Should Canada and the United States adopt a common currency? Should Mexico or other Latin American nations adopt the U.S. dollar? Suddenly, monetary union in one form or another is very much on the Hemispheric agenda. Debate focuses on the advantages and disadvantages of either merging two or more national monies into one (currency unification) or supplanting one national currency with another (dollarization). Assessments range from the highly favorable to the distinctly critical, as other contributions to this volume amply demonstrate. Consensus remains elusive.

Among the reasons for discord is a tendency for most analysis to focus narrowly on just the economic gains and losses associated with monetary union. No discussion could be complete, however, without some consideration of the politics involved as well. The purpose of this chapter is to highlight the political dimension of monetary union: key domestic and international political issues implicated when sovereign states consider either currency unification or formal dollarization. The role of politics will be explored in the context of two distinct stages: first, at the time when the initial decision is made whether to create a monetary union (i.e., to unify separate currencies or to dollarize); and second, as part of what determines the subsequent sustainability of such a joint endeavor. Economic factors, which figure so prominently in most of the other contributions to this volume, are of course a necessary part of the analysis. But they are not sufficient. By highlighting the political dimension, I hope to help clarify what is really at stake today for the nations of the Western Hemisphere.

Monetary union can be created in one of two ways. First, two or more countries can agree among themselves to merge their separate currencies into a wholly new joint money, as members of Europe's Economic and Monetary Union (EMU) have done with the euro. This is currency unification. Other examples of currency unification in the twentieth century include the East Caribbean Currency Area (ECCA, with the East Caribbean dollar) and the now defunct East African Community (EAC, with the East African shilling). Close variations, involving separate currencies so tightly tied together in an exchange-rate union that they effectively function as one money, include two in Africa, the CFA Franc Zone, with fourteen francophone countries in West and Central Africa, and the Common Monetary Area (CMA) joining together South Africa and three of its poorer neighbors (Lesotho, Namibia, and Swaziland). They also included until recently one in Europe, the Belgium-Luxembourg Economic Union (BLEU), now superceded by EMU. Second, one country can unilaterally or by formal agreement supplant its own currency with the already existing money of another. This approach, generically referred to as dollarization, is typically characteristic of tiny enclaves or micro-states such as the Marshall Islands and Micronesia in the Pacific (using the U.S. greenback) or San Marino and the Vatican in Italy (using the Italian lira). The three largest formally dollarized economies are Panama, which has used the dollar ever since gaining independence in 1903, along with Ecuador and El Salvador, both of which decided to switch to the greenback in 2000. Though in all three cases token amounts of local money continue to circulate, the dollar is now the main legal tender in each country.

Each of the two approaches, currency unification and dollarization, can be considered separately.

Currency Unification

First, consider the possibility of currency unification as advocated by Courchene and Harris (2001) and Grubel (2001). Since monetary union is by definition an economic matter, it seems only natural to focus attention mainly on economic considerations-the material costs and benefits associated with a merger of separate national monies into one. Would the citizens of Canada and the United States be better off sharing a single money between them? Or would real welfare losses, whether at the microeconomic or macroeconomic level, exceed any conceivable gains that might accrue? In practical terms, however, such a narrow focus is not only incomplete but potentially misleading. A common currency is not just about economics but is quite obviously also about politics-about gains and losses that fall outside the standard cost-benefit calculus of economics. Ultimately, it is about the exercise of power and the ability of a national community to control its own affairs.

Economic Analysis . Would a common currency be good or bad policy? In the formal economics literature the decision to create a joint currency is addressed as an optimization problem limited mostly to issues of general economic welfare. Analysis, as other contributions to this volume emphasize, is based on the familiar theory of optimum currency areas (OCAs) dating back to Robert Mundell's pioneering article published nearly four decades ago (Mundell 1961).

In its first incarnation, OCA theory was strikingly apolitical. Following Mundell's lead, most early contributors concentrated on a search for the most appropriate domain of a currency irrespective of existing national frontiers. The globe, in effect, was treated as a tabula rasa. The central issue was to find the best criterion for the organization of monetary space. But as the practical limitations of the so-called "criteria approach" (Tavlas 1994:213) became clear, an alternative--and, in political terms, seemingly less naïve--approach eventually prevailed, focusing instead on material gains and losses, as seen from a single country's point of view, stemming from participation in a common currency or equivalent.

