Political Economy in Macroeconomics ALLAN DRAZEN ,c,, Copyright © 2000 by Princeton University Press Published by Princeton University Press, 41 William Street, Princeton, New Jersey 08540 In the United Kingdom: Princeton University Press, Chichester, West Sussex All Rights Resm;ed Library of Congress Cataloging-in-Publication Data Drazen, Allan. Political economy in macroeconomics I Allan Drazen. p. em. Includes bibliographical references and index. ISBN 0-691-01670-4 (cl: alk. paper) Macroeconomics. 2. Decision making. 3. Policy sciences. 4. Economics. I. Title. HB172.5.D73 2000 339-dc21 99-41735 account libcralization as a signal of commitmcnt to economic reform. Wc also summarizc thc empirical rcscarch on political determinants of capital controls. Another major issue in open-economy macroeconomics is sovereign debt, that is, the debt owed by a government to foreign creditors. hsucs of sovereign debt arc substantially different than those concerning nonsovereign debt and arc inherently political. For example, since it is oWed by the government, repayment decisions are not connectcd with any question of the ability to repay. With few exceptions, a borrower country has the technical ability to repay the debt, so that non-repayment is a political issue. In Section 12.8, we analyze basic models of sovereign borrowing and its repayment, especially the role of penalties in enforcing repayment. We also consider the importance of political versus nonpolitical penalties in the decision of whether to issue debt at home or abroad. The final topic considered is foreign assistance, especially lending by governments and international financial organizations to developing countries for the purpose of structural adjustment. Our point of departure is the strikingly disappointing results that foreign aid programs have had in alleviating poverty and stimulating growth in the recipient countries, a failure that we argue reflects the political nature of aid. Foreign assistance is inherently political for a number of reasons. First, the incentives of the donors may be political, not only in the obvious sense that aid is often given for strategic political reasons, but also because the nature of aid (and especially its ineffectiveness) often reflects political and bureaucratic conflicts within the donor organizations. Second, the ineffectiveness of aid is also very much attributed to misappropriation by the recipients, where this is widely believed to reflect political factors. There are a number of models which formalize the role of political factors, as well as much empirical work detailing the role of political factors in explaining the ineffectiveness of aid. PART I-EXCHANGE-RATE ARRANGEMENTS 12-2. PRINCETON PRINCETON, UNIVERSITY NEW JERSEY PRESS FIXED VERSUS FLEXIBLE EXCHANGE RATES The choice of optimal exchange rate arrangements has long been a key issue in open-economy macroeconomics. From an economic point of view, the general question may be put simply: which exchange rate arrangement is best for economic performance? The choice of the optimal regime from a, purely economic point of view is beyond the scope and purpose of this chapter. Hence, we will simply review thc key issues, suggesting that readers interestt.:d in the primarily economic arguments refer to the voluminous literature. The brief discussion of issues that arc primarily economic is meant simply as a prelude to a discussion of political issues. C II A I' T 1i N T H' I·, I. I' 1,· T II /c,' 530 · argument m · f·avor o f fi xe d n aminal exchange rates is Tl1 e main economic . . trade that they Imply. f · . t. · 1 d d transactiOn costs or mterna 10na . . t he re uce - · 1· re uncertam . Floating exchange rates arc seen as more volat1 1e, Imp ymg mo 1 This uncertainty may reduce the volume of mternarea I exc ha nge rates· · · · T · f · t tiona! trade, discourage investment, and reduce the poss1b1 1t1es_ or m ernational diversification of risks. The role of fixed exchange rates m ｾｮ｣ｯｲﾭ a ing trade may extend to trade arrangements. In the ｅｵｲｯｰ･｡ｾ＠ Umon, or ･ｾ｡ｭｰｬ＠ a fixed-rate system (or, more precisely, monetary umon) IS ｾＭ･＿＠ as ｭ｡ｸｩｾｺｮｧ＠ the gains from unified goods and labor ｾ｡ｲｫ･ｴｳ＠ ｡ｾ､＠ e Iminating the changes in competitiveness between ｣ｯｵｮｴｾｳ＠ stemmmg fr?m ersistent exchange rate movements, the latter tendmg to u_ndermme ｾｵ＠ ort for a single market. This last argument ｳｵｧｾｴ＠ an_ Important PJftical argument for fixed exchange rates, namely, as mcreasmg ｳｵｰｾｯｲｴ＠ for cooperative arrangements more ｾ･ｮｲ｡ｬｹＮ＠ We explore connectiOns between monetary and political union m Section ＱＲＮｾ＠ . , Another argument often made for fixed rates IS that ｾ･ｭｧＮ＠ one s at of a low-inflation country will help to restram mflat10nary currency to th · · d'ISCip · 1·me t o e nable a ed rates are seen as providmg r . v· .ng pressures. Hence ' fix . . government to resist the temptation to follow mflat10nary _po ICies. . Iewi f fixed rates as a commitment device brings us back to precisely the ｉｓｾｲｯ＠ commitment discussed in Chapter 5. We discuss ｴｨｾｳ＠ lfl?e of ｾｲ･＠ I _IIty ar ument below. A number of the issues discussed will mmor ､ｩｳｾｵｮ＠ in gCha ters 4, 5, and 6, so that we rely at points on ｾ＠ more mtmtive discussion, referring the reader who wants a more techmcal treatment to those chapters. . The main argument in favor of a flexible exchange_ ｾ｡ｴ･＠ IS th_e monetary policy independence it allows, implying a greater abihty to adJust to ｢ｯｴｾ＠ domestic (or country-specific) and to foreign ､ｩｳｴｵｲ｢｡ｮｾ･Ｎ＠ ｔｾ＠ co_sts ? ivin u monetary sovereignty for purposes of ･｣ｯｮ＿ｾｵ＠ ｾｴ｡｢ｩｨｺ｟ｏｮ＠ m ｾｨ･＠ ｦｾ｣Ｏｯ＠ country-specific shocks are considered explicitly m Secuon 12.4{ Mussa (1979) and Marston (1985) present good surveys ｾｦ＠ the role. o exchange rate regimes in optimal response to both domestic and foreign disturbances arguing that though flexible rates are generally presumed ｾＭｯ＠ provide ｢･ｴｾｲ＠ insulation against foreign shocks, there are cases where t IS need not be true. . h t We now turn to primarily political aspects of the chOice of exc ange ra e arrangements, which will be our focus. need not ｲ･｡ｬｾｹ＠ red,uce Ｎ･ｸＬｨ｡ｮｾ＠ A key point to note ｩｾ＠ that fixing the exchange ｲｾｴ｣＠ risk, but simply greatly alter the form of the probabihty dJstnbullon ｯｾ＠ ･ｸ｣ｨｾｮ＠ from continuous but small ｣ｨ｡ｮｧ･ｾ＠ to infrequent JUmps m the exchange rate. this below in considering currency crises. I ＮＺＬｾ｣＠ ｲｾｴ［｣Ｚ＠ r:.le I N T ,,. R N A 1' I 0 N A I. E C 0 N 0 M I' 531 Pegging the Exchange Rate to Gain Credibility Taken together with the economic arguments for flexible exchange rates, pegging the exchange rate to gain credibility for a low-inflation policy can be thought of as an application of the question of commitment versus flexibility introduced in Section 4.6 of Chapter 4 and further discussed in Chapter 5. Giavazzi and Pagano (1988) were perhaps the first to make explicit the argument for pegging the exchange rate to gain credibility. They consider a policymaker who has the same incentives towards surprise inflation that we first introduced in Chapter 4-unanticipated inflation may lower unemployment or may reduce the real value of the stock of outstanding debt. A social welfare-maximizing policymaker therefore has the temptation to engineer a surprise inflation; since this incentive to time-inconsistent behavior is understood by the public, the resulting equilibrium implies lower welfare than if he could commit himself to zero inflation. The problem is how to make such a commitment credible. Giavazzi and Pagano suggest joining an exchange rate mechanism with fixed but adjustable parities as a way to do so. They use the example of the European Monetary System (EMS), which we take as a representative of a fixed exchange rate system simply for expositional convenience? With a fixed nominal exchange rate, inflation in excess of the EMS average translates into an appreciation of the real exchange rate, a real appreciation that would not occur if domestic inflation induced an equal nominal depreciation. The real exchange rate implications of high inflation thus change the governments' incentives to inflate. One interesting aspect of the analysis is that it gives a specific real-world structure to the question of institutional commitment first raised in Chapter 5. The specific structure raises some new issues. For example, since inflation differentials between countries generally still exist in a fixed exchange rate system such as the EMS, fixing the exchange rate will imply a continually appreciating real exchange rate for the relatively high-inflation countries. Though inflation differentials may be far lower for these countries than if they did not join the EMS and let their exchange rates float, the flip side of the fixity of the nominal exchange rate is real overvaluation. If the cumulative real appreciation gets large enough, a realignment will be necessary, so that the welfare implications of commitment may be less clear. Giavazzi and Pagano consider a continuous-time model in which the domestic policymaker in the EMS chooses a level of inflation 7r1 to maximize the welfare of the representative individual, where welfare 2 There is a distinction between unilaterally pegging the exchange rate in the attempt to achieve anti-inflation credibility versus joining a multilateral fixed exchange rate system. As our main focus is on fixed rates as an anti-inflation device, we downplay this distinction here, but return to it in later sections. C II A I' r f: II 'f 11· li I. 1· t: 532 depends on inflation and unemployment (as in the Barro-Gordon ｣Ａｯｳ･ｾﾭ cconomy model considered in Chapter 4), as well as the (log of ｴｬｾ･Ｉ＠ ｲｾ｡ｬ＠ ,fined as the foreign price level over the domestiC pncc t zl' dc exc hangc rae · . . · h· t 1 1 The inclusion of the real exchange rate 1s motJvated by assummg t a ｴｾＺ･＠ domestic policymaker has the ｯ｢ｪ･｣ｴｾｶ＠ of increasing profits m ｴｨｾ＠ ex ort sector, implying a desire for a higher real exchan_ge rate .. It IS ｡ｳｾｵｭ･､＠ that there is no capital mobility and a fixed ｰｾｮｯ､＠ of ｴｮｾ｣＠ T between realignments, at which time a nominal devaluation returns ｴｾ･＠ real exchange rate to its level at the last realignment, denoted zo. With ｦｾｲ･ｩｧｮ＠ inflation assumed to be zero, the (log of the) real exchange rate at each point of time is thus given by Zo- Jt 7T, t ds, E (jT,jT + T), j = 1,2,3, .... (12.1) jT Both the return value of the real exchange rate zo ｡ｮｾ＠ the ｾｩｭ･＠ between realignments T are not choice variables of the ､ｯｾ･ｴＱ｣＠ ｰｯｨ｣ｹｭ｡ｾ･ｲＬ＠ but are determined by the EMS. In fact, if membership m the EMS IS to be ·nable these two parameters cannot be chosen independently of one sust a1 , · ·h h where another. Following Giavazzi and Pagano, we begm w1t. t e case . 0 so that at each realignment the exchange rate IS set ｢｡｣ｾ＠ to the ｾｶ･ｬ＠ ｾｯｮｳｩｴ･＠ with purchasing power parity (PPP), and consider the . . . . effects of changes in T on welfare. The domestic policymaker's objective is to ｟ｭｴｾｉｺ･Ｎ＠ a loss func!1on representing individual welfare. Given the discussion m the previOus paragraph, the domestic policymaker's loss function may be represented by ;, .t... f(j+ . l)Te -pt[ j=O JT (7T ,- 7Te) t + !!_7Tt2 2 KZ I 1dt (12.2) ' 8 > 0 3 This implies an optimal path for inflation under. adwh ere K, • . • r ptlmal justable parities (where a "tilde" above an inflation rate 1mp 1es an o value): fT peg = -1 [ 1 t 8 K p (12.3) (1 Hence, under pegged rates with regular ｲ･｡ｬｾｧｮｭｴｳＬ＠ inflation ｾｬｯｮｧ＠ an optimal path rises monotonically between realignments and then falls back to its minimum level on the date of the realignment. ｔｾ･＠ real ･ｸｾｨＺｏｧ＠ thus gradually falls from one realignment to the next, bcmg always bdow PPP, except at realignment dates. .1 llf 'cignioragc is an 1. ·ystem -ra c s · · G ·. . . · . d l'ag<lnO 3 ]so consider the possibility that the collection .Jtavaz7:1 an . . . r J ;,dJitional inn:ntivc 10 inflate. Below, we consider scigmnra!!,C tn a lXl.! I' II F I .\' I' 1; ll ;\' A I' I 0 ·' I /. 