Schnabl 2009), even though most of the countries officially declare their exchange rate regimes as (managed) floats.

Table 6.5 gives an overview of East Asian countries’ official and de facto exchange rate regimes. The estimated weights of the US dollar in the hypothetical currency baskets of East Asian countries presented in column 3 show that Cambodia, China, Hong Kong, Malaysia, Myanmar, and Vietnam have basically maintained fixed exchange rates vis-à-vis the dollar.67 Lao and the Philippines also manage tight pegs to the dollar, with the weights of the dollar in their hypothetical currency baskets above 90 percent. Indonesia, Korea, Singapore (and hence also Brunei, which maintains a currency board vis-à-vis the Singapore dollar), and Thailand allow considerably more flexibility toward the dollar than the previously mentioned countries but nevertheless display a strong dollar orientation in their exchange rate regimes, with dollar weights ranging from 70 to 90 percent. The only exception to this pattern is Japan, with a dollar weight of only 66 percent.

The situation can be illustrated in a stylized manner by the well-known n – 1 problem: given n currencies, there are n – 1 independent bilateral exchange rates. Maintaining these n – 1 bilateral exchange rates fixed ties down as many national money stocks. Hence, there is only one degree of freedom left. In the absence of exogenous rules, the price level is (p.134)

Table 6.5 De jure and de facto exchange rate regimes in East Asia

IMF classification

Estimated weight of the USD in hypothetical basket (%)

Brunei dollar

Currency board arrangement (vis-à-vis the Singapore dollar)


Cambodian riel

Managed floating with no pre-announced path for the exchange rate


Chinese yuan

Crawling peg: daily fluctuations in the RMB-USD exchange rate are limited to +/−0,3%, against other currencies +/−3%


Hong Kong dollar

Currency board arrangement (vis-à -vis the US dollar)


Indonesian rupiah

Managed floating with no pre-announced path for the exchange rate


Japanese yen

Independently floating


Korean won

Independently floating


Lao kip

Managed floating with no pre-announced path for the exchange rate


Malaysian ringgit

Managed floating with no pre-announced path for the exchange rate


Myanmar kyat

Managed floating with no pre-announced path for the exchange rate (officially pegged to SDR)


Philippine peso

Independently floating


Singapore dollar

Managed floating with no pre-announced path for the exchange rate


Thai baht

Managed floating with no pre-announced path for the exchange rate


Vietnamese dong

Managed floating with no pre-announced path for the exchange rate


Note: The estimates in column 3 were calculated using Frankel and Wei (1994) as presented in section 6.2 with daily exchange rates from 1/1/1999 to 8/14/2008. All dollar weight estimates are significant at the 1 percent level. The yen was regressed only on the dollar and the euro. Data are from Datastream (Reuters and Tenfore). Classifications are from IMF (2008c).

(p.135) determined by the remaining nth country. In the case of today’s East Asia (excluding Japan), this means that East Asian currencies are the n – 1 currencies, and the US dollar is the nth currency through which the price level is determined. In other words, policy autonomy rests in the hands of the United States, while East Asian countries more or less have to follow, depending on the degree of rigidity in their exchange regimes.

Although most East Asian countries maintain some form of capital controls that, in principle, should allow for monetary policy autonomy in the face of an exchange rate target, historical evidence suggests quite strongly that capital controls are porous and are easily circumvented.68 Their effectiveness erodes over time as domestic and international investors and traders find channels, such as “favorable” trade invoicing, to elude them (Edwards 1999).