On the positive side, currency unification can be expected to reduce transactions costs and increase money's underlying network externalities. The usefulness of a currency is enhanced in its standard functions as a medium of exchange, unit of account, and store of value. With replacement of local monies by a single regional currency, there is no longer a need to incur the expenses of currency conversion or hedging in transactions between the partner economies. Trade, as a result, could be increased substantially - by as much as a factor of three, according to empirical estimates by Andrew Rose (2000)-generating considerable efficiency gains. On the negative side, governments are assumed to consider the disadvantages of the corresponding surrender of monetary autonomy: the potential cost of having to adjust to domestic or external disturbances without the option of changing either the money supply or exchange rate. Monetary policy now becomes a matter of collective rather than unilateral decisionmaking, limiting what the country itself can do to cope with transitory or cyclical disturbances. As Paul Krugman has neatly summarized the calculus, it "is a matter of trading off macroeconomic flexibility against microeconomic efficiency" (Krugman 1993:4).

A diverse range of variables is stressed in the OCA literature, including: wage and price flexibility, labor and capital mobility, commodity diversification, geographic trade patterns, size and openness of economies, levels of development, inflation trends, and the nature, source and timing of potential shocks. Each country characteristic arguably affects the magnitude of losses at the macroeconomic level by influencing either the severity of potential disturbances or the ease of consequent processes of adjustment. The explanatory power of OCA theory, however, appears to be quite limited. There are simply too many permutations possible among the many factors cited. As one source puts it, quite bluntly, "theoretical ambiguities abound" (Argy and De Grauwe 1990:2). Not all of an economy's features may point in the same direction, making forecasts difficult; nor are the variables necessarily mutually independent or easy to measure or compare for relative importance. Concedes one recent study: "Overall the country characteristics do not help very much to explain the countries' choice of exchange rate regime. It might be that the choices are based on some other factors, economical or political" (Honkapohja and Pikkarainen 1994:47-48). Indeed--why should we be surprised that politics might also enter into such a critical decision?

In fact, political factors enter in two ways. First, the policy calculus is manifestly affected by domestic distributional politics: the tug and pull of organized interest groups of every kind. As political scientist Jeffry Frieden has emphasized, "domestic distributional considerations are also central to the choice of exchange rate regimes" (Frieden 1993:140). The critical issue is the familiar one of whose ox is gored. Who wins and who loses? The material interests of specific constituencies are systematically influenced by what a government decides to do with its currency. Producers of tradable goods, for example, as well as internationally active investors, are all apt to be favored by a currency regime that maximizes the predictability of exchange rates. Currency volatility, for such groups, is an anathema. Domestically oriented sectors, by contrast, are more likely to benefit from stability at home and thus to attach higher priority to preserving as much national policy autonomy as possible. Such groups stand to lose most from a common currency insofar as their local interests are submerged in a broader monetary area. Government choices are bound to be sensitive to the interplay among such domestic political forces.

Second, even apart from distributional concerns, the policy calculus includes much more than just macroeconomic performance alone. Plainly, diverse political goals at the national level must weigh at least as heavily as economic welfare in the strategic calculations of policymakers. It is by no means unreasonable to assume that governments are sensitive to the balance between macroeconomic flexibility and microeconomic efficiency. But in considering whether to share their monetary sovereignty, states are unlikely to limit their thinking to that one trade-off alone. There must also be a strong political incentives involved to persuade policymakers to make the kind of firm commitment that is demanded.

Certainly that appears to be the lesson of history, where it is impossible to find a single example of a common currency motivated exclusively, or even predominantly, by the concerns highlighted in OCA theory. As Paul De Grauwe has observed: "Not a single monetary union in the past came about because of a recognition of economic benefits of the union. In all cases the integration was driven by political objectives" (De Grauwe 1993:656). Of the half dozen currency unions that have been attempted among states of any significant size during the twentieth century, one-BLEU, founded in 1922-grew out of the security needs of a small and vulnerable mini-state; and four others-the ECCA, the CFA Franc Zone, the now defunct EAC, and southern Africa's CMA-derived from arrangements initially imposed by colonial powers.

And what of Europe's newly established Economic and Monetary Union, which began in January, 1999? After decades of debate, it is by now clear that the purely economic case for EMU is inconclusive at best. The real issues, most observers concur, are undoubtedly political, relating first and foremost to the European Union's declared goal of an "ever closer union." To its critics as well as to its advocates, Europe's newborn currency, the euro, is seen as a harbinger of eventual political integration. In the words of one careful survey: "Although there are surely economic benefits to be expected from a monetary union, the main driving force for [EMU's] resurgence remains the quest for the political integration of Europe.... The main objections to monetary union have also been largely political" (Fratianni et al. 1992:1-2). In brief: economics matters, but politics matters more.