533 1; (' (} :V 0 M l In contrast, under flexible rates, Giavazzi and Pagano assume that PPP always holds and that the policymakcr cannot affect the real exchange rate z,. With exogenous z,, the policymaker chooses a constant inflation rate: 1T !lex t (} (12.4) Hence, fTrg < 1rrcx for all t < T with equality for t T. The crucial question is not whether membership in an exchange rate system with an adjustable peg implies lower inflation, but whether it implies higher welfare. One may calculate the difference in equilibrium welfare under the two regimes, given the implications of the regimes for the dynamics of inflation and the real exchange rate. After some manipulation (the exact calculations for this expression, as well as for (12.6) and (12.8), may be found in Giavazzi and Pagano [1988]), one obtains the welfare difference ｾ＠ for a permanent choice between the two regimes: (12.5) where this difference is nonnegative and uniformly increasing in T, the time between realignments, with a finite asymptote as T approaches infinity. Hence, as T increases and the system approaches one with permanently fixed rates, the welfare gain increases. In this open-economy version of the Barro-Gordon model, the incentive to inflate is the only source of inefficiency. The negative real exchange rate consequences of inflation attach an extra penalty to inflation, thus reducing the incentive to inflate. Since this disincentive is public knowledge, the low-inflation policy will be credible. Giavazzi and Pagano point out that this disincentive system makes the fixed-rate regime unsustainable in the long run. Since the return point z 0 equals 0, the real exchange rate is always below PPP, except at realignment dates. Hence, in between realignments, a high-inflation country experiences a worsening trade balance; in the absence of capital mobility, it is either steadily losing foreign exchange reserves or steadily accumulating foreign debt. Hence, it would eventually have to abandon the fixed-rate regime. The calculation in (12.5) is thus misleading for a return point of z0 0. As one alternative, they consider the case in which, given z 0 0, the fixed-rate regime is known to collapse after a time interval of length T. For temporary membership to yield a credibility benefit as discussed above, the commitment to remain in the system for an interval T must be credible, which it will be if the decision to join a temporary adjustable parity system improves the domestic policymaker's welfare. One may calculate that the l'Jil:.' 534 welfare gain relative to flexible rates of membership for an interval T is INl'l;RNAl'JONAL 535 ECONOMI' taken as given, the welfare gain from permanent membership as T approaches infinity can be derived as (12.6) where tJ. is given by (12.5). . An even harsher set-up would be one where successive devaluations. arc insufficient to make up the entire loss of competitiveness ｾｵｭｬ｡ｴ･､＠ s.mce the last realignment. That is, in the adjustable peg mechamsm, a relatively high-inflation country moves to a lower value of z 0 , so that the real exchange fluctuates around a trend of real appreciation. Since ｾｨ＠ cost of · flation is now higher in the form of persistent real appreciatiOn, the ｾｳ｣ｩｰｬｮ･＠ conveyed by joining this ｳｹｴ･ｾ＠ is h.igher.. Hence, the welfare gain from joining is also higher, a pomt ｇＮｉ｡ｶｾｺｬ＠ . and Pagano ｳｨｯｾ＠ formally by means of a simple example. The flip side IS that th.e accumu lated loss of competitiveness means ｴｨｾ＠ the ｳｹｴ･ｾ＠ ｭｾｹ＠ be ｖｉｾ｢ｬ･＠ for a shorter period of time. This is an essential trade-off m th1s set-up. a system with a greater penalty in terms of loss of competitiveness and ｾ･ｮ｣＠ loss of net foreign assets has a larger welfare Ｎｧｾｩｮ＠ (and. is hence credible) as ｬｾｮｧ＠ as a country can afford to stay in, but 1t IS sustamable for a shorter penod of time. · bl th To make a fixed nominal exchange rate permanently sustama e, e around PPP, rather. than real exchange rate must be allowed to ｦｬｵ｣ｴ｡ｾ＠ below it. In this case, a high-inflation country will be allowed a ｳｵｾ｣Ｑ･ｮｴｬｹ＠ large depreciation at each realignment that in the early part of the mterval between realignments they will run trade surpluses large enough to offset the later trade deficits. Assuming that the trade balance (and . ｨ･ｮ｣ｾ＠ change in reserves) is a linear ｦｾｮ｣ｴｩ＿＠ of ｴｨｾ＠ .log ?f !" 1 , Giavazzl ｾｮ､＠ Pagano impose the following sustamab1hty conditiOn hnkmg Z 0 and T. z = 0 ｾ＠ rT f' irrg ds dt, (12.7) T 10 10 be meaning that on the date of realignment, the real exchange rate ｭｵｾｴＮ＠ above ppp by an amount exactly equal to the average ｬｯｾｳ＠ of competitiveFrom ｾＱＲＮＷＩＬ＠ ness until the next realignment. One obvious proble.m ｡ｮｳｾＮ＠ one sees that, given T, the return ｰｯｩｾｴ＠ z 0 .is an mcreasmg ｦｵｾ｣Ａｮ＠ of domestic inflation in the interim: the higher IS the chosen path 7Tt '.the more the policymaker will be compensated ｾｮＮ＠ ｾｨ･＠ ｳｵ｢･ｾｮｴ＠ nommal devaluation. If this can be exploited, the ｣ｲ･､ｩ｢ｴｹＭｮｨ｡ｾ＠ benefits of EMS membership would vanish, as the cost of real ｡ｰｲ｣ｾｴｊｏｮ＠ ｢･ｴｷ｣ｾ＠ realignments would be offset by the ｡､ｪｵｳｴｾ･ｮ＠ at . rcah.gnment. ＨＧＡｨｩｾ＠ problem is simply an analogue of the trade-off Issue raised m the previous paragraph.) Giavazzi and Pagano therefore suggest that for ｅｾ＠ Ｎｭ･ｾｲﾭ ·h'p to be sustainable in the long run, but still yield crcdlhlhty ｧ｡ｭｾＬ＠ 1 . , . )able Jf Z IS S (l 7) , · domestic policymakcrs cannot view z 0 m 2. as mampu · n lim tJ. + --';(1- T --><n (}p ｾＩＬ＠ p (12.8) where tJ. is given by (12.5) and the last term is the discounted value of the initial competitiveness offset. When K/ p > 1, meaning the discounted value of the penalty for higher inflation exceeds the incentive to create inflation surprises, Giavazzi and Pagano argue that the EMS regime is no longer unambiguously superior. Though this may sound counterintuitive, an example will help explain why. As T----) oo, (12.3) implies that irrg is negative forK/ p > 1. Whereas this causes no problem in the earlier set-up with no competitiveness offset, it now implies that a country starts with a real appreciation, that is, with a competitiveness penalty to ensure long-run sustainability. If the cost associated with this penalty exceeds the benefit associated with reducing the temptation to inflate, the fixed-rate regime may be inferior to flexible rates. Giavazzi and Pagano suggest a similar argument will apply when T is finite. I"evocably Fixed Exchange Rates? The sustainability issue that Giavazzi and Pagano raise has led some authors to question the commitment value of fixed rates. Tornell and Velasco (1995) argue that viewing fixed exchange rates as not irrevocably fixed may significantly change our view of credibility-enhancing properties of fixed rates. They in fact argue that flexible rates may be a stronger disciplining device than fixed rates. If the exchange rate cannot be fixed forever, fixed rates simply postpone the inflationary consequences of overexpansionary fiscal policy, while flexible rates mean that some of the inflation cost comes today. Bearing the costs currently may put a stronger constraint on fiscal authorities if they care about welfare primarily during their terms of office, rather than after. Exchange Rate Pegs under Incomplete Information The Giavazzi and Pagano model considers disciplining a policymaker with a known incentive to inflate. It is thus like the complete-information models considered in Chapter 4 and Section 6.3 of Chapter 6. An alternative is a model of incomplete information about a policymaker's preferences. Andersen and Risager (1991) consider a mimicking model of the sort presented in Section 6.4 of Chapter 6 to study the possible contractionary effects of pegging the exchange rate as a disinftationary device. As in the Backus and Driffill (1985a) model considered there, there are two C II .I I' 1' li II r 1' II' E /. I. 10 536 types of policymakers: a tough type who assigns no loss to unemployment, caring only ahout reducing in1lation, which is equal to the rate of exchange rate depreciation in their small open economy; and a weak type who assigns a loss to both unemployment and inflation, and who therefore sees a gain to an exchange rate devaluation as a means to increase employment. Uncertainty about the policymaker's type means that the announcement of a fixed exchange rate policy with the goal of price stability is not fully credible. The government's reputation is updated using Bayes' Law (see (6.9)), and over time the probability that the government is tough rises, implying a fall in expected inflation and in domestic interest rates. However, unlike the Backus-Driffill model, Andersen and Risager assume there is a positive probability that a weak type will abandon the peg and inflate (that is, devalue) in the first period. (In the model in Chapter 6, a weak type was certain to masquerade as tough with a sufficiently long horizon.) This makes a crucial difference. If the public is sure that a new government will choose zero inflation for a period of time at the beginning of its tenure, inflationary expectations will be zero immediately after the new government takes over, so that the move to zero inflation can be achieved with no output loss in the short term. In contrast, if there is a positive probability that a new government will initially inflate and break its fixed exchange rate promise almost immediately, the disinflationary fixed exchange rate policy will lead to a drop in output in the short term. Under this scenario, long-term interest rates will decline by less than short-term interest rates with the implementation of a program, reflecting its initial lack of full credibility. 12.3. CURRENCY CRISES AND CONTAGIOUS SPECULATIVE ATTACKS In practice, countries that have chosen to fix their exchange rates often find their currencies coming under attack by speculators hoping to profit from a devaluation or a move to floating exchange rates. On the surface, a speculative attack appears to be a pure economic event, motivated by profit seeking by speculators, independent of political overtones. In fact, there are several political aspects to currency crises. The decision of whether or not to defend an exchange rate peg is a decision about trading off conflicting objectives. The decision to attack in turn depends on the credibility that speculators assign to the government's announced commitment to the peg. The importance of credibility leads to the possibility of self-fulfilling currency crises, whereby a given level of fundamentals may he consistent with either the maintenance of the fixed exchange rate or with its collapse. In this case, the expectation that the government is committed to defending the currency leads to this being an equilibrium and vice versa. Contagion in speculative attacks also has political aspects, not only because the decision to defend is political, but also because it may II h I ,'\i r J:' /l ,'v' A 1 I 0 ' A I. 537 lc' C 0 /II 0 M l' he factors that link curre ncics . together so that c · · ac · political , 1 I oss t lem. We consider these . . . . . turn. , nsJs spreads · Issues 111 Political Models of Speculative Attack !he ｳｾｭｩｮ｡ｬ＠ paper on speculative attack i Kru mconsJstency in fundamentals induces a s ｧｭ｡ｾ＠ . . (1979), m which an an abandonment of fixed r t F steady loss 111 reserves, ending in Ｐ deficit, and is financing ｾ＠ ebs. ｾ＠ ｾ｡ｭｰｬ･Ｌ＠ the government is running a t f . . y pr111tmg money Th expansion IS inconsistent with the fixed . e . ra e o monetary the short run, individuals do not want exchange ｲｾｴ･＠ m the long run; in currency, and they exchange it f f to _hold the htgher level of domestic The peg rate must be aband ; ｯｾ･ｴｧｮＭ｣ｵｲｹＮ､ｭｩ｡＠ assets. Ｚ｡ｲｾｮ＠ ｲ･ｳｾ＠ htt a minimum level, which is common knowledge ｴｾｮ｡ｬ＠ However, the peg collapses not at the date implied b sim let ｰ｡ｲｴ｣ｾｳＮ＠ of reserves, but in a speculative ｾｴ｡｣Ｏ＠ ; extrapolatt_ng the steady decline ･ｾｲｨ＠ date, namely, the first date at which optimal investor b : ｾｯｭ｟･＠ succeed. e avtor tmphes such an attack will "first-generation" model r . . In ｾｵｧｭ｡ｮＧｳ＠ 111g wtth current mutually inconsistent Ｇｴｾ＠ Icymakers are ｾ｡