Furthermore there is a growing amount of empirical literature that puts into doubt even the traditional argument that countries with flexible exchange rates are able to isolate their domestic interest rates from changes in international interest rates. Frankel, Schmukler, and Servén (2004), for instance, find that while floating regimes afford greater monetary independence than fixed regimes, floating regimes offer only temporary monetary independence.69 That is, while the speed of adjustment of domestic rates toward the long run, one-for-one relation with international interest rates is generally lower under floating than under fixed regimes, even floating regimes cannot exert autonomous monetary policy. Their findings suggest that besides the United States, Germany (now the euro zone) and Japan appear to be the only countries that can independently choose their own interest rates in the long run. Similarly Reade and Volz (2009b) who measure monetary independence across various regions find little monetary independence for smaller nations; even large emerging countries like Brazil, China and India exhibit dependency in their monetary policy whereas big industrialized countries, such as Japan, the United States, and the euro area, do exhibit considerable independence. Fratzscher (2002) also observes that even under flexible exchange rate arrangements it becomes ever more difficult to pursue independent monetary policy.70 These studies seriously put into doubt the idea that all countries with a national currency automatically enjoy monetary autonomy. This is particularly the case for highly open economies—like those in East Asia (see section 5.3).71

Moreover, the underdevelopment of domestic financial markets hampers the conduct of monetary policy, creating a situation wherein the vast majority of East Asian countries have not been able to effectively pursue independent monetary policy. The People’s Bank of China (PBC), for instance, did not raise interest rates for more than a decade until October 2004, and even this change was very modest and more symbolic than of practical import.72 Similarly Malaysia raised its rates in November 2005 for the first time since 1998. The most extreme case of a loss of monetary autonomy in East Asia is Hong Kong. To maintain its currency board vis-à-vis the US dollar, the Hong Kong Monetary Authority has to move in tandem with the Federal Reserve, even though local inflation development has been very different from US inflation over recent years.

(p.136) To formally investigate the degree of monetary policy independence that East Asian countries have, a cointegration framework is employed. In particular, a vector equilibrium-correction model is used to identify whether East Asian countries belong to steady-state relationships involving interest rates of other countries and whether they adjust to such relationships. If Xt, is a vector of money market rates, the model is described as

Δ X t =α β Δ X t1 + k=1 K1 Γ k Δ X tk + ε t ,

where K is the number of autoregressive lags and βXt−1 are other cointegrating vectors.73 To illustrate, assume a big country a and a small country b so that

X t =( r a r b ).

If testing suggests that r = 1, there is one cointegrating vector in this system. It is possible to then carry out likelihood-ratio tests to test for the form of the cointegrating vector. Suppose that the restrictions required for the following system are accepted:

r a,t = β 0 + r b,t + ε t .

The two countries respond one-for-one in interest rates; if the interest rate of country b rises by 25 basis points, so does the rate of country a. This can be shown in the vector equilibrium-correction model of (6):

( Δ r a Δ r b ) t = ( α 1 α 2 ) t ( β 0 β 1 β 2 ) ( 1 r a r b ) t1 + k=1 K1 Γ k Δ X tk + ε t .

With the restriction β1 =−β2 = 1, then

( Δ r a Δ r b ) t = ( α 1 α 2 ) t ( r a r b β 0 ) t1 + k=1 K1 Γ k Δ X tk + ε t .

It may be expected that country a would not adjust to this cointegrating vector. As it is a large country, it might be expected to exert monetary policy independence, so α1 = 0. Country b, which is a small country, may be expected to adjust, so α2 ≠ 0. Furthermore α2 describes how much of any disequilibrium is corrected each period, as α = ΔXt/(βXt−1); hence (ceteris paribus) a speed of adjustment can be calculated. The smaller is this coefficient, the more independent is a country’s monetary policy, since it devotes less of its attention to correcting to what other interest rates are doing. As such, the α matrix is very informative about the nature of monetary policy independence. A country not adjusting to a cointegrating vector in which it appears is said to “drive” the system: the level that (p.137) the country’s interest rate is at is not constrained by the cointegrating relationship but in fact dictates what level that cointegrating relationship takes.

Table 6.6 presents estimates of this model for eight East Asian countries (China, Hong Kong, Korea, Thailand, Indonesia, Malaysia, Singapore, and the Philippines). As the large a countries we chose the world’s three main currency areas, namely the United States, the euro area, and Japan. The East Asian countries are then the small b countries. We use monthly three-month interbank interest rates from IFS, Datastream, and the Global Financial Database for the periods listed in table 6.6. Along each row are the results relating to each country. Under α̂i are the estimated speed of adjustment coefficients, followed by the estimated cointegrating vectors under β̂i. Underneath each estimated coefficient is its standard error, which indicates whether that particular country enters the cointegrating vector, while bold typeface for the coefficient value denotes that that interest rate adjusts significantly to that particular cointegrating vector. The cointegrating vectors relating to each country b (reported as β̂1) show that in each system there was a cointegrating vector relating to that country. Generally there is one steady-state relationship relating to the small country, and a wider cointegrating relationship.