Political Analysis . What politics? Most relevant are the direct political benefits of monetary sovereignty, all of which are likely to be compromised by the creation of a common currency. A distinct national money may not be an essential attribute of state sovereignty, but along with the raising of armies and the levying of taxes it has long been regarded as essential. As one observer has argued, with only a touch of sarcasm:

A government that does not control money is a limited government... No government likes to be limited.... Governments simply must monopolize money if they are to control it and they must control it if they really are to be governments (O'Mahony 1984:127).

It is easy to see why a monetary monopoly is so highly prized. Genuine power resides in the privilege that money represents. Apart from the instrument that a monetary monopoly provides to help manage the macroeconomic performance of the economy, three direct political benefits are derived from a strictly national currency: first, a potent political symbol to promote a sense of national identity; second, a potentially powerful source of revenue to underwrite public expenditures; and third, a practical means to insulate the nation from foreign influence or constraint. All are important elements of the fundamental purpose of the state: to permit a community to live in peace and preserve its own social and cultural heritage.

At the symbolic level, a national currency is particularly useful to rulers wary of internal division or dissent. Centralization of political authority is facilitated insofar as citizens all feel themselves bound together as members of a single social unit--all part of the same "imagined community," in Benedict Anderson's apt phrase (1991). Anderson, a cultural anthropologist, stresses that states are made not just through force but through loyalty, a voluntary commitment to a joint identity. The critical distinction between "us" and "them" can be heightened by all manner of tangible symbols: flags, anthems, postage stamps, public architecture, even national sports teams. Among the most potent of these tokens is money, as Italian central banker Tommaso Padoa-Schioppa has noted:

John Stuart Mill once referred to the existence of a multiplicity of national moneys as a "barbarism"... One could perhaps talk of a tribal system, with each tribe being attached to its own money and attributing it magical virtues... which no other tribe recognizes (Padoa-Schioppa 1993:16).

Money's "magical virtues" serve to enhance a sense of national identity in two ways. First, because it is issued by the government or its central bank, a currency acts as a daily reminder to citizens of their connection to the state and oneness with it. Second, by virtue of its universal use on a daily basis, the currency underscores the fact that everyone is part of the same social entity--a role not unlike that of a single national language, which many governments also actively promote for nationalistic reasons. A common money helps to homogenize diverse and often antagonistic social groups.

A second benefit of a national currency is seigniorage--the capacity a monetary monopoly gives national governments to augment public spending at will. Technically defined as the excess of the nominal value of a currency over its cost of production, seigniorage can be understood as an alternative source of revenue for the state beyond what can be raised via taxation or by borrowing from financial markets. Public spending financed by money creation in effect appropriates real resources at the expense of the private sector, whose purchasing power is correspondingly reduced by the ensuing increase of inflation--a privilege for government if there ever was one. Because of the inflationary implications involved, the process is also known as the "inflation tax."

Despite the economic disadvantages associated with inflation, the privilege of seigniorage makes sense from a political perspective as a kind of insurance policy against risk--an emergency source of revenue to cope with unexpected contingencies, up to and including war. Decades ago John Maynard Keynes (1924) wrote: "A government can live by this means when it can live by no other." Generations later another British economist, Charles Goodhart (1995:452), has described seigniorage as the "revenue of last resort"--the single most flexible instrument of taxation available to mobilize resources in the event of an sudden crisis or threat to national security. It would be the exceptional government that would not wish to retain something like the option of an inflation tax.

Finally, an important political benefit is also derived in a negative sense--from the enhanced ability a national money gives government to avoid dependence on some other provenance for this most critical of all economic resources. A national monetary monopoly draws a clear economic boundary between the state and the rest of the world, promoting political authority. The more effectively a government is able to enforce its monopoly within its own territorial frontiers, the better it will be able to insulate itself from outside influence or constraint in formulating and implementing policy.

That sovereign states might use external monetary relations coercively, given the opportunity, should come as no surprise. As political scientist Jonathan Kirshner has reminded us: "Monetary power is a remarkably efficient component of state power.... the most potent instrument of economic coercion available to states in a position to exercise it" (1995:29,31). Money, after all, is simply command over real resources. If a nation can be denied access to the means needed to purchase vital goods and services, it is clearly vulnerable in political terms. The implication is simple: If you want political autonomy, don't rely on someone else's money.