ｳｴｶ･Ｌ＠ stickrate reflexively when the critical . . po ｾ｣ｩ･ｳ＠ and abandonmg the fixed approach to modelin curren ｭｴｵｾ＠ evel is reached. An alternative d ｾ＠ _collapsfle IS to treat it not as inevitable, but as the result of a ｰｾｬｩ｣ｹ＠ eCISIOn re ecting the . ht h places on the objective of m . t . . wetg t e government objectives. That is a fixed ratam atlnl mg a fixed ER, relative to other · f ' e co apses not because ·t · t h · 111 easible to maintain it· as i Kru t IS ec meally longer sees it as ｯｰｴｩｾ｡ｬ＠ i: r ｾｵｴ＠ ｨｾｭ｡ｴＬ＠ because the government no importance of other objectives If t e costs of doing so and the "second-generation" m d I d . ln . s approach, often referred to as a o e ' eva uatmn IS a politic l d . . t h e balancing of conflicting objectives. M . a ecision, reflecting modeled as having an obJ·ective fu t' ore ｾｰ･｣ｩｦ｡ｬｹＬ＠ the government is . nc mn With more target th . . . s an mstruments, as m the closed model of an extensively in Part II of the book or the ＺｵｴｰＭｭｾ｡ｯｮＮ＠ trade-off used the previous section. Examples include ｏ｢ｾ］Ｚ､ｯｦ＠ Giavazzl and Pagano in son (1994) Masson (19 ) k d (1994), Drazen and Mas95 ' 0 z an an Sutherland (1995) Ob f ld ( ) ' an d Bensaid and Jeanne (1 997). 1996, ' st e t: In these models, speculator behavior . . tions of a devaluation rationally d'f IS ｾｵｭ｡ｮｺ･､＠ by their expectaｲｾｳｰｮ･＠ to a single' underlying ｏＺｯｾ･､＠ on the government's optimal dtstnbution of that shock Draz d n on the ｣ｯｾｭｮ＠ knowledge in detail in Section 6.S of Cha ･ｾ［Ｉ＠ Masson (1994) (whtch was discussed the policymaker's with uncertainty about fundamen , , an cons! er ow It will IS persistence in the effectS of }' tals. ｾｨ･ｹ＠ show that when there defend a fixed excha po tcy across pcnods, tough policy meant to ngc rate may actually make the peg less credible, as it ｾｮｴ･ｲ｡｣＠ ｾ｢ｯｵｴ＠ ｯ｢ｪ･｣ｾｶｳ＠ ＨｨｩｾｮＧｴｹＺ［ｯ＠ ｾＹＵＩ＠ ｾ､＠ ｾｮ｣･ｲ｡ｩ＠ CHAPTER TWELVE 538 is known to worsen the future trade-offs policymakers will face. Shocks that are not offset through a devaluation today have further unfavorable effects in the future, increasing the probability that a government of any type will devalue. The resulL is descriptive of the sometimes failed attempts of governments to defend a fixed exchange rate via displays of toughness. Consider, for example, the experience of the United Kingdom or of Sweden in their attempts to maintain a fixed exchange rate in 1992 and the response of investors to their displays of toughness. In the case of the United Kingdom, in August and September of 1992, Prime Minister Major and Chancellor of the Exchequer Lamont vowed to defend the pound vigorously as speculative pressures mounted. Despite their tough statements, and the increases in short-term rates to back them up, reserves flowed out, with it being reported that the Bank of England may have spent half of its reserves in a single day in an unsuccessful attempt to defend the pound. In Sweden, when there was heavy pressure against the krona in September 1992, the Prime Minister and the Finance Minister both emphasized that they would defend the currency at any cost. Overnight interest rates were raised first to 24%, then to 75%, and finally to 500%, but the Riksbank had to engage in massive borrowing to meet the demand for foreign exchange. Though the September crisis passed, in November there was renewea speculation against the krona, with large capital outflows, followed by the Riksbank allowing the krona to float. In both cases, tough policy meant not simply to defend the exchange rate, but also to send the message that policymakers were firm in their commitment to the fixed exchange rate, appeared to have the opposite effect. One may also expand this approach, in analogy to the closed-economy work on reputation, to models of speculators' optimal behavior in which they solve a more complicated, dynamic signal extraction problem in which there are several types of shocks. Bensaid and Jeanne (1997) and Drazen (1999b) present models in which the probability of devaluation is derived via Bayesian updating on the policymaker's type. Self-Fulfilling Currency Crises THE INTERNATIONAL 539 ECONOMY The level of fundamentals will determine whether there are possible equilibria or not.4 multiple A Model of Multiple Equilibria and Escape Clauses Obstfeld (1996) presents a simple model of multiplicity, based on the trade-off ｾ･ｴｷｮ＠ ｾ･＿ｉｰｬｯｹｭｮｴ＠ and inflation that we have used re eatedly. eqmhbna can arise due to a fixed cost of a or ｲｫ･ｶ｡ｬｾＮｴｩｏｮ＠ of the .exchange rate. Equilibria differ by the degree of market s eptic1sm regardmg the exchange rate peg and the · 1· t' f d" . ' lmp lCa Ions or unempl t oy_rnen , con ｾｴｯ｡ｬ＠ on the maintenance of the current parity. ｈｾｮ｣･Ｌ＠ d1fferent ｾｱｵｩＱ｢ｮ｡＠ are associated with different probabilities of ｾｯ＠ ｬ｡ｾｳ･Ｎ＠ He cons1ders a small open economy, where the exchan e rate is to the domestic price level, so the change in the rate is ｾｲｭｰｬ＠ the rate of domestic inflation or deflation. In addition to the ｭｦｬｾｴｩｯ＿＠ an.d possible unemployment costs, changing the exchange rate panty 1mphes a ｣ｾｳｴ＠ to the government ( ( 7T) = ( for a devaluation (an upward change m the exchange rate) and ((7T) = ( for a revaluations Hence, the government's objective function may be written · ｍｾｬｴｩｰ･＠ ､･ｶ｡ｬｵｴｩｾｮ＠ ｬｾ･ｊ［Ｎｴｩ｣｡＠ ･ｸ｣ｨ｡ｮｾ＠ (12.9) where both tpe unemployment rate U and the government's unem lo _ ?Ient target U are measured relative to the natural rate (so that 6 < y 1s the rate of devaluation of the exchange rate (or revaluation if 7T < and 8 > 0. Unemployment is determined by ' ｾＩ＠ U= (7Te- 7T) + e, 'o) (12.10) inflation .expectation of domestic wage setters based on where ｾ＠ e is ｴｨｾ＠ lagged mformat10n and e 1s a mean-zero i.i.d. shock which is observed b it chooses 7T, but by wage setters only after wagZs the government are set. For tractab1hty of computation assume that · 'f 1 distributed over [- v, v ]. ' e lS um orm Y n.o fixed-cost term (( 7T ), the rate of devaluation would be . If there ｷｾｲ＠ ｾｨｯｳ･ｮ＠ ｴｾ＠ m1mm1ze (12.9) subject to (12.10), yielding an optimal rate of evaluation (or revaluation) as a function of the unemployment shock and ｢･ｦｯｾＮ＠ Models of self-fulfilling attacks, as in Obstfeld (1994, 1996), provide 4 Ｎｾ＠ ｾｯ､･Ａ＠ based on an optimizing government is not identical to one · h · another political link, due to their emphasis on problems of credibility. In equthbna and the resultant possibility of self-fulfilling crises A "new" crt"st· dwtlt mhulttple · "l'b · · · s mo e can ave a models of self-fulfilling crisis, a given level of fundamentals may- be umque equtt num,. as m Drazen and Masson (1994) whereas a nono f · · consistent with either the maintenance of the fixed exchange rate or with have multiple equilibria, as discussed by Krugman (19g6). p tmtzmg model can its collapse. The speculative attack equilibrium thus has a self-fulfilling as ::he ｣ｯｾｴ＠ ｯｾ＠ de:iating from the fixed parity make this model an "escape-clause" model . ｴｳ｣ｵｾ＠ m ｾ｣ｴｯｮ＠ 4.6 of Chapter 4. The alternative formulation of an esca e clause ｡ｾ＠ characteristic, in that the exchange rate collapses if attacked, but survives m SectiOn 6.8 of Chapter 6, yields a unique equilibrium, as indicated in ｾｨ･＠ ｰｲ･ｶｩｾｵｳ＠ if speculators do not attack. Hence, fundamentals do not unambiguously ｾＺ［Ｎ＠ determine whether a peg can survive or not; market beliefs are critical. CHAPTER TWELVE 540 the expected rate of inflation: 7T 1 - --(€- U = 1 + (} + (12.11) 'TTe). Substituting (12.11) and (12.10) into (12.9), one obtains the loss under discretion (i.e., flexible rates) as a function of the unemployment shock and the expected rate of inflation: (} - .:.?'flex= - - ( e U o 1+ 2 (12.12) + 7r•) . When the fixed rate is maintained, so that 7T = 0, the loss is (12.13) If there is a fixed cost of changing the exchange rate, the_!e will be a devaluation only if the shock e is large enough that 2"ftex + { < yfix and. a revaluation only if e is small enough that 2"ftex <2"fix, in which ｣｡ｾ･＠ the change in the exchange rate willbe determined by (12.11). That ts, ､･ｶ｡ｬｾｴｩｯｮ＠ occurs when e > e and revaluation when E < ｾＮ＠ where +; € ir- ｾ＠ = U 'TTe 'TTe + ((1 + 0)()1/2 (12.14) ((1 + O)f)l/2• Given this escape-clause rule, the rational expectation of 7T next period is E(7r) = E(7r!E < ｾＩｐｲＨｅ＠ THE INTERNATIONAL ECONOMY 54 Specifically, for a low enough 7r•, ｾ＠ > -u, so that € ｾ＠ is independent c 'TT•. One may calculate dE(7r)/d7re = 1/(1 + 0). Similarly, for a hig enough 'TTe, e is at the lower limit - v, so that the government's behavior: described by (12.11) and dE(7T)jd7T• 1/(1 + 0). (These correspond t the unique Barro-Gordon solution.) However, for "intermediate" valm of 7T e' ｾ＠ = - v' e is interior to [- u, v] and is described by (12.14), so th: E(7T) is a quadratic function of 'TT•, implying the possibility of multip intersections of (12.16) with the rational expectations line E( 7T) = 7T Each of these intersections is an equilibrium. Intuitively, if markets expect a low probability of a devaluation ar hence a low 'TT•, a given (positive) unemployment shock is consistent wi1 lower unemployment under fixed rates, so that the government is le likely in fact to devalue. Conversely, high devaluation expectations imp higher unemployment for any realization of e, making the governme1 more likely to incur the fixed cost { and devalue to improve welfare. Eac of these expectations may be self-fulfilling. It is easy to show that multiple equilibria exist, the devaluation expectation in the highest prob bility of devaluation equilibrium is equal to that under floating rates. Among other things, the Obstfeld model provides a strong cautiona note on why fixing the exchange rate may not buttress the credibility of government with an incentive to inflate. It may instead lead to self-fulfi ing currency crises if the expectation of a devaluation remains high. Tl credibility-reducing effect of fixed exchange rates in this model is differe than that stressed by Drazen and Masson (1994), in which there is unique equilibrium. There, keeping with a fixed exchange rate today m so worsen the expected short-run trade-off tomorrow between inflatic and unemployment that it makes the policy less credible. Alternative Models of Multiple Equilibria < ｾＩ＠ + E(7riE > e)Pr(E > €), (12.15) which, using (12.11) and the uniform distribution for e, may be written c:) - ｺＴｵｾｬﾷ＠ e-- - (7Te- U)- - - 2v (12.16) In a rational-expectations equilibrium, E( 7T) = 'TT•. In the. absence ｯｾ＠ the cost ((7r ), this would be determined simply by (12.11), ｬ･｡､ｭｾ＠ to a umque devaluation expectation in equilibrium, namely, 'TTe - U /0, and an associated actual change in the exchange rate for each realization of the shock e. This is precisely the Barro-Gordon solution, derived in Section 4.4 of Chapter 4. The presence of a fixed cost, however, leads to an of 'TT.". Ｈｒ･ｭｾﾭ expectation E( 7T) in (12.16) which is a nonlinear ｦｵｮ｣ｾｩｯ＠ ber that, via (12.14), € and ｾ＠ are themselves decreasmg functions of 7T .) As Obstfeld (1996) points out, there are several other mechanisms whi could lead to multiple equilibria. Multiplicity of equilibria is inherent government debt or other assets if the return that one investor earns is increasing, rather than a decreasing, function of the amount that otl investors choose to invest. This phenomenon is important in the analysis banking panics, as argued by Diamond and Dybvig (1983). Consider a ba that holds less than 100% reserves against deposits, whose redempti price is not allowed to vary. The belief that there will be a bank run c then be self-fulfilling on the part of rational depositors. If a deposi believes that a significant number of other depositors will withdraw th deposits, so that the bank must suspend convertibility, he will choose withdraw his deposits as well. If many depositors share the initial belief will be self-fulfilling. This idea may be applied to government debt if the attractiveness holding the debt depends on the probability that it will be repaid in f1 542 • • r: CHAPTER TWELVE whtch m turn depends positive! on h uiret ｴｾ＠ amount of ｾ･｢＠ which is held. In the simplest case, this would cost rise more than proportionally to the amot!t of d ｾ＠ the ｲｾｰｵ､ｴｩｯｮ＠ to be repudi· ated. For example, suppose it was belie: t that IS ｡