The United States enters all cointegrating vectors. It does not adjust to any of those, which means that it “drives” the system. This suggests that the United States plays a role in determining monetary policy of all the countries considered, to varying extent, depending on the α̂ coefficients of the b countries. Japan also enters several cointegrating vectors, while the euro area doesn’t seem to affect the monetary policies of East Asian countries at all. Hong Kong has a one-to-one relationship with US rates and each period moves to correct a substantial amount in response to an US move in interest rates. China responds to US interest changes to a lesser extent than Hong Kong, but China’s policy is influenced by world interest rates, in particular the US rate. Korea seems to respond quite strongly to changes in both Japanese and US rates. Singapore also moves with US and Japanese interest rates, which is in line with its basket regime in which both the USD and the yen are included. Indonesia, Thailand, and Malaysia respond to US interest rate movements (Indonesia also responds to Japanese rate changes), but rather modestly, suggesting that they do enjoy some monetary policy autonomy.

If the argument is correct and the status quo level of East Asian monetary independence is limited, then the costs of monetary integration in East Asia, at least for the economically small and developing countries, are much lower than commonly assumed. A common agency approach to monetary integration, as pursued in Europe, could pave the way for greater monetary policy flexibility in East Asia. Through the creation of a common monetary arrangement, or even a common currency, which could float freely against the dollar and the euro, East Asian countries could potentially gain some degrees of shared monetary independence. In terms of the n – 1 problem, creating a regional monetary arrangement that would take a regional currency or a basket of regional currencies as reference would mean that the one degree of freedom would be shifted to the region. While East Asian (p.138)

Table 6.6 Cointegrating vectors and adjustments for East Asia
































6.58 [0.23]

Hong Kong











11.76 [0.16]





−2.80 (0.24)

0.09 (0.02)

−2.66 (0.12)







5.70 [0.22]















8.94 [0.18]

















14.98 [0.04]












K(4), AC(1)

12.66 [0.12]














K(4), AC(1)

9.86 [0.13]

















4.80 [0.31]

Source: Own calculations.

Note: β1 is the first cointegration vector, α1 the adjustment of East Asian country b to that vector; only the vector pertaining to that country is reported. A bold typeface reveals a significant parameter. The right panel reports any misspecification issues with the model; K(n) signifies n equations in that model had an excess kurtosis problem, while AC(n) denotes that n equations had problems with autocorrelation. Dummies records how many dummies for identified outliers were required. Test is the test of overidentifying restrictions, with the p-value in square brackets.

(p.139) countries would still face an external constraint on domestic economic policies, the great difference between a regional currency arrangement or a multilateral monetary union in East Asia and a continued (informal) dollar pegging under the East Asian dollar standard (or even full dollarization) is that the former would give East Asian countries a say in their common monetary and exchange rate policy or policies.74

In the case of multilateral monetary union in East Asia, each member country could have a share in the common central bank’s policy-making through a pooling of sovereignty. In contrast, continued dollar pegging (or dollarization), while basically requiring the same sacrifices in domestic policy autonomy as monetary union, would mean that all monetary policy influence is abandoned (permanently).75 One potentially delicate political problem, however, would be the institutional design of a common central bank and the apportionment of power in an East Asian monetary union. A representational structure akin to that of the EMU, with more or less equal rights between all members, would be unrealistic if either China or Japan were involved, making such an arrangement less attractive for the smaller Southeast Asian countries due to the potential dominance of the bigger member countries. Conversely, neither Japan nor China (nor Korea) would be likely to accept the disproportional representation of the economically smaller Southeast Asian member countries the way Germany did in Europe.76 Even though they are also highly heterogeneous, a solution would probably be easier to reach among ASEAN member nations.