Although generally ignored by OCA theory, these three political benefits are all obviously affected by a decision to create a common currency. In effect all three are diluted at the national level and recreated at the group level, to be shared and in some manner managed collectively by the partner countries involved. Though this may on balance represent net gain for the group as a whole, for each partner individually it necessarily implies a significant degree of loss of sovereign power and privilege. Money's magical virtues can no longer be relied upon to enhance a unique sense of national identity. Insofar as value continues to be attached to loyalty to a distinct political community, it can no longer be promoted through the tangible symbol of a separate state-sponsored currency. Likewise, neither the revenue of last resort nor the political autonomy that a national money offers is as readily available when replaced by a common currency. Responses to crises are now subject to collective, not unilateral, decisionmaking; joint management, in turn, leaves each country individually more open than before to overt influence or constraint by its partners. Such considerations are bound to figure prominently in the calculations of governments.

Good or bad policy, in short, is not defined by strictly economic considerations alone; nor can the value of sovereignty be measured solely by its ability to deliver higher material standards of living. What is to be optimized is welfare in a much broader sense-in political-economy terms, welfare in the sense of a community's overall sense of identity and control of its own destiny. Even if the economic case for currency unification can be made, policymakers must still ask: What will be the effect on the "imagined community?" On the government's ability to handle unexpected emergencies? On the society's insulation from external coercion? To omit such key political concerns is to risk rendering analysis seriously deficient if not downright irrelevant.

Dollarization

The same point also applies to dollarization, which has become a topic of intense public debate in Mexico and several other Latin American nations since Carlos Menem of Argentina spoke out in favor of the approach in early 1999. In 2000, Ecuador and El Salvador actually did choose to dollarize. Should other Latin Americans follow the Ecuadorian and El Salvadorian examples, largely abandoning their local monies in favor of the greenback (or perhaps a future joint U.S.-Canadian currency)? Here too attention tends to remain misleadingly focused on technical economic considerations rather than on welfare in a broader political-economy sense; a rare exception is the chapter by Schuldt (2001) in this volume. Hence here too most analysis runs the risk of being deficient if not irrelevant.

Like the decision to create a common currency, dollarization in the economics literature tends to be addressed as an optimization problem limited mainly to issues of strictly material welfare. As in OCA theory, income gains or losses are evaluated from a single country's point of view - a country that might consider replacing its own money with the money of a larger and more successful partner (such as the United States). Typically, the foreign partner's money already circulates widely, albeit unofficially, within the domestic economy as a result of a market-driven process of currency substitution (in analytical terms, distinguished as informal dollarization). The question is: Should the government take the process to its seemingly logical conclusion by making the foreign money legal tender and officially withdrawing most or all of the local money from circulation (formal dollarization)?

On the positive side, as in OCA theory, broad benefits are expected from reduced transactions costs and enhanced usefulness of money in all its functions-the standard efficiency argument for monetary integration. Three additional gains could also be important. First is a reduction of administrative costs, since the dollarizing governments will no longer be obliged to incur the expense of maintaining an infrastructure dedicated solely to production of a separate local money. That saving would of course be of most interest to poorer or more diminutive sovereignties because of the diseconomies of small scale involved in monetary management. Second, as a supposedly irreversible institutional change, dollarization could also establish a firm basis for a sounder financial sector-a benefit that would be of particular value to states that previously have not enjoyed much of a reputation for price stability or fiscal responsibility. And finally, with dollarization there could be a substantial reduction of interest rates for local borrowers in countries that have not yet succeeded in establishing a sound credit rating in international financial markets. With luck, the reduction of interest rates could result in substantially higher levels of domestic investment and future economic growth.

Conversely, on the negative side, three costs tend to be emphasized in the economics literature. None of the three is insignificant. All are likely to be greater than if the two countries mutually agree to create a new common currency.

First, there is a loss of monetary autonomy, since the dollarizing country can no longer exercise unilateral control over its money supply or exchange rate. That is true with a common currency too, of course. But as compared with currency unification, the degree of loss is greater since dollarization as such implies no direct part in the making of monetary policy. With currency unification, each country presumably has a seat at the table where joint policy is made. With dollarization, unless based on an agreement with explicit provision for power-sharing, all authority is simply ceded to the partner country's central bank. The relationship is not one of parity but of hierarchy, with no promise at all that the dollarizing country's specific circumstances or needs would be taken into account when monetary decisions are made.