ｶｾｉｬ｢･＠ would be imposed only if th . ed that sanctwns for repudiation repudiated was above a certain level. It would then be optimal ｡ｾｯｵｮｴ＠ enough other investors were doi! government debt only if one believed ｭｵｬｴｾｰ･＠ equi!i?ria. The key ｦ･｡ｴｵｲｾ＠ Ｚｾ｡ｲｮ･Ｌ＠ leading to the possibility of multiple eqmhbria in financial k s .example, as for many models of individual investors perceive m· c rna! ets, IS that the set·up is such that . reasmg returns to · of aggregate mvestment Tbes . . mvestment as a function .. . · e mcreasmg retur b t at they arise because of some d . . ns are po1Ittcal tn the sense eciston on the part of government affecting returns. Calvo (1988) applies a similar idea to the . . ､ｾｴＺｭ｡ｩｯｮ＠ of the interest rate on government debt to generate th ｰｯｳｴ｢ｩｬｾｹ＠ of n:uitiple equilibria. He considers a two-period model in ｷｨｾ＠ can be partially repudiated (ore uival ch debt. Issued m the first period proportional to the size of the ｾ･ｦ｡ｵｮＬ＠ ｰ｡ｲｾｉｬｹ＠ taxed away), at a cost lowers the effective mterest rate actuaiiy received i tb . ･ｰｵ､ｴ｡ｾｯｮ＠ penod for any interest rate contracted in the first period Hn e ｾ｣ｯ､＠ ｾ｢｡ｴ＠ investors perceive, the ｧ［･｡ｾ｣ｨ＠ t e higher. the probability of default m the first period. Calvo makes th f . e contract mterest rate they demand ｴｾｨ･ｲ＠ ｲ･｡ｳｯｮｾｉ＠ assumption that the probability of default depends ?overnment being more likely to ､･ｦ｡ｵｬｾｯｶｲＨｮｭｴ＠ s budget position, the . Its debt the higher its level of e ｯｾ＠ or tax ｾｷ｡ｹ＠ the proceeds of) component of government outla s t xpend.ttures. Wtth debt se!Vice one of debt repudiation in the second period is an increasui'g'f he ｾｰｴｩｭ｡ｬ･ｶ＠ Hence, interest rates depend ｯｳｩｦｵｾ｣＠ on of the first-period Interest rate. turn depends positively on int!est ｾＺ＠ y o; expected repudiation, which in values, one gets two equilibria, one in which es. or. reasonable. ｰｾｲ｡ｭ･ｴ＠ the other in which the debt is rt' II there ｾ＠ no repudtatwn of the debt ' Tb pa ta Y repudtated ｾｲ･＠ are other mechanisms for multi le ... . . . hmgmg on expected effective nominal interest rates High p ･ｾｵｩｬｴ｢｡＠ pressure on financial intermed' : . . nommal mterest rates will put . Iartes mcreasing th e pressure for a govern- . ment bailout. Increased expectati , f th lender of last resort and the .ons o e government stepping in as · . • associated monetary . ･ｾ｡ｮＮｳｴｯＬ＠ thus translate mto mcreased expectations of a curr market expectations. 6 Though em iricaency_ depreciatiOn, thus ratifying on self-fulfilling crises due to. credibility problems is stift re:i ･ｾ､ｮ｣＠ suggestive studies. Kaminsky and Rp 'nhmmary, there are a number of crises and currency crises are stron I ei. art ＨＱＹ＿ｾ＠ suggest that banking ｾｯｮ＠ (1994) find that ERM credibili; ｾＮｮｫ･､＠ emp.trtcally. Rose and SvensJust before the currency crises h' th ｾｯｴ＠ detenorate until August 1992, ' w IC Is taken as suggesting that the io o: 6 0 h r, t e government may try to avoid a bailout by a quick devaluation. THE INTERNATIONAL ECONOMY 543 pre·crisis parities were not viewed as unsustainable by market participants until the crisis occurred. Contagious Currency Crises Contagion of currency crises across countries has become a major focus of research, propelled in part by a number of seemingly contagious crises in the second half of the 1990s. Existence of contagion appears to be supported by solid empirical evidence/ and there is significant effort being devoted to construction of convincing theoretical models of contagion. Beyond starting with "second-generation" models of currency crisis stressing a trade-off of objectives, most current research on contagion gives little or no role to explicitly political factors. "Contagion" is the phenomenon of a currency crisis itself in one country making a currency crisis (or currency weakness) in another country more likely.8 The emphasis is meant to differentiate true contagion from a common shock (other than a currency crisis) which affects countries differentially because of their differential susceptibility to infection. When differential vulnerability to an unobserved common shock reflects unobserved characteristics, we may get what looks like true contagion, since a crisis in one country will be followed by a crisis in another, with no apparent explanation other than the original crisis itself. Drazen (1998b) presents a model of intrinsically political contagion, in that the objectives which give rise to contagion are primarily political. In the absence of the political objective, devaluation in one country would not affect speculative pressure on another country's currency. Specifically, he considers a policy of holding the exchange rate fixed at a significant economic cost in order to further the objective of political integration. As discussed in Section 12.4 below, monetary union has often been adopted for political objectives, the move towards European Monetary Union being but the most recent example. One may think of membership in a "club," 9 whether explicit or implicit, where the benefits of membership are heavily political and the condition for membership is the maintenance of a fixed exchange rate. To obtain contagion, one must make the further assumption that .the · on who else is value of membership in the arrangement depends ｰｯｳｾｴｩｶ･ｬｹ＠ or may be a member. Hence, if a country learns that other potential See, for example, Eichengreen and Wyplosz (1993), Gerlach ｡ｮ､ｾ＠ Smets (1995), and, especially, Eichengreen, Rose, and Wyplosz (1995). 8 Masson (1998), in contrast, argues that the term "contagion" should be applied only to cases where a crisis in one country triggers a crisis elsewhere for reasons unexplained by macroeconomic fundamentals, suggesting "spillover" be used when a crisis in one country affects the fundamentals in another country. 9 The term "club" here is not used in the sense of Chapter 9, that is, primarily as a mechanism for provision of public goods. 7 I' 545 544 T II li members of the arrangement place less weight on meeting the conditions required to join, and hence arc less likely to participate, it will find it less advantageous to join as well. It will therefore assign a lower value to maintaining a fixed exchange rate, especially when doing so requires sacrificing domestic goals. If speculators arc uncertain about the value a country places on membership in the "club," but arc aware of both th_e no-devaluation condition for membership and the dependence of this value on who else is a member, rational behavior on their part will then imply that a successful attack on one currency creates an externality in the form of a lower commitment of all other potential members. They will therefore be more vulnerable to attack, a phenomenon the paper terms "membership contagion." The paper further presents empirical evidence that membership contagion may have played a role in the 1992-1993 EMS crisis. politics, rather than simply the economics, of these measures. From an economic point of view, joining a monetary union would imply a weaker commitment than dollarizing or adopting a currency board, as devaluations to deal with external imbalances are often allowed. By giving up control of the currency, all three measures largely remove from public debate the possibility of a devaluation in response, for example, to a real appreciation. One way in which they are thought to differ is in their implications for sovereignty.•Joining a monetary union implies an apparent loss of some national sovereignty in making policy, while adopting a currency board does not (at least explicitly) appear to cede sovereignty to other countries. On the other hand, to the extent that a country retains bargaining power within the monetary union, it can still affect its own monetary policy, however partially. Dollarization is seen as providing more credibility than a currency board since it is harder to undo the arrangement. Drazen (1999c) argues that both measures should be seen as examples of constitutionalism, as they make it difficult to revisit or reverse policy decisions. 12.4. MONETARY UNIONS I lV F E R lV A T I 0 S A I. /:' C (} N 0 M Stronger Institutional Arrangements Optimum Currency Areas One theme of the discussion so far is that comparing the political benefits of fixed versus flexible rates may be of limited relevance, as fixed exchange rates are, to use Obstfeld and Rogoffs (1995) term, a "mirage." The credibility-enhancing properties of committing to fixed rates require that fixed rates themselves be credible, with the discussion in previous sections suggesting a number of reasons why this may be open to question. The massive losses that central banks have suffered in trying to defend a fixed exchange rate when their commitment was less than fully credible suggests how costly fixed rates may be. As Obstfeld and Rogoff argue, the incidence of speculative attacks suggests that even systems with exchange rate bands pose problems so that, in their words (1995, p. 74), "there is little, if any, comfortable ground between floating rates and the adoption of a common currency." In short, since the announcement to maintain a fixed exchange rate need not itself be credible, countries that view it as crucial to demonstrate their commitment to fixed rates may opt for mechanisms that are stronger than simply turning exchange rate management over to an independent central bank that has an announced commitment to fixed rates. Basically, this means giving up control of the domestic currency-by joining a monetary union with other countries, by replacing domestic 10 currency by foreign currency ("dollarization") as the medium of exchange, or by adopting a currency board requiring that domestic currency be 100% backed by foreign currency assets. As in previous sections, we consider the In a monetary union, countries adopt a single currency with a single central bank (and no internal exchange rates). As in the case of fixed versus flexible exchange rates, it is useful to distinguish the primarily economic arguments for a single-currency area from the more political ones. In discussing political-economic arguments about monetary unions, there is obviously significant overlap with the discussion of fixed exchange rates. When the argument is simply the same point but stronger, it will not be repeated. For example, if a credibly fixed exchange rate is a way to tie one's hands to gain anti-inflation credibility, giving up a single currency and joining a monetary union can be seen as credibility enhancing, only more so. The economic arguments for a common currency are usually discussed in terms of optimum currency areas. The concept of an optimum currency area (OCA) was first developed by Mundell (1961), with other important early contributions being McKinnon (1963) and Kenen (1969). The move to European Monetary Union (EMU) has revived interest in optimum currency areas, with numerous recent papers. (See, for example, Bayoumi [1994], Bayoumi and Eichengreen [1996], and Frankel and Rose [1997]. Melitz [1995] presents a critical analysis of the state of research.) There are a number of factors that are central in determining the economic desirability of a common currency area. First, there is the extent of trade between the countries; the higher the volume of trade, the greater the saving in transaction costs resulting from the adoption of a common currency. Second, there is the question of how similar the countries are in the shocks they face and in the economy's response to these shocks. A 10 Panama is the standard example of a dollarized economy, hut the original decision to adopt the dollar as its currency was not an anti-inllation measure. 