The discussion here and in the preceding sections should therefore not evoke the impression that regional monetary integration—and especially monetary unification— comes without problems and costs. Monetary integration is not a free lunch, and it would be wrong to simply dismiss concerns about the costs of monetary integration—as they are emphasized in traditional OCA theory—and only point at potential trade-creating effects and a likely convergence of business cycles.77 Especially the surrender of national monetary sovereignty in the case of monetary unification involves uncertainties, not at least regarding the partner countries’ future policies. Moreover there would still be the political cost of giving up formal monetary independence, such as the loss of monetary policy autonomy illusion and the loss of the national currency as a symbol. Giving up de jure independence could also involve diplomatic costs in the form of political dependency on foreign nations.

But still, the analysis in this chapter showed that the costs of monetary integration in terms of the loss of independent domestic macro policies and the exchange rate policy as a reliable adjustment mechanism are lower than commonly thought, and that the benefits go beyond a mere saving of transaction cost and a better allocation of resources. It tried to highlight monetary integration as a potential strategy to overcoming structures that present obstacles to economic development and to create a stable macroeconomic environment that is conducive to investment and growth. In particular, the chapter drew attention to the potential contribution of monetary integration to developing regional financial (p.140) markets in East Asia and to overcoming the problems of asset and liability dollarization that are possible sources of macroeconomic instability. Moreover the discussion of the Lucas critique tried to make clear that monetary integration needs to be understood as a dynamic process, in which the criteria for successful monetary integration that were laid out in OCA theory are likely to change through a policy of integration. Rather than being a precondition for successful monetary integration, fulfillment of certain OCA criteria should be regarded as the desired result of integration. In this regard the results of our analysis of the effects of real and financial integration on output movements indicate that further economic integration between East Asian countries would make their output cycles more synchronous, so that the region would increasingly fulfill this important OCA criterion. The overall assessment of the costs and benefits of monetary integration in East Asia that results from this analysis is therefore positive.

This positive conclusion, however, does not imply that a policy of monetary integration in East Asia should necessarily be directed at monetary unification. As was made clear by the definition of monetary integration in chapter 1, different degrees of monetary integration are possible, requiring different degrees of political commitment and entailing different degrees of loss of national autonomy and sovereignty. It is therefore important not to center the discussion about monetary integration in East Asia too narrowly on monetary unification. Indeed, focusing the discussion too much on monetary unification could be counterproductive, as it might evoke the impression that monetary integration is a predetermined path that inevitably leads to monetary union. The following chapters will consider possible exchange rate options for East Asia and, based on an analysis of the European experiences with monetary integration, develop a strategy for monetary integration for East Asia that takes into consideration the specifics of the region.


(1.) One need not necessarily go as far as Buiter (2000: 222) who condemns OCA theory as “one of the low points of post–World War II monetary economics” or as Schelkle (2001b: 2) who describes it as a “dead end for debates on monetary integration.”

(2.) This development is related to the paradigm change in monetary theory that was triggered by Friedman’s (1968) essay “The Role of Monetary Policy.” The dwindling influence of the Keynesian school of thought since the 1970s and the simultaneous growing prominence of the monetarist school have also strongly changed the thinking about monetary integration, which, as a result, “has shifted from Keynesian-inspired skepticism about monetary integration towards a monetarist-induced sympathy for the cause of monetary union. This shift in intellectual environment has certainly been influential among policy makers and helps to explain the initiatives taken in Europe at the end of the 1980s that ultimately led to monetary union.” (De Grauwe 2001: xii)

(3.) The Phillips curve originates from an analysis by Phillips (1958) on the long-run relationship between unemployment and nominal wage changes in the United Kingdom. Phillips found that high wage rises correspond to low unemployment, and vice versa. Strictly speaking, the relationship between unemployment and inflation refers to the “modified Phillips curve,” which was described by Samuelson and Solow (1960). Under certain conditions, the change in price level and nominal wages are identical so that the original meaning remains unchanged (see Bofinger 2001: 97ff).