Admittedly, in practical terms, much of the country's monetary autonomy may already have been lost as a result of informal dollarization. The greater the degree of currency substitution that has already occurred, reflecting market pressures and preferences, the greater is the degree of constraint already imposed on a government's ability to manage macroeconomic conditions. As a recent IMF report concluded: "[Informal] dollarization can complicate ... monetary policy by introducing a foreign currency component into the money supply.... Dollarization may complicate stabilization and cause additional volatility" (Baliño et al. 1999:1,3). But retreat is one thing, complete surrender quite another. As compared with a strategy of keeping the national currency in circulation, which enables a government to retain at least a residual degree of control, formalization of the informal does imply an extra loss. Henceforth there will be absolutely nothing that the country can do to directly influence monetary conditions within its borders.

Second, there is a loss of a steady portion of seigniorage, since the dollarizing country must finance replacement of local currency still in circulation with a new issue of the partner's money. Circulating cash represents in effect a non-interest obligation of the central bank, matched on its balance sheet by interest-bearing reserve assets that are an ongoing source of revenue. Dollarization automatically terminates that revenue unless explicitly offset by some kind of agreed formula for seigniorage-sharing, as would logically be expected in the case of currency unification. In a classic contribution, Stanley Fischer (1982) divided this seigniorage loss into two parts: first the "stock cost," equal to the one-time expense of obtaining the new notes and coins needed to replace local currency in circulation; and second the "flow cost," representing the continuing flow of income foregone because the partner's money pays no interest to the local government. Both stock and flow costs will be smaller, of course, the greater is the degree of prior informal dollarization. But unless domestic circulation of local currency has already dwindled to the vanishing point, the seigniorage loss is unlikely to be trivial.

Finally, a dollarizing country is said to lose a lender of last resort, since in adopting a foreign currency it also formally gives up a central bank capable of discounting freely in times of financial crisis. Domestic banks thus are said to be more exposed to potential liquidity risks. To a large extent, however, this alleged cost is a red herring inasmuch as the loss of a lender of last resort can be rather easily offset on a unilateral basis. Dollarization, for example, reduces the overall need for international reserves, since a share of external transactions that previously required foreign exchange can now be treated as the equivalent of domestic transactions. A portion of the central bank's assets, therefore, could be dedicated instead to a public stabilization fund to help out domestic financial institutions under stress. Alternatively, a contingency fund could be built up over time from tax revenues, or flexible credit lines with foreign banks or monetary authorities could be negotiated, using future tax revenues as collateral. These possibilities stand in stark contrast to dollarization's other economic costs--the losses of monetary autonomy and seigniorage--which require explicit sharing agreements if they are to be contained or reduced significantly.

Once again, however, the question is not whether economic considerations of these kinds matter-of course they do-but whether they are all that matter. Here too one can legitimately argue that politics matters at least as much to governments, if not a good deal more. To repeat: good or bad policy cannot be defined by strictly economic considerations alone.

On the political side, as on the economic side, there are costs as well as benefits. For Schuldt (2001), the political advantages are paramount and suggest that most Latin American nations are certain to follow the examples of Ecuador and El Salvador within the next few years. Dollarization will be chosen because it serves the interests of dominant political elites and because it can be used by conservative governments as an instrument of social discipline, to force through unpopular structural reforms of the economy. But this reckons without the political disadvantages, which could be even more considerable. Involved here are the same three issues as with a common currency: losses of a symbol of national identity, a revenue of last resort, and a measure of political insulation. In contrast to currency unification, however, which implies some degree of parity and sharing among the partners, dollarization is a hierarchical relationship that offers little in the way of direct compensation and no necessary role at all in decisionmaking. The United States would enjoy not only all the status and prestige that goes with more extensive cross-border use of its money but also an enhanced capacity to mobilize fiscal resources when needed. What Latin government would not factor such vital concerns into its calculus of the pros and cons of abandoning its own currency?

Perhaps most importantly, the United States would gain a powerful instrument of influence or coercion over the dependent dollarized economy. Consider, for example, the case of Panama, which since its independence has used the U.S. dollar as legal tender for most domestic monetary purposes. Panama owes its existence to the United States, which encouraged secession from Colombia to facilitate construction of a canal across the isthmus, and has always maintained a special relationship with Washington. Although a national currency, the balboa, exists in principle, only a negligible amount of balboa coins actually circulates in practice. The bulk of the money supply, including all paper notes and most bank deposits, is accounted for by the dollar. In the late 1980s, Panamanians learned just how exposed they were to external coercion under this monetary arrangement.