546 'CHAPTER· TWELVE common currency means adopting a common monetary policy; the loss in flexibility has low cost if the optimal monetary policy is the same across countries.U Third, similarity is .less important the greater is the flexibility of wages and prices, so that the need for relative price adjustment can be accommodated by movement in domestic prices rather than exchange rate changes or increases in unemployment. Fourth, in a similar vein, the greater is the mobility of the labor force across countries, the better candidates the countries are to form a monetary union, since differential shocks within the area will then result in labor movement rather than unemployment. Finally, there is the question of the extent to which there is a system of compensatory fiscal transfers across regions differentially affected within the union (sometimes termed "fiscal federalism") and, more generally, the possibilities for risk sharing. (We return to this issue in Section 14.6 of Chapter 14.) Bayoumi and Eichengreen (1996) present a recent survey of these issues. The Connection between Monetary and Political Union Monetary sovereignty may be prized by a country as a sign of political sovereignty. As John Stuart Mill put it in an oft-quoted phrase, nations "assert their nationality by having, to their own inconvenience and that of their neighbors, a peculiar currency of their own." Political aspects of monetary sovereignty and monetary unification is both a classic topic and a very active area ofcurrent applied research in international economics, the latter beqause of EMU. We postpone a discussion of the politics of EMU until the very end of this section, after we have considered the broad range of research on the political economy of common currencies. Related to the question of the connection between monetary . and political sovereignty is the question of whether monetary union may be seen as a way to achieve political goals. Nineteenth-century Germany and Italy are often taken as examples, though the case is far from clearP Bordo and Jonung (1997) point out that political unification in Italy was completed in 1861, but financial arrangeme9ts were quite disparate prior to this, with currency unification occurring only in 1862. This is similar to the United States, in which monetary unification. followed soon after political union. They argue that in Germany, both political and monetary unification proceeded stepwise, with scholars disagreeing about when the most important step towards monetary unification occurred. floltfereich (1993) suggests it was unification of coinage in 1857, while Kindleberger 11 Frankel and Rose (1997) correctly point out that these criteria are not .. independent, since the correlation of business cycles across countries depends on the degree of trade intefration. . · · 1 We return to economic determinants of the political union or breakup of nations in Chapter 14. THE INTERNATIONAL ECONOMY 547 (1981) and others view the creation of the Reichsbank in 1875 as the crucial step. Hence, it is unclear whether monetary unification preceded political unification in 1871, or vice versa. In all these cases, one can argue that the creation of a national monetary union was closely associated with the creation of independent nation-states (or with reunification, as in the recent case of the former West and East Germany). All three national monetary unions were primarily arrangements to reduce seigniorage competition, standardize coinage, and set up a national unit of account. There may be effects working in the other direction as well, from politics to the sustainability of the exchange rate arrangement. A difference between a monetary union and a fixed exchange rate (or even a currency board) concerns reversibility. Policy pronouncements notwithstanding, fixed exchange rates are never irrevocably fixed, as discussed in the previous section. Because leaving a true monetary union or a dollarization arrangement would mean reinventing a separate currency and central bank, that is, creating or resuscitating institutions, the commitment is more permanent. (It is for this reason that a currency board is more reversible than dollarization or a monetary union.) The breakup of strictly defined monetary unions has occurred because of the breakup of the associated sovereign state, never simply for monetary reasons. In contrast, those (multinational) monetary unions that have collapsed without the dissolution of the associated sovereign entities typically were not full in the sense given above, in that members maintained separate monetary authorities. This, in fact, facilitated the dissolution of the unions once the member countries were subject to large shocks. Cohen (1993) suggests another link from politics to the sustainability of a currency union. He examines the history of a number of currency unions and argues that monetary cooperation within the union is most likely to be sustained eith.er when there is a single dominant member who is willing and able to use its influence to sustain monetary cooperation or when there is a broad network of institutional linkages between members such that the loss of monetary autonomy is outweighed by the gains of cooperation. These two possibilities have obvious relevance for EMU. Neumeyer (1998) suggests a different sort of link from politics to economics in a monetary union. The primary purpose of his paper is to give a choice-theoretic basis to the argument that an important benefit of a common currency is reduction of exchange rate uncertainty. Though it is widely believed that "excessive" exchange rate volatility reduces welfare, this is not easy to prove in a formal setting. The problem is that fluctuations in exchange rates in response to economic shocks (such as preference and technology shocks) generally are welfare improving relative to less exchange rate volatility in the face of the same shocks. However, exchange rates fluctuations caused by certain other, "nonfundamental" factors will indeed reduce the efficiency of financial markets. 548 CHAPTER TWELVE THE INTERNATIONAL ECONOMY 549 ;here Trf:u is the expected common rate of inflation and is a common 71 uropean shock. The common European rate of inflation will beB To study the benefits .of reducing exchange rate volatility, Neumeyer uses a general equilibrium model with incomplete asset markets to study the implications of a monetary union for hedging of risks; specifically, he is concerned with the benefit of the reduction of nominal volatility compared with the cost of reducing the number of currencies with which to hedge. Political influences on monetary policy, say in the timing of inflation stabilization, are a prime example of "nonfundamental" factors causing exchange rate volatility.. The possibility of political interference implies that, given the realization of an economic shock, there is still uncertainty about monetary policy, as it will be influenced by future political events. The excess fluctuations in price levels that result are socially costly because they "contaminate" the real payoffs of nominal financial assets, thus reducing the ability of investors to use these assets to hedge against economic shocks. He thus argues that currency unions and permanently fixed exchange rate regimes may improve welfare by insulating money from domestic politics. TrEu =- 1 1 ｾ＠ -UEU + - - 7 1 1+ e e · (12.19) f ｅ｡ｾｨ＠ n:emb:r country i evaluates the policy on the basis of its own loss unctwn, Identlca! to (12.17),_ but with e replaced by (Ji and tJ; re laced ｾｹ＠ a ｣ｾｮｴｲｹＭｳｰ･ｩＺｦ＠ ｾ｡Ｚｧ･ｴ＠ lf;, and with country-specific ｵｮ･ｭｾｉｯｴ＠ U epen mg. on. unanticipated European inflation and a count -s ･｣ｩｦｾ＠ shock Ei, y1eldmg an expected loss: ry p E.2j = E[(Treu - TrEu Monetary Union without Political Union-Policy Conflicts ＬｓｉＲＮｾＩ＠ If a group of countries form a full political and monetary union, so that not only is policy joint but so is the evaluation of policy, the issues within the union are basically identical to the monetary and fiscal policy issues in · a closed economy. An alternative is the existence of a monetary union without a political union, so that there is a common monetary policy, but differences across countries in what the desired policy should be. + €· • fl)2 + (J.Tr2EU 1• t (12.20) , wher.e E is the eXJ?ectation operator. Hence, a country with 6. < e uts ｲ･ｬ｡ｴｶｾｉ＠ more weight on unemployment fluctuations than the' Eurotean ｾｴｲ｡＠ hank and thus has a higher temptation to inflate. When countries ｾｮｯ＠ ｃｾＩＺ､＠ loss for membership can be found by substituting E£':mem = E[(€ _ _!__ ｾＩ 1 + e 7l- U; .. • Ｒ＠ + i ｾ＠ )2] . e1uEu ( 1 6; 1'+"871- (12.21) If instead each country were to choose its· own optimal inflation oli autonomously, analogous to (12.19), the expected loss would be p cy Stabilization Policy Following Alesina and Grilli (1992), we begin with an analysis of policy conflicts within a monetary union in the absence of political union. Suppose there is a single monetary entity ("Europe") which sets policy to minimize a loss function trading off unemployment and inflation (as in (12.10), but with no devaluation cost ((Tr )): E£':aut = E[(_!l_ -) 1 + 6i €i - lf; l 2 (· 1 + 6i 1 + 0; €; 1 ｾＩＲ｝＠ - 6; lf; ' (12.22) a:2 (12.23) implying a difference of (12.17) where TrEu is the common inflation rate across members of the union, UEu is the central bank's unemployment target measured relative to the natural rate of unemployment (so that UEu < 0 if the central bank's target is below the natural rate), and e is the relative weight put on inflation fluctuations. This is minimized subject to an unemployment relation for the entire union: (12.18) 1 +-a:2 + 1 + 6; t! 1+ (}. l (1 + €>)2 '1 2 ｃｨ｡ｰｾ･Ｚ＠ -T'+"eu"'l' 13 • results are Identical to those derived for the closed economy in Section 4.4 of 550 CHAPTER TWELVE where o;. 2 is the variance of €; and o;, 11 is the covariance between €; and '17·. Differences in economic welfare from participating in a union versus retaining autonomy in policymaking stem from two sources: differences in o_bjective ｾｮ｣ｴｩｯｳＬ＠ reflected in differences 0 and 8;, as well as between Ueu ｾｮ､＠ U;; and differences in shocks across economies, so that o;,2 ¢ a"'2 • Consider first the case where the shocks are identical across countries that is, €; = '17, so that u112 = u/ = o;.'l = u 2 • In this case, the right-hand side of (12.23) may be written (J. I ( - )2 (Ueu @2 - ( - )2 ) + .!:l_ 0/ . (1 + @ 1+ 2 oJ (1 + 8)U _1_ - _1_ . _ 2 a: ) ｾＱＩ＠ • 551 minimize its own expected 2j, that is, to minimize (12.21). One can then show that the lower is the variance of unemployment in a country, the more conservative a European central bank it will prefer. The same holds true the lower is the correlation between domestic and European unemployment fluctuations. When the correlation is low, a country would prefer a central bank which is relatively inactive, as it will generally stabilize in the wrong direction from country i's perspective. We return below both to specific application of these results to EMU and to general issues of choosing representatives in transnational organizations. The Decision to Participate 2 > (12,24) so that a coul!.try wi!h an incentive to high inflation, namely, one for whom 0; < e and IV; I> IUeul. will unambiguously benefit from participation in a because of the higher credibility it bestows. This is simply monet.ary ｵｾｯｮ＠ the Gmvazzt-Pagano argument discussed in Section 12.2 above, extended to the case of an economy facing stochastic shocks. ｴｨ｡ｾ＠ countries have identical objective functions, In .contrast, ｳｵｰｯｾ＠ that 1s, 0; = 0 and U; = Ueu for all i, but face different shocks. The difference in objectives (12.23) becomes 1 E.:?;,. mem _ E£aut = ___ (a: 2 + a: 2 ' 1+@ • 11 THE INTERNATIONAL ECONOMY (12.25) A nu!l;lber of results emerge. Suppose, for example, countries differ in the magnitude of their shocks, but shocks are perfectly correlated across countries, so that o;, 11 = o;. u'l and the term in parentheses in (12.25) · becomes ( o;. u11 ) 2 > 0. Thus; membership in a union is always welfare reducing. If the variance of shocks in country i exceeds the European ｡ｶ･ｲｾｧ＠ (so that o;. 