(4.) As will be discuss later, the monetary and exchange rate regime, by its influence on the quality of a nation’s currency, does have an important impact on financial market development and investment conditions and thereby on economic growth and development. But these effects have nothing to do with “fine-tuning” the economy. (p.227)

(5.) On what monetary policy can achieve and what not see also Bernanke et al. (2001: 11–6)

(6.) As Buiter (2000) points out, the OCA theory was developed in an era when international financial integration was limited and where foreign exchange and capital controls were the norm. At the time Mundell’s considerations therefore served their purpose.

(7.) For an overview of exchange rate theories, see Sarno and Taylor (2002).

(8.) The seminal contribution to the portfolio balance theory of the exchange rate is Branson (1979). A comprehensive treatment of the portfolio balance model is given in Branson and Henderson (1985).

(9.) Keynes highlighted the principal uncertainty of the future. In this tradition, the monetarist-Keynesian approach emphasizes the role of money as a bridge between the present and an uncertain future. See Nitsch (1999a).

(10.) This argument will be developed for East Asian countries in section 6.5.

(11.) The Lucas critique was given great prominence in the discussion on monetary integration by Frankel and Rose’s (1998) article on “The Endogeneity of the Optimum Currency Area Criteria.”

(12.) In the words of Lucas (1976: 42): “In short, it appears that policy makers, if they wish to forecast the response of citizens, must take the latter into their confidence.”

(13.) Examples are reactions to changes in progressive taxation, employment programs, or the central bank’s monetary policy (Schelkle 2001b).

(14.) See Schelkle (2001a). For a comprehensive treatment of the interaction between wage bargaining and monetary policy in the EMU, see Dullien (2004).

(15.) Structural inflation denotes that the causes of inflation have their roots in structural factors such as the nature of the monetary system (i.e., a lack of independence of the central bank) or the wage bargaining system (i.e., an indexation or fragmentation of wage bargaining). As Buiter (2000: 223) puts it: “[T]he duration of nominal wage and price contracts and… the extent to which they are synchronised or staggered is subject to an obvious application of the Lucas critique. These contracting practices are not facts of nature, but the outcomes of purposeful choices. Changes in the economic environment conditioning these choices will change the practices.”

(16.) The disciplining effects of the EMS will be discussed in chapter 10.

(17.) For an analysis of business networks in East Asia, see Tachiki (2005) and Hamilton-Hart (2005).

(18.) See Fernández-Arias, Panizza and Stein (2004). These problems are nicely illustrated for Brazil and Argentina in the 1990s by Eichengreen (1998).

(19.) NAFTA is a case in point.

(20.) With reference to the European Commission’s study “One Market, One Money” (Emerson et al. 1992), Rose (2000) called his article “One Money, One Market: The Effect of Common Currencies on Trade.”

(21.) The countries are Cambodia, China, Hong Kong, Indonesia, Japan, Korea, Lao, Malaysia, the Philippines, Singapore, Thailand, and Vietnam. Taiwan was included to enlarge the sample size. The sample is divided into three subsample periods, 1980 to 1989, 1990 to 1996, and 1999 to 2003, to capture the change over those years, with the crisis years 1997 and 1998 excluded.

(22.) For an overview of the theoretical foundations of the gravity equation see Frankel (1997: ch. 4).

(23.) GDP and population data are from WDI, except for Taiwan, which are from the Taiwan Statistical Data Book 2004. Distance was calculated using the distance calculator of the US Department of Agriculture/Agricultural Research Service Phoenix, Arizona (http://www.wcrl.ars.usda.gov/cec/java/capitals.htm).

(24.) As an alternative measure for the similarity of currency regimes, we compute bilateral exchange rate volatility, namely the standard deviation of the first differences of the natural logarithms of the nominal exchange rate (p.228) between the two countries in question. The estimates for the gravity model using this measure, which are not reported here, are almost identical to those in table 6.1, confirming the robustness of the results. For both indicators we use monthly exchange rates from IFS and the Central Bank of China for Taiwan.