In economic terms, most observers have rightly had only praise for Panama's currency dependence. Reliance on the dollar has created an environment of stability that has both suppressed inflation-a bane of most of Panama's hemispheric neighbors-and helped establish the country as an important offshore financial center. But in political terms Panama has been extremely vulnerable in its relations with Washington, which of course could sour at any time.

Such a moment came in 1988, following accusations of corruption and drug smuggling against General Manuel Noriega, the country's de facto leader. In March 1988, Panamanian assets in U.S. banks were frozen, and all payments and dollar transfers to Panama were prohibited as part of the Reagan administration's determined campaign to force Noriega from power. The impact was swift. Most local banks were forced to close, and the economy was squeezed by a severe liquidity shortage. The effect on the economy was devastating despite rushed efforts by the Panamanian authorities to create a substitute currency, mainly by issuing checks in standardized denominations that they hoped recipients would then treat as cash. The country was effectively demonetized. Over the course of the year, domestic output fell by a fifth.

As it happens, the sanctions turned out to be insufficient to dislodge Noriega on their own. Ultimately, in 1989, Washington felt it necessary to mount a military invasion that led to a temporary occupation of the country until a new, friendlier government could be installed. But there can be no doubt that the liquidity squeeze was painful and contributed greatly to Noriega's downfall. The message is obvious, as economist Lawrence Klein has prudently suggested:

Panama ... uses U.S. dollars for its monetary units. As long as relations remain cordial, this is not a bad arrangement .... But for Panama the risk price is very high for having the convenience of U.S. dollars. The small country would be in a better and more independent position if it had not let some of its monetary actions be governed by foreigners (Klein 1993:112-113).

No country concerned for its political sovereignty is likely to ignore Klein's point. Are we really to believe that Latin governments will optimistically discount the high risk price involved?

Consider now the sustainability of a monetary union. In principle, monetary union, whether via currency unification or dollarization, is supposed to be permanent-a formally irrevocable commitment. But if the history of world politics teaches us anything, it is that in relations between sovereign states one can never truly say never. No inter-state commitment, no matter how seemingly firm, may be regarded as truly irrevocable in practice. Hence we must also look beyond the initial decision to the unfolding of subsequent events. What conditions determine whether a commitment to a currency union or dollarization, once made, can be successfully sustained over time? Again, each of the two approaches can be taken up separately.

Currency Unification

We start once more with currency unification. Some currency unions have been successfully sustained for decades, including BLEU, ECCA, CFA, and CMA among those already mentioned. (The jury is of course still out on EMU.) But many others, by contrast, have gradually eroded over time or even totally failed, including not only the East African Community but also the Soviet Union's ruble zone and other former federations like Czechoslovakia and Yugoslavia, which disintegrated almost as soon as their constituent members gained political independence. Much the same was also true of the nineteenth century's two major experiments in currency unification, the Latin Monetary Union (LMU) and Scandinavian Monetary Union (SMU). Each functioned more or less effectively until World War I (a not inconsiderable achievement), yet ultimately both were formally terminated. If we look carefully at this variety of experience, we find that here too politics plays the leading role in determining final outcomes.

Economic variables offer little assistance in explaining the sustainability of monetary unions. As I have noted elsewhere (Cohen 2000), for every one of the characteristics conventionally stressed in OCA theory, there are contradictory historical examples--cases that conform to the expectations suggested by OCA theory and others that do not. None seems sufficient to explain observed outcomes. This is not to suggest that economic factors are therefore unimportant. Clearly they do matter insofar as they tend, through their impact on economic welfare, either to ease or exacerbate the challenge of sustaining a common currency. But equally clearly, more has gone on in each case than can be accounted for by such variables alone.

Nor is much help offered by an analysis of institutional design--that is, the legal provisions concerning issue of currency and management of monetary policy. Such organizational formalities have differed sharply in various cases. Only in three instances, apart from fragmenting federations, have members at any time relied exclusively on a true common currency--in the East Caribbean and, more briefly, in East and southern Africa. In all the others, including EAC after 1967 and CMA after 1974, arrangements have featured national or regional currencies that were officially linked together to one extent or another. And in parallel fashion monetary institutions have also varied greatly, ranging from a single central authority in two cases (ECCA and, before the mid-1960s, EAC) to two regional authorities in one case (CFA) and to separate national agencies in all the others (including the ruble zone and other such federations after their break-up). No systematic relationship is evident, however, between these organizational differences and the success or failure of various monetary alliances.