2 > u112 ), the European central bank will be stabilizing too httle from country i's perspective, while if unemployment shocks show less variability than the European average, the central bank will be stabilizing too much. As a second possibility, suppose the variance of unemployment is the same across countries, but shocks are not perfectly correlated. The lower is the correlation between country-specific unemployment shocks and the common European shock, the greater will be the loss from participating in the union. In the polar case of perfect negative correlation with country i's shock, the European central bank will be contracting when country i experiences a negative shock and expanding when it experiences a positive shock. . Alesina and Grilli also consider the case .where 0, rather than being gtven, reflects the outcome of majority voting in the setting of the common European monetary policy, each country choosing its preferred 0 to Comparing the benefits from participation versus nonparticipation leads to a more explicit discussion of the incentives to participate. Casella (1992) considers the decision of whether to participate in a monetary union from the perspective of the distribution of power within the common central bank. Whereas most papers on international cooperation consider equalsized countries, as in the previous subsection, Casella concentrates on the implications of unequal size for decisionmaking within the union, and, ｯｾ＠ the basis of that, for the decision to participate. Naively, one might think that a country's weight in decisionmaking should be proportional to its size, measured along some dimension. (This is implicit in the rule for splitting seigniorage revenues in EMU, whereby each member country's seigniorage share depends half on its population share and half on its GDP share in the total.) Suppose, however, that two countries contemplating a cooperative arrangement are of very different size. Under the proportionality rule, the much larger country would then basically determine the outcome in a cooperative arrangement, implying that the much smaller country would be worse off than in an equilibrium with no cooperation. 14 If participation in the union is voluntary (rather than required as a condition for something else, such as political union in), a small country may choose not to participate. It may then be in the interest of both countries to share power more equally than implied by the proportionality rule, so that the small country finds it optimal to participate. This point is shown formally in Section 12.5 below using reaction functions to illustrate noncooperative versus cooperative equilibria. Fiscal Aspects and Seigniorage Instead of considering the implications of a common monetary policy in terms of stabilization and response to shocks, one can consider the public 14 Note that this is different than saying that noncooperation dominates all cooperative agreements from the point of view of each country. It only says that noncooperation dominates a specific cooperative agreement, namely, one where power is proportional to size, from the point of view of the small country. 552 CHAPTER TWELVE finance implications of a common monetary policy and a common inflation tax rate. Giving up the right to print money means giving up the right to determine the level of seigniorage. Of course, seigniorage does not disappear; it simply accrues to a different entityP Hence, there is the question of how the level of seigniorage is to be determined and how seigniorage revenues are to be split. Canzoneri and Rogers (1990) reexamine criteria for an optimal currency area from a public finance perspective. Countries rely on seigniorage revenues to a different extent, even from the perspective of optimal tax packages. Hence, they recommend an additional criterion in specifying whether two countries should form a monetary union, namely, whether their optimaUax structures imply a similar reliance on the inflation tax. Among other things, this suggests that a high level of government spending may make it less likely that a group of countries will .form an optimal currency area, as high revenue requirements may make it more likely that different countries will have different optimal tax packages. Splitting the seigniorage has political implications as well. Casella (1992), as discussed above, considers the allocation of seigniorage revenues in partially determining the decision of whether to participate in a monetary union. Implicit in the rule for splitting seigniorage in EMU is the fact that a country's weight in decisionmaking should be proportional to its size, measured ,along some dimension. Hence, if participation is voluntary, a small country may choose not to· participate. Allocating to it a more-thanproportional share of seigniorage revenues may then be necessary to induce its participation. With a common currency, more-than-proportional influence of the small country is equivalent to a transfer of seigniorage revenues in its favor. Sibert (1994) argues that if the common central bank is able to choose both the level and the allocation of seigniorage, inflation will be suboptirrially high. Suppose provision of public goods is financed with both income taxes set at the national level and seigniorage at the "European" level. If fiscal policy is set before monetary policy, a national fiscal authority has an incentive to set income taxes too low, as a welfare-maximizing common central bank will then give it more seigniorage revenues. In equilibrium, income taxes will be too low in each country and the community-wide inflation rate will be too high. In contrast, if the European central bank chooses only the level of inflation, Sibert argues that it may· be too low, due to the negative spillover effects from income taxes. (Sibert assumes that the taxed factors are not mobile across countries, so there is no problem of tax competition, a phenomenon we discuss in Section 12.5 below.) An increase in a country's income tax rate lowers its disposable 15 As the move to a single money will change the demand for money of the union relative to the sum of demands for individual countries' monies, total seigniorage can, of course, change. The direction of the effect could be positive or negative. THE INTERNATIONAL ECONOMY 553 ｩｮ｣ｾｉ＿･＠ ｾｮ､Ｌ＠ ultimately, community-wide seigniorage revenues. In the Nash equtlibnum between two fiscal authorities, their failure to internalize fiscal ｾｸｴ･ｲ＿｡ｬｩｳ＠ leads to too high a level of income taxes and too low a level of ｾｮｦｬ｡ｴＱＰＮ＠ Note the Ｎ｣ｯ＿ｴｲ｡ｳｾ＠ ｾｩ＠ ｴｾ･＠ situation where there are multiple Issuers of m??ey withm a JUriSdiction, in which the failure to internalize from monetary spillovers leads to an overissuance of the ･ｸｴｲｮｾｨｩｳ＠ money, as m Casella and Feinstein (1989) or Aizenman (1992). The European Monetary Union As indicated at the beginning of this section, a current area of research is the recently formed European Monetary Union, the EMU. The move to a common currency has been a major topic of research, especially in Europe throughout the 1990s. The literature is literally voluminous· ｗｹｰｬｯｳｾ＠ (1997) is but one of several excellent recent surveys. ' The ｅｵｲｯｰ･ｾｮ＠ Economic Community (EEC, sometimes simply denoted EC) as an en.tlty dates to the 1957 Treaty of Rome. Since then, various plans for a smgle European currency were devised and discarded· real progress towards economic unification can probably be dated to 1986, with ｾｯ＠ key events. Three additional countries, Greece, Spain, and Portugal, ｊｾｉｮ･＠ the ｾｅｃ［＠ and, the Single European Act (SEA) was adopted, dictatmg a smgle market and removal over time of all barriers to the movement of goods, capital, and people within the EEC by the end of 1992. The Treaty of European Union (commonly referred to as the "M.aastricht Ｚｲ･｡ｾＢＩ＠ in 1992 formally set out the goal of a European Umon (El!) mv_olvmg both economic and political union. The centerpiece of economtc union (over and above the SEA) was the creation of EMU a European Monetary Union. The nature of political union was left ｭｾｲ･＠ vague by ｴｨｾ＠ ｾ｡Ｌｴｲｩ｣ｨ＠ Treaty, the goal being eventual joint foreign and ､･ｦｮｳｾ＠ pohctes. c::mvergence criteria" were also established requiring countnes to show evtdence of "good" macroeconomic behavior in order to be allowed ｴｾ＠ enter the future monetary union. Three criteria concerned ｭｯｮ･ｴｾｲｹ＠ policy, two concerned fiscal policy. More specifically, the monetary Ｎ｣ｮｴ･ｾ｡＠ for a country to join the single-currency area were: first, that 1! percentage points of the average of the Inflation rate must be ｾｴｨｩｮ＠ the three lowest EU countnes; second, the long-term interest rate must be no greater than 2 percentage points higher than the interest rates of the three countries with the lowest inflation rates; third, the exchange rate rate bands of the existing must have remained within the ･ｾ｣｡ｮｧ＠ European Monetary System (EMS) Without severe tensions" for at least two years. The fiscal criteria were: first, a ceiling on the ratio of debt to GDP of 60%; and, second, a ceiling on the ratio of the government deficit 16 to GDP of 3%. When the Maastricht Treaty was signed, only Luxem16 The " · d fi · d ･ｸｾｳＱｶ＠ e .c.It ｐＬｾｯ｣･＠ ure" makes this last entry criterion permanent, except " under exceptiOnal cond1trons when a country is temporarily allowed to exceed the ceiling. 554 CHAPTER TWEI.VE bourg met all five criteria. We return below to the politics of the convergence criteria. The Maastricht Treaty contained an explicit timetable for EMU, envisioned as a three-stage process. Stage I, which began in 1992, comprised formal ratification of the treaty. Stage II, initiated in January 1994, was aimed at achieving various monetary and fiscal conditions for the establishment of a European Central Bank (ECB). First, national central banks were to be granted greater independence by their respective governments. Second, the European Monetary Institute was established, both !o pave the way for the creation of a European central bank and to momtor the convergence criteria set out above. Stage III was the initiation of the monetary union itself, which was to take place no later than January 1, 1999, as it did, with 11 out of the current 15 EEC members (Austria, Belgium, Finland, France, Germany, Ireland, Italy, Luxembourg, Netherlands, Portugal, and Spain) joining the union. Exchange rates between the member countries were permanently fixed, both to one another and to a European currency unit, the euro. National currencies may still be used for a transition period, the euro starting to circulate as the medium of exchange after January 1, 2002, with national currencies losing legal tender status on July 1, 2002. As in our discussion of fixed exchange rates in Section 12.2, we begin with a discussion of the economic arguments for EMU as a background for discussion of the political aspects. The common, though not universal, view is that the economic arguments in themselves do not make the case for EMU. First, there are the benefits from lower transaction costs from of ｾｵｲ･｣ｹ＠ moving to a single currency, both in terms of lower ｣ｯｳｴｾ＠ conversion and the effect of lower exchange rate uncertamty mducmg greater trade and investment within the EEC. It is generally ｡ｲｧｾ･､＠ ｴｾ｡＠ the saving in transaction costs is not large (see, for example, the discussion in Eichengreen and Frieden (1993]), though Melitz (1997) argues that the economic case stands by itself, on the basis of the lower transaction costs it implies. A second argument is that the ECB will act as an anti-inflationary commitment mechanism as discussed in detail in Section 12.2 for fixed exchange rates and earlier in this section for stronger institutional arrangements. This is seen as a benefit especially to the high-inflation countries in the EEC. There are several reasons why this should be seen as a political, rather than simply an economic, argument. First, question.s of credibility are inherently political, though the ultimate goal, of course, ts to reduce inflation, that is, economic. Second, the decisionmaking process within the ECB (a board consisting of representatives from each member country, with "one-man, one-vote" decisionmaking) means that the con:mon monetary policy need not be that of a "conservative central banker ' as discussed in Section 5.4 of Chapter 5 above. Third, independent of institutional design, the choice of a common monetary policy will reflect the balancing of divergent country interests, as discussed in the model of Aksina and Grilli (1992) above. We return to this issue below. 