(25.) Using an error correction model, Poon, Choong, and Habibullah (2005) also find that exchange rate volatility has a statistically significant negative impact on real exports in most East Asian countries. The results are also consistent with Baak (2004) who finds a significant negative impact of exchange rate volatility on the volume of trade among Asia Pacific countries. Hayakawa and Kimura (2008) find that intra–East Asian trade is discouraged by exchange rate volatility more seriously than trade in other regions because intermediate goods trade in production networks, which is quite sensitive to exchange rate volatility compared with other types of trade, occupies a significant fraction of trade.

(26.) Population is expected to control for self-sufficiency and size of the respective countries, because a higher population relates to less dependency on exports and imports due to a larger domestic market and accessible natural resources (see Frankel 1997: ch. 4). As discussed before, the smaller East Asian economies are more open than the large ones. Singapore’s share of intraregional trade, for instance, is particularly high due to the fact that it engages considerably in intraregional entrepôt trade.

(27.) Nitsch (2002) has re-examined Rose’s analysis and concluded that a common currency doubles instead of trebles trade. In his response to Nitsch, Rose (2002: 479) sees his general message confirmed: “If I had originally estimated that currency unions leads trade to rise by ‘only’ 100 percent, I still think that Nitsch would have been provoked to write his paper. The mystery remains.”

(28.) Rose and Engel (2002) point out that since well-integrated countries are more likely to establish a common currency union, the trade effect might be somewhat overestimated. That is, the causality may flow from real integration to monetary unification rather than the reverse.

(29.) See Rose (2002, 2004), Frankel and Rose (2002), Rose and Engel (2002), Pakko and Wall (2001), Flandreau and Maurel (2005), Nitsch (2002, 2004, 2008), Berger and Nitsch (2008), Baldwin (2006), and Frankel (2006).

(30.) See also Rose and van Wincoop (2001).

(31.) Frankel and Rose (2002: 461) point to the tendency of their message: “Our estimates seem very large, and we try not to take them too literally. Rather we hope they shift the terms of the debate on common currencies toward a more serious consideration of the somewhat neglected trade benefit.”

(32.) Some of these factors are usually used as explanatory variables in the gravity model.

(33.) Several other papers deal with business cycles in East Asia. Girardin (2004), for example, analyzes correlations of East Asian countries’ growth with Japan.

(34.) Arguably, Shin and Sohn’s (2006) results are affected by the way they measure financial integration, a rather intricate task that will be discussed in greater detail below and that is differently addressed here.

(35.) Again, Taiwan is included to increase the sample size.

(36.) We also tried linear detrending to see if the detrending method affects the estimation results. Moreover, we tried out other measures of economic activity, namely, GDP at constant USD, GDP at PPP, industry value added at constant local prices, and unemployment as a percentage of the total labor force. The results were broadly in line with the ones presented here.

(37.) Shin and Sohn (2006) construct an indirect measure of financial integration based on the returns on financial assets, namely correlations of monthly interest rates for a given year.

(38.) One could also use de jure measures of financial integration. Imbs (2004), for instance, uses the IMF’s classification of capital account restrictions to construct an indicator of bilateral financial integration. But as he (p.229) acknowledges, de jure measures of financial openness say little about effective bilateral financial integration. As a proxy for effective bilateral capital flows, he uses Lane and Milesi-Ferretti’s (2001) measure of international financial integration based on aggregate foreign assets and liabilities. But while the use of such measures is very convenient with respect to data availability, they are rather an indicator for international financial openness than for bilateral integration.

(39.) The BIS actually collects data for cross-border banking activities, but for East Asia only aggregate data are on hand.

(40.) The industries contained in the UN Statistical Yearbook are “agriculture,” “mining,” “manufacturing,” “electricity,” “construction,” “wholesale,” “transport,” and “other activities.”

(41.) For information on the indicators, see Kaufmann, Kraay, and Mastruzzi (2005). Data are available at http://www.worldbank.org/wbi/governance/govdata/.

(42.) For other problems associated with the approach taken by Frankel and Rose, see Kose and Yi (2002).

(43.) The results presented here were obtained using the second measure for the similarity of currency regimes, namely bilateral exchange rate volatility. When using the Frankel and Wei variable, we obtain almost identical results (except, of course, that the signs of α4 and β2 have opposite directions).