In principle, such differences might be thought to matter insofar as they affect the net costs of compliance or defection by individual states. Recent theoretical literature on transactions costs emphasizes the key role that institutional design can play in promoting credible commitments, by structuring arrangements to match anticipated incentive problems (North 1990). From this perspective, creation of a single currency would appear to be superior to a formal linking of national currencies because of the higher barriers to exit involved: the greater cost of reintroducing an independent money and monetary authority.

That was also the conclusion of policy discussions of alternative strategies for EMU before its creation, which directly addressed the relative merits of full currency unification versus a simple exchange-rate union (Gros and Thygesen 1992:230-233; von Hagen and Fratianni 1993). Most analysts express doubt that a system retaining existing currencies and central banks, no matter how solemn the political commitments involved, would be as credible as a genuine joint money, precisely because the risk of reversibility would presumably be greater. The implication is that compliance mechanisms are likely to be weaker to the extent that governments continue to exercise any control at all over either the price or quantity of their currency. Thus one might have expected to see in historical experience a direct correlation between the degree of centralization of a monetary union and its practical sustainability over time. In fact, however, no such relationship can be found. Contradictory examples abound.

The lesson of history is clear. The degree of centralization surely must matter insofar as it influences the potential cost of exit. But high barriers to exit or not, the evidence is that commitments can be--and, indeed, frequently have been--broken when governments decide it is in their interest to do so. Institutional design is probably no less important than economic characteristics. But it is equally clear that there is still something else at work here that overshadows them both. That something, of course, is politics.

From a political perspective, two characteristics stand out as crucial to the fate of monetary unions. One, suggested by traditional realist approaches to international-relations theory, is the presence or absence of a powerful state willing and able to use its influence to keep such an arrangement functioning effectively on terms agreeable to all. The other, suggested by more institutional approaches, is the presence or absence of a broad constellation of related ties and commitments sufficient to make the loss of policy autonomy, whatever the magnitude of prospective costs, seem basically acceptable to each partner. The first, which implies a degree of subordination as well as a sharing of monetary sovereignty, calls for a locally dominant country--a "hegemon"--and is a direct reflection of the distribution of inter-state power. The second calls for a well developed set of institutional linkages and reflects, more amorphously, the degree to which a genuine sense of solidarity--of community--exists among all the countries involved. Judging from the historical record, it seems clear that one or the other of these two factors is necessary for the sustainability of monetary union among sovereign states. Where both are present, they are a sufficient condition for success. Where neither is present, unions tend to erode or fail.

In short, when it comes to sustaining currency unification, the issue is only secondarily whether countries meet the traditional criteria identified in OCA theory or whether monetary management and currency issue happen to be centralized or decentralized. The primary question is whether the necessary political conditions exist: either a local hegemon or a fabric of related ties with sufficient influence to truly neutralize the risk of exit. Sovereign governments require incentives to stick to bargains that turn out to be inconvenient. The evidence from history suggests that these incentives may derive either from side-payments or sanctions supplied by a single powerful state or else from the constraints and opportunities posed by a broad network of institutional linkages. One or the other of these political factors, it appears, must be present to serve as an effective compliance mechanism.

Dollarization

With dollarization, the leading role of politics is even more nakedly obvious. Since the relationship is inherently hierarchical, the presence of a dominant state is guaranteed. The only question is what, if anything, the hegemon will do when faced with the prospect of exit by a subordinate partner. Will the hegemon actively deploy its influence to sustain use of its currency or will it remain passive, more or less calmly acceding to its partner's altered preference? Much depends on how vital the dependent partner is felt to be to the dominant country's political or economic interests.

Two historical examples will suffice. One involves the Common Monetary Area, whose origins go back to the 1920s when South Africa's currency, now known as the rand, became the sole legal tender in three of Britain's nearby possessions, Bechuanaland (later Botswana), British Basutoland (later Lesotho), and Swaziland, as well as in the United Nations trust territory of South West Africa, previously a German colony (now Namibia). In effect, all of South Africa's four closest neighbors were officially dollarized. Once decolonization began in the region, however, the monetary arrangement began to loosen, despite formalization first in 1974 as the Rand Monetary Area and later in 1986 as the CMA, as distinct national currencies were introduced by each of the junior partners. Diamond-rich Botswana eventually dropped out altogether, preferring to promote its own national money. Likewise Swaziland, even though remaining linked to the rand, ultimately chose to withdraw the legal-tender status of South Africa's currency within its own borders. As a result, today only Lesotho and Namibia remain formally dollarized insofar as the rand continues to circulate in each country alongside the national currency as legal tender.