1"1/Jo I!VTJiRN.< TION.<I, liCONOMY 555 . A third economic argument for EMU is that the move to a truly market in goods, labor, and capital dictated by the mtegrated_ ｅｵｲｯｾ･｡ｮ＠ SEA ｲｾｱｵｴ･＠ a smgle European currency. The removal of capital controls makes ｭｦ･｡ｾＱ｢ｬ＠ any_ exchange rate arrangement other than either a single currency or fully flex1ble rates between European countries, as discussed at the very beginning of this section. Flexible rates are seen as inducing high ･ｸ｣ｨ｡ｾｧ＠ rate uncertainty; more importantly, they are basically incompatible With the overall goal of an econ01pically fully integrated Europe. This ｡ｲｧｵｭ･ｾｴ＠ ｯｾ＠ the necessary relation between monetary union and true economtc umon more generally is persuasive; however, it simply shifts the foc.us to the economic benefit'> of full economic integration, which must be we1ghed against the economic costs. The main economic cost of a common currency is the loss of autonomous monetary policy for individual countries, central in the discussi:m ｯｾ＠ optimum currency areas and considered more formally in the d1scusswn of the Alesina-Grilli (1992) model earlier in this section. The general view is that Europe does not satisfy the standard conditions to constitute an optimum currency area. There is disagreement about the extent of asymmetry in the shocks hitting Europe: Bayoumi and Eichengreen (1994) argue that aggregate demand and supply shocks are much more asymmetrically distributed across the EEC than across the United ｾｴ｡･ｳＮ＠ On ｴｾ･＠ other hand, many industries cut across Europe, so that mdustry-spectfic shocks affect countries in similar ways; Wyplosz (1997) ｳｾｧ･ｴ＠ that European economies are well integrated. There is little d1sagreement about relatively low labor mobility across Europe, suggesting overall _that the conditions for Europe to be an optimum currency area are not satisfied. Overall, it is not easy to make a clear case for EMU on purely economic grounds. The Political Economy of EMU ｾｴｨ･＠ above discussion makes clear, the issues surrounding EMU are not stmply economic, but also political. Although the political nature of EMU is widely recognized, there has been relatively little formal political-economic analysis. A good reference is Eichengreen and Frieden (1993). As they point out, the decision to create a single currency is a political phenomenon, in that it was made not by a social planner, but is "the ｾｵｴ｣ｯｭ･＠ of a political process of treaty negotiation, parliamentary ratificatiOn, and popular referenda. Interest groups support or oppose the initiative depending on how it is likely to affect their welfare, not the welfare of the nation or of the [European] Community as a whole." (p. 85) There are ｴｨｲｾ＠ general aspects to the political nature of EMU: its specifically pohttcal goals; the conflict over member countries' objectives; and, implications for the structure of EMU of the political process of resolving these conflicts. We consider these in turn. CHAPTER TWELVE 556 The .most basic political aspect of EMU is the importance that ｾｮｹ＠ observers put on monetary union ｳｰ･｣ｩｦ｡ｬｾ＠ as. a ｰｲ･ｾｵｳｯ＠ to poh:tcal union. That is, the argument for monetary umon ts that tt helps t? ｡ｾｨｴ･＠ the goal of political union. This issue was discussed generally earher ｾｮ＠ thts ?nly m ｾｨ･＠ section; it is a major argument made for EMU, apparent ｮｾｴ＠ Maastricht Treaty, but even in the whole process of economtc mteg.ratlon A ｾｭｲｮ＿＠ argument ts that beginning with the Treaty of Rome in ＱｾＵＷＮ＠ economic union implies political cohesmn, whtch wtll serve to prevent future military and political conflicts in Europe. Feldstein (1997), a vocal opponent of EMU, argues that the effect m.ay actually be ｾ＠ the ｾｰｯｳｩｴ･＠ direction with the freedom to pursue thetr own economiC, soctal, and ｩｮｴ･ｲ｡ｾｬ＠ objectives that countries have in the absence of EMU serving to reduce conflict relative to a situation in which countries with diverse This ｡ｲｧｾｭ･ｮｴ＠ gets. to experiences are forced to pursue common ｰｯｬｩｾｹＮ＠ the heart of the issue, namely, whether enforcmg economic cooperation will make political cooperation more or less easy. . . . . Conflict over objectives is inherent in any economic umon. Thts mcludes not simply the conflict between member countries over ｯ｢ｪｾ｣ｴｩｶ･ＮｳＬ＠ but also conflict among interest groups ·within member countries which will play out differently because the country has less autonomy ｾ＠ policymaking. ｾｯｲ＠ example, in responding to the attack on ｴｨｾ＠ ｾｯｵｮ＠ m.1992 ｟Ｈｾ･＠ Sectt?n 12.3 above), the British government was hmtted m tts abtlity to ratse interest rates because the structure;;of mortgage lending meant that an increase in short-term rates would be passed on to mortgage rates. This would hurt property ｩｮｴ･ｲｾｳＬ＠ a key constituency of the ruling Conservative party, Their influence on policy will ｯ｢ｶｩｾｵｳｬｹ＠ differ in the con:ext of a monetary union. Similar examples can be gtven for other countrt.es concerning the influence of domestic interest groups on monetary po.hcy. . Distributional conflicts within countries are in turn reflected m pohcy differences across member countries of a mo11etary union. In EMU, the key policy conflict is over the degree of anti-inflation ｾｯｭｩｴ･ｮＮ＠ Geras ｂ･ｬｧｴｾｭＬ＠ ｌｵｾｭ｢ｯｲｧＬ＠ many and countries in its monetary sphere Ｈｳｾ｣ｨ＠ and the Netherlands) favor a stronger commitment to pnce stabthty; some other EC countries favor a weaker commitment, with greater stress on other objectives. The Alesina -Grilli (1992) model above illustrates the nature of this conflict. Interestingly, the move from the EMS to EMU can be seen as a move away from the highly anti-inflationary policies associated with the Bundesbank rather than a move towards tougher monetary policy. With free ｣｡ｰｩｴｾｬ＠ flows within Europe, the EMS was compatible only with a single ｇ･ｲｭｾ＠ monetary policy for member countries, w}fich increasingly ｾ･｣｡ｭ＠ monetary policy. (This reflected a number .of ｦ｡｣ｾｲｳＬ＠ n?t stmply ｇ･ｲｭｾｮｹ＠ s economic size, but also the acceptance of Jts ant1-1nflat10nary leadership by like-minded countries.) For example, monetary tightness in the EMS in the 1991-1993 recession in Europe largely reflected the policy preferences of THE INTERNATIONAL ECONOMY s: Germany; other EMS members wanted less restrictive monetary polic but the exchange rate system .forced them to follow Germany's lea; Hence, countries such as France and Italy wanted to find a way to reasse1 a degree of control over the setting of their monetary policy. The creatio of a European monetary authority that would supersede the Bundesban and in which they would have a larger voice was seen as the way to regai some degree of monetary control. Hence, they strongly favored the en ation of a European central bank with decisions made by majority vote, t which Germany was resistant. We return to this conflict below in conside1 ing how politics affected the nature of bargaining over EMU, as well as th outcome of this bargaining. The conflict over objectives can also be seen in conflict over the natur of the transition. The length of stages I and II in the transition fror Maastricht to monetary union reflected the resolution of a conflict be tween two different views of the relation between macroeconomic condi ti.ons and monetary union. One view, favored by central banks and espe ctally the Bundesbank, was that low inflation and fiscal balance in membe countries was a precondition for a successful monetary union. Otherwise macroeconomic imbalances would make monetary union unsustainable This "coronation approach" held that countries should satisfy strict con vergence conditions before being allowed to join the single currency. Th• competing view was that convergence would be the result of the creation o a monetary union with an independent central bank, in line with the vie.,. that tough monetary institutions will impose a discipline on countries tha would enable them to reduce inflation and fiscal deficits. The stress put 01 the five convergence conditions set out above, and the length of time givet to allow them to be achieved, was the victory of the "coronation approach' over the "discipline from institutions" approach. Finally, EMU can only be understood from a political perspective in tha1 the nature of the transition, of its institutional structure, and of the policies that are expected to emerge, reflect the process by which the above conflicts of interests were and are resolved. German acceptance ol the European Central Bank as it is structured in place of the ｂｵｮ､･ｳ｢｡ｾ＠ provides a good example. The conventional wisdom is that the ECE recreates the Bundesbank at the EC level, with a similar ｳｩｮｧｬ･Ｍｭ､ｾ＠ commitment to price stability. The nature of the ECB governing board (numerically dominated by country representatives rather than members o1 the ECB Executive Board, with decisions made by majority rule) calls into question the argument that the ECB's institutional structure will automatically imply a tough anti-inflation outcome. (This is same point as made that the .move from EMS to EMU may imply a weakening above, Ｎｾ｡ｭ･ｬＡＬ＠ of antt-mflattonary poltcy.) Germany favored an institutional structure which would have implied a more anti-inflationary ECB, for example, more decisionmaking power in the hands of the Executive Board, and unanimity required for certain decisions, giving individual countries veto 558 CHAPTER TWELVE power. As structured, the ECB could be characterized as reflecting not a victory for German interests, but concessions on her part. Garrett (1993) argues that Germany made such concessions because it had other goals that gave it a strong interest in making sure that EMU came into being, even if it meant giving up some of its economic objectives. Specifically, he argues that Germany was especially interested in gaining international acceptance of rapid reunification; that united Germany be accepted into the EC without treaty revisions; and, that both Europe and the United States not view reunification as presaging a more militaristic Germany. Making concessions was seen as smoothing European acceptance of reunification on terms Germany preferred; reunified Germany's tying itself to a united Europe was strongly seen as calming international fears that reunification might arouse. The nature of the CB can thus be understood only against the backdrop of the bargaining process that led up to it. This case provides but one example of a general phenomenon denoted linkage politics by political scientists, whereby otherwise unconnected objectives are tied together so that agreements can be reached. Parties make concessions in unrelated areas to reach agreements, so that the nature of outcomes reflects not only the different objectives of the ｰ｡ｾｴｩ･ｳＬ＠ but also the process by which agreement is reached. Linkage po1itics played a crucial role in the formation of EMU. . The outcome of the conflict over whether macroeconomic convergence should precede or follow monetary union can now be better understood. Garrett argues that Germany viewed a strong anti-inflation commitment as deriving not only from the institutional characteristics of an independent monetary authorit)r, but also from the interaction of these institutional features and a strong anti-inflationary constituencyP Hence, an independent ECB would be strongly anti-inflationary only if the members of EMU shared anti-inflationary objectives themselves. This view was the basis of Germany's insistence on "convergence before union," rather than vice versa. Moreover, once this German view on the nature of the transition to· stage III was adopted, implying an EMU composed of members whose anti-inflation preferences have been demonstrated (or created), Germany was. more willing to yield on the nature of decisionmaking within the ECB once established. Alesina and Grilli (1993) show formally how the nature of decisionmaking within EMU can affect the relation between the composition of EMU and the policy outcome. They use the Alesina-Grilli (1992) model (exposited earlier in this section) of deciding whether to join a monetary union when countries differ in the relative weight they put on their inflation and unemployment goals, as summarized by a parameter 8; for 17 Posen (1995b) makes an analogous argument for the United States, namely, that the effectiveness of monetary institutions in providing anti-inflation credibility depends on the political support given to their goals. See Section 5.4 of Chapter 5. THE INTERNATIONAL ECONOMY country i. They define a feasible monet . . 