(44.) Note that a negative sign for β2 in equation (2) in table 6.3 implies that an increase in exchange rate volatility reduces trade. Accordingly, a reduction of volatility should increase trade.

(45.) Except for trade in case 12.

(46.) There is, however, an indirect negative effect of FDI on output correlations through the trade channel if China is omitted (β1α1 〈 0).

(47.) This view dates back to the 18th century, when David Hume (1752) initiated the famous “oil-in-the-machine” illustration of the neutrality of money.

(48.) Fritz and Metzger (2006) hence distinguish “south–south” cooperation from “north–north” cooperation, where “north” refers to a country’s ability to accumulate debt in its own currency and “south” refers to its disability to do so.

(49.) See also Calvo (2001) and Calvo and Mishkin (2003).

(50.) Similarly Hicks (1969: 89–90) distinguishes “local currency” from “great” currency, which are “widely used in international or distant commerce,” whereas the circulation of local currency is “confined to the area which the government controlled.”

(51.) Data are for 2006 and are taken from WDI. The relatively low level of real interest rates in some East Asian countries is due to two factors. First, capital controls make domestic assets imperfect substitutes for foreign assets in private portfolios. Second, directed lending partly outplays the market mechanism. Access to credit, however, remains a problem due to underdeveloped financial markets in most countries, especially for small and medium businesses. For an explanation how the “threat of currency appreciation” has pushed the interest rates of China and several other East Asian countries below the US interest rate level, see McKinnon and Schnabl (2009).

(52.) The term “original sin” has been criticized by Goldstein and Turner (2004) and Nitsch (2006) because it suggests that it cannot be overcome.

(53.) Dooley et al. (2009a, b) maintain that East Asian countries keep their currencies pegged to and resist appreciation against the US dollar primarily for mercantilistic reasons, namely to increase domestic growth and employment. (p.230)

(54.) An IOU is a promise of money, goods, services, or other items of value, which may be either written or verbal. The name derives from the phonetic pronunciations of the respective letters, which sound like the phrase “I owe you.”

(55.) See, for instance, Corsetti and Mackowiak (2005).

(56.) Eichengreen et al. (2005) argue that Switzerland and the United Kingdom can be regarded as special cases that achieved key currency status due to their unique historical roles.

(57.) After entering monetary union, however, the governments of several EMU member countries have manifestly shrugged off fiscal discipline and effectuated a weakening of the SGP

(58.) The US case is quite different from that of the other four former colonies. See Bordo et al. (2005).

(59.) This size argument certainly does not apply to China and Japan (which, in any case, already has an internationally accepted currency), and only to a limited extent to Korea. Korea has tried for quite some time to promote the international use of the won by granting trade credits in won to importers, albeit with rather limited success. In May 2006 Korea announced a strategy for “facilitating the internationalization of the won,” which includes a schedule to liberalize the won and capital flows. See Korean Ministry of Finance and the Economy (2006).

(60.) Mundell (2002) distinguishes between “policy sovereignty” and “legal sovereignty.”

(61.) Goodhard (1995) does mention the costs of losing seignorage, but he argues that the value of seignorage to a stable country with low inflation is small and that the arrangements made under the Maastricht Treaty for returning seignorage to the constituent national central banks suggests that the net loss or gain to most European countries is of secondary importance. One should also highlight the loss of the national central bank’s ability to act as a lender of last resort, but in the EMU this function is still fulfilled by the European System of Central Banks with the ECB at the center.

(62.) Data are monthly and are taken from IFS and Global Financial Data. As is usual for backward-looking Taylor rules, the estimates are obtained using OLS with heteroskedasticity- and autocorrelation-consistent standard errors. See, for instance, Carare and Tchaidze (2005). Changing the lags does not significantly alter the results.

(63.) The Greek drachma joined the EMS only in March 1998.

(64.) This result is confirmed by Frankel, Schmukler, and Servén (2004) and Chinn, Frankel, and Philips (1993), who find that interest rates in European countries had become completely insensitive to US interest rates but fully sensitive to German interest rates. Using co-integration analysis, Reade and Volz (2009a) also confirm German monetary hegemony within the EMS.