Why have Lesotho and Namibia remained dollarized while the others have not? Essentially it is because South Africa has been willing to bribe them to avert defection. Side-payments are provided in two forms. First is direct compensation for the seigniorage the two countries forego by permitting circulation of an amount of rand within their domestic economies. Compensation is based on an estimate of the income that would accrue to each if they instead held reserves of equivalent amount invested in rand-denominated assets. And second is an assurance that the South African central bank will stand as lender of last resort for their domestic banking system in the event of need. Pretoria's willingness to make such critical concessions is clearly related to the importance the government attaches to good relations with its immediate neighbors. It is also the only reason why any degree of dollarization at all is still sustained in the southern African region.

The second historical example involves Liberia, which from the time of its independence in 1847 has always, like Panama, had a special relationship with the United States. During World War II, as the United States built up its military presence in the country, Liberia made the U.S. dollar sole legal tender, replacing the British West African coinage that had previously dominated the local money supply. Though supplemented, beginning in the 1960s, by a limited issue of small-denomination Liberian coins (also named the dollar), America's greenback maintained its dominant role until the mid-1980s, when political turmoil and fiscal deficits led the Liberian authorities to issue large amounts of higher-denomination coins as well as notes--a classic example of a government resorting to seigniorage as a revenue of last resort. Also true to form, Gresham's Law quickly went to work, and by the end of the 1980s U.S. dollars had almost completely disappeared from circulation, though the monetary agreement with Washington remains nominally in effect. Technically, the greenback is still the principal currency of Liberia, though it no longer actively circulates much as a medium of exchange. In practice, dollarization has not been sustained, and for one simple reason-Washington has not felt particularly moved to do anything about it. Liberia simply does not rank very high among U.S. vital interests around the globe.

What are the implications of all this for Canada and Latin America? Highlighting the political dimension of monetary union suggests a paradox of sorts for both our northern and souther neighbors. Politics would appear to favor sustainability but to hinder creation. The problem is not how to preserve a common currency for Canada or dollarization for Latin Americans, once such a commitment is made. The real challenge is how to achieve the commitment in the first place--how to get from Here to There. Political barriers could prove to be insurmountable.

In the case of Canada, for instance, both political conditions needed to sustain a common currency would appear to be satisfied. Certainly there already exists a sufficiently broad constellation of institutional linkages between Canada and the United States-manifested most obviously in NAFTA-to ease the risk of possible unilateral exit by either partner. If not yet a single community, the two countries do not lack for a significant sense of solidarity. Likewise there already exists a locally dominant country, the U.S., with an undoubted ability to keep a joint arrangement functioning effectively if it so chooses. Once a mutual commitment to currency unification is made, it should not be particularly difficult to ensure compliance on both sides.

But there, of course, is the rub. Can the necessary initial commitment be attained? For Canada, clearly, the key issue is how monetary policy would be made if North America's two dollars are merged. Will Ottawa be able to negotiate a share of power over a joint currency sufficient to compensate Canadians for the political losses involved-not least, the loss of one of the most tangible symbols of their carefully guarded distinctiveness as a nation? Conversely will the much larger United States, so accustomed to having its own way in monetary affairs, be prepared to cede much more than token influence to Canada? It is not difficult to imagine that the gap between the two countries over this issue might turn out to be simply too wide to bridge.

Similarly, in the case of Latin America, the basic condition needed to sustain dollarization-a committed hegemon-would also appear to be satisfied. Few regions of the world are seen in Washington as more vital to U.S. interests. If the U.S. and Latin governments formally agree to replace local currencies with the greenback, there seems little doubt that Washington would subsequently make every effort to keep the arrangement from failing. But again, the question is: Can the necessary initial commitment be attained? For Latin America, unlike Canada, the key issue would probably not be a direct role in decisionmaking. As junior partners in a distinctly hierarchical relationship, Latin Americans could not realistically aspire to seats at the Federal Reserve. Rather, attention is more likely to focus on such matters as seigniorage and lender-of-last-resort facilities as a way of gaining compensation for the obvious political risks involved in becoming other Panamas writ large. But will Washington be willing to play the role of South Africa, offering either a share of seigniorage revenues or a safety net to Latin American banking systems? Here too it is not difficult to imagine an unbridgeable gap between negotiating positions.

Yet these are precisely the gaps that will have to be addressed if dreams of currency unification or dollarization are to become reality. They are what is truly at stake today. Easily overlooked by