559 member country is not wors ff . aty umon as one In which each o m the .umon than outside. With EMU's inflation policy defined by · fl · oo, representmg the ·h m atiOn fluctuations (see (12.17) and (12 19)) We1g t the ECB puts on for country i depends on ® r 1 f . ' the benefit of joining EMU suppose that there are three ･ｰｾＺｶ＠ /oi 8; (see (12.23)). For concreteness ･ｾ＠ Ｑｾ＠ members, say Italy, France, and Germany, with 8IT < eFR < 8 the shocks are the same ｡｣ｲｯｾｩＧＺ＠ tt. at both unemployment targets and f, n nes. ｔｨｾｲ･ｦｯＬ＠ conflict of interests reflects only different inflation all i. Note that in this case th/re. er:nces, as m (12.24), with 6; = Ueu for 8; = ®. Hence, the minim'um Jat om membership is identically zero if representing the gain from h . or a country to join is ei, with ® > (}. with ｬｾｳ＠ of an inflation ｢ｩ｡ＺｾＡｮ＠ ｾ･ｴ｡ｰｯｬｩ｣ｹ＠ se.t by a central bank a max:tE.tum ® consistent with it b . g . rnment Itself has. There is noted ®;. emg opttmal for country i to join, deFeasibility depends on there b . a ｾｯｮ･ｭｰｴｹ＠ range of possible values for ®,such that membership is ｰ［ｾｧ＠ ＮｾＱ＠ ?onmen:bership. For example, when countries face identical ｳｨｯ｣ｫ･ｾ＠ that the high-inflation country It east ttty reqmres that the highest ® membership at ®IT is preferred t a y would accept (that is, for which That is, there exist values ® E oe ｮｯｾ･ｭ｢ｲｳｩｰＩ＠ is no less than Bae· chosen depends on the votin f GE• ®IT 1. Whtch value of ® will be Fo 1 · g procedure and the nat f ｵｲｾ＠ o preferences. r examp e, if the preferred ® of both F ?utcome of majority voting will be ® = e ranee and Italy ts below Oae, the JUst tough enough to keep G . GE· Monetary policy will be made convergence criteria lead to a stermany m. ｾﾷ＠ To the extent that the and Italy before EMU is ina . ron;e; anti-mflation constituency in France ｾｯＧ＠ ; ' ｲ･ｦｬ｣ｾ､＠ in higher eFR and 8IT• ECB monetary policy will be ｾｦｯＺ＠ and Grilli also consider cases h erhmany s preferred policy. Alesina coun tnes . or where countries assi w ere . th s bocks are not 1'd entlcal across gn er enefits to EMU. ! i ° PART II-MACROECONOMIC INTERDE.PENDENCE 12.5. INTERNATIONAL POLICY COOPERATION A major issue in open-economy macro . . economic disturbances across countri ･｣ｾｮｦＺＱｳ＠ IS the transmission of governments are a main source of th esd. . o Ictes undertaken by foreign of soveretgn · governments aff t ese tsturbances. Wh en po1Icy . actiOns . c.oordination becomes central. ｉｾ｣ｴｨＺ＠ another, the question of policy hves, there would of course b o governments had identical objecBut si?ce the two ｾｯｶ･ｲｮｭｴｳ＠ ｾＺ･＠ ｲｾ｡ｳｮ＠ fo.r Ｎ｣ｯｲｾｩｮ｡ｴ＠ to be an issue. there IS no reason to suppose that th .y ｢･ｾｭｉｏｮＬ＠ d!fferent constituencies etr o ｾ･｣ｴｶｳ＠ Will overla . ' CHAPTER TWELVE 560 (1976) observed in his pioneering contribution on international policy game, and coordination, policymaking then becomes a ｮｯ｣ｰ･ｲｾｴｩｶ＠ there are likely to be gains from cooperation. This reasoning implicitly assumes that countries are governed by benevolent social planners. Given the external effects from macroeconomic and trade policies, uncoordinated policies are likely to lead to inefficient outcomes, relative to what could be achieved under coordination. For example, if two countries each engage in expansionary monetary policy in order to depreciate their currency relative to the other in order to improve the current account, neither achieves the goal of a change in relative prices, and both end up with undesired inflatiQn. Assuming that countries can make binding commitments, a more efficient outcome can be obtained. There have been numerous papers on both theoretical and empirical aspects of policy coordination. Canzoneri and Henderson (1988) and Persson and Tabellini (1995) present excellent surveys of models of coordination, stressing political economy aspects. Coordination can take several forms. The most obvious form, which will be our focus of attention, is direct policy coordination. It is not, however, the only possibility. We begin with an alternative. Policy Assignment Following Mundell (1968), one can take the policy assignment approach, which, following Tinbergen (1952), considers the optimal assignment of instruments of economic policy to targets. By analogy, rather than two countries using all instruments to affect all targets, one could consider the assignment of instruments and targets to each country. For example Mundell (1971) advocates that under the then-prevailing dollar standard, the United States would adjust its money supply to peg the world price level, while the "rest of the world" would use monetary policy to keep the balance of payments in equilibrium. Such a view is based on the "redundancy" or "N - 1" problem, whereby N ...:. 1 countries can achieve their balance of payments objective, with the Nth country acting·as the residual country. (See also Swoboda and Dornbusch [1973) for an applica- · tion of this approach.) Though the assignment approach is attractive in its simplicity, it can be misleading for a number of reasons. First, optimal policy often requires a mixing of instruments to achieve objectives. Second, under uncertainty, achieving policy goals may require more than simply having as many instruments as targets. (See, for example, Brainard [1967).) Furthermore, the simple assignment of objectives to countries ignores a basic incentive problem: will it be credible that a national authority will always have the incentive to use policy for the world good according to its THE INTERNATIONA.L ECONOMY . 561 assigned role? Hence one wants to kn ment would find it in 'its best interests ｴｾｷＺｏＡＮ＠ ｾｮ＠ optimizing govern- Optimizing Governments-A Simple Monetary Model M • mteracting governments as 0 t' . . . 0 d ･ｹＭｾｧ＠ 11 mternat10nal policy coordination in t P un:mg ｡ｧｾｮｴｳ＠ ｳｵｾ･ｴ＠ analyzing 0 ･ｲｭｾ＠ strategic behavior on the part of each country That is co by ､･［ｩｶｮｾＺ＠ ｾ･ｨ｡ｯｲ＠ ｾｮ､＠ the resultant equilibrium would be ｭｯ､･ｬｾ＠ of their own optimizing ｢･ｨ｡ｶｩｯｾＮ＠ Ｚｯｴｾ｣ｬｏｮｳ＠ f?r countries on the basis ｲｾ｡ｳｯｮ＠ for modeling governments as optimizing agents al d ･ｲ､ｾｵ｣ｩ｡ｬ＠ currency crises is that count . .' rea y Iscussed m our treatment of ｴ｡ｲＮｧ･ｾ＠ It ther;fore seems ｭｯｾ･Ｚ］ｩ｡ｶ＿＠ fewer ｩｮｳｾｭ･ｴ＠ than as havmg conflicting obJectives that require trade-offs a h Iew ｴｨ･ｾ＠ ｴｾ･ｲｯｧｮｩｹ＠ paradigm suggests. All the more so between nations' ｾｯｵＮｲ＠ who asks whether countries ｵｲｳｾｩｮ＠ st IS. e approach.of !Iamada (1976), dependent world will lead topan ｾ＠ ..Ｚｾｉｲ＠ ｮｾｴｩｯｬ＠ ObJeCtives in an intermaximized, and who finds not ｳｾｱｭＮｉ＠ . ｮｾ＠ m which worldwide welfare is ｾｨ･ｴｲ＠ this result will be' altered ren:mg y, ｴｨｾ＠ it ｷｾｬ＠ not. I_Ie then asks tlonal changes ("rules of the g ｾＩ＠ ｾ･＠ ｡､ｯｰｾｉｮ＠ of mternatlonal institueffect of institutions from theame. . nf ｓｾ｣ｴｬｯｮ＠ 12.2, we considered the t ' . . pomt o view of one Ignonng. strategic interactions.'. here, we me . lude such m· te count'ry s welfare, as f ｯ｣ｵｳｭｾ＠ more on worldwide welfare. rae mns, as well The basic arguments can be re re d. quadratic loss functions and lin:a sente ｾｮ＠ terms of a simple model with (We consider the implications of alternative specifications ｾＺｏｴｲｮｳＮ＠ ｉｶ･ｾ＠ .and economic structure in the next two subsections.) Su ose ｾｩ･＠ Simplicity, ｾ｡ｴ＠ the world is made up of two "countries": the urJfed St ｾｵｲｯｰ･Ｌ＠ denoted with a subscript ｅｾ･ｯｴｷ＠ a subscript US, and Is, as above, a function of unemploym.en t an e dos.smfl atlon: n.ctton for each country ..'?:I = u,z + 0·7TI2 , I I fore·= U.S ' EU' (12.26) where 7T is the rate of inflation of the co . . pnce mdex and, as always, the unemployment rate is measured rel ｾｵｭ･ｲ＠ ployment. Note that we have assumed tv: to ｴｨｾ＠ natural rate of unemployment rate is equal to the natural t a. hｨ･ｰｯｾ｡ｫｲＺｳ＠ target unema surprise inflation, an assum tion ra e. ｾ＠ as no m:entlve to engineer ｷ･Ｎｲｴｵｮｾ＠ ､･ｴｾｨ［ＺＡＮ｢＠ crucial for our simple results, and to ｷｨｩ｣ｾ＠ The consumer pnce mdex is a function f . o assume the domestic output price levels, which for simprIct'ty we enter and withforeign equal t: 18 .. Instrument assignment to different levels • pohttcal union) is widely discussed in th bl' fiof gove_mment tn a federal system (or a our discussion of fiscal unions in Sectione lp4u6Icf enhance literature. We return to this issue in · o apter 14. CHAI'TER 562 1'Wt,:t,VE 1'HE INTERNATIONAl, ECONOMY 563 weights. Assume that the exchange rate is fixed and that output price inflation in each country is equal to the rate of domestic monetary growth, mus for the United States and mw for Europe. Hence, the CPI inflation rate in each country is simply 71" 1 = .5mus + .5mEu• fori US, EU. (12.27) Unemployment in each country is related to unanticipated output price inflation according to (12.28) where a;> 0 is a slope coefficient and c: is an unemployment shock common to both countries. Each country has two targets but only one instrument, so that it cannot achieve its objectives independent of the action of the other country. In the absence of cooperation, the relevant solution concept is Nash noncooperative equilibrium. One begins by deriving reaction functions for each country, which give the optimal setting of the policy instrument under its control for each policy choice of the other country. In each country, we minimize the loss function (12.26) with respect to its own money growth rate subject to (12.27) and (12.28), holding constant the other country's ｭｯｾ･ｹ＠ growth rate. Taking expectations to eliminate m£u and mbs (which will equal zero), this yields reaction functions mus mEu = nusc €, mEu) =A us€ + DusmEU• (12.29a) = AEu€ + DEumus• (12.29b) nEU(c:, mus> where D.= .259, ' Solving (12.29a) and (12.29b) simultaneously, we derive the noncooperative Nash equilibrium: The solution and some of its properties are illustrated in Figure 12.1, which should be studied carefully. The reaction function of each country is given by the thick solid lines, with the most preferred point for the United States, for example, given by Ous (and similarly for Europe by Ow>· The Nash equilibrium is given by the intersection of the two reaction functions FIGURE 12.1. Reaction functions. at N. Indifference curves centered on Ous and OEu radiate outward, with utility being lower the farther away from 0 1 is the curve. The U.S. indifference curves are vertical when they intersect the US reaction function, since along the reaction function, mus is chosen so that its marginal effect on US welfare is zero, given mEU. Analogously, the European indifference curves are horizontal when they intersect the European reaction function. In this simple example, the reaction functions are necessarily downward sloping, with the European reaction function steeper than the US reaction function for mEu on the horizontal axis. More generally, in models of international spillovers, reaction functions could be either positively or negatively sloped. For example, in a model where a monetary expansion abroad increases unemployment at home via a depreciation of the foreign country's currency and a resultant trade balance improvement abroad, it would be easy to obtain a positively sloped reaction function. (Frankel and Rockett [1988] present a simple general mathematical representation of policy spillovers and reaction functions, and discuss the possibilities.) The case of negatively sloped reaction functions, where an increase in the setting of one country's policy instrument induces a decrease in the other country's policy instrument, is the case of strategic substitutes. When an increase in mu5 , for example, induces an increase in mEu• we have a case of strategic complements, where multiple intersections of the reaction functions, that is, multiple Nash equilibria, are possible. Independent of