(65.) The ECB statutes determine that “from the time when the number of governors of national central banks [NCBs] exceeds 15 and until it reaches 22, the NCB governors will be allocated to two groups, according to a ranking of the size of their Member State’s share in the aggregate GDP at market prices and in the total aggregated balance sheet of the monetary financial institutions of the Member States which have adopted the euro.… The first group will be composed of five governors with five voting rights and the second group of the remaining governors with the remaining votes.” Once the number of NCB governors reaches 22, the governors will be allocated to three groups according to a ranking based on the above criteria. See Article 10.2 of the Statute of the European System of Central Banks and of the European Central Bank.

(66.) On the Swedish case, see Reade and Volz (2009c). The situation is different for the United Kingdom, where the Bank of England has managed to continue its independent monetary policy tailored to the national economy’s needs. This can be attributed to the size of the UK economy as well as its financial sector, which give it a greater pull compared with smaller economies like Denmark and Sweden.

(67.) On July 21, 2005, the Chinese central bank officially announced that it had abandoned the eleven-year-old peg to the dollar and instead linked the yuan to an undisclosed basket of currencies. Annex 9 presents annual (p.231) regression results for all currencies. It shows that in the case of China the dollar weights have gone down since 2004 (when the dollar still had a 100 percent weight) to about 93 percent in 2007 and 2008. The Chinese move was followed by Malaysia. Several of other countries have also loosened their dollar pegs a bit.

(68.) See table 7.1.

(69.) This is not inconsistent with Kim and Lee’s (2004) finding that the sensitivity of local interest rates to US rates has declined for Korea and Thailand since they moved toward less rigid exchange rate regimes after the Asian crisis.

(70.) See also Calvo and Reinhart (2001 and 2002).

(71.) Buiter (1999b: 24) even calls the insistence of “a rather small economy like the UK, quite open to international trade in goods and services and very open to international financial flows” to maintain a national currency “an expensive luxury—a costly way of indulging a taste for national sovereignty.”

(72.) China’s rapid accumulation of foreign exchange reserves in recent years demonstrates the constraints that the dollar peg has imposed on its monetary policy. One must note, however, that China has managed to sterilize the large increase of foreign reserve holdings very well and has thus been able to control inflation. The PBC has also made use of other monetary policy instruments, such as reserve requirements for domestic banks and credit ceilings. When announcing the raise of benchmark rates on one-year yuan loans to 5.58 percent from 5.31 percent and the rate on one-year deposits to 2.25 percent from 1.98 percent in October 2004, the PBC said in a statement that “[t]his interest rate rise… is to make bigger use of economic measures in resource allocation and macro-adjustment” (Xu 2004), indicating the country’s intention to increasingly deploy its macroeconomic policy instruments. Still, even after loosening the dollar peg, Chinese officials acknowledge the monetary restrictions posed by the exchange rate regime. For instance, Yu Xuejun of the China Banking Regulatory Commission was quoted in the Financial Times as saying that the currency system had made the PBC “passive” (McGregor 2007). For a discussion of the domestic and external constraints on Chinese monetary policy, see Goodfriend and Prasad (2006).

(73.) For details on this methodology, see Reade and Volz (2009b).

(74.) In a similar fashion, DeMartino and Grabel (2003: 269) argue in a nonmonetary context that “regionalism represents not a further loss to policy autonomy, but a means to recoup it. Working together, states might be able to pursue a policy that each desires but each lacks the ability to achieve independently. In sum, then, regionalism may undermine or enhance policy autonomy, depending on the context and on the policy area at issue.” They also maintain that “regionalism per se does not entail any sacrifice of state sovereignty” (DeMartino and Grabel 2003: 269).

(75.) On the advantages of multilateral monetary union over dollarization, see Alexander and von Furstenberg (2000) and Angeloni (2004).

(76.) On the power structure within the EMU, see Berger and de Haan (2002).

(77.) As Schelkle (2001a) points out, few would negate Mundell’s (1961) assertion that the optimal currency area is not the world. However, Mundell (1968, 2003a) himself later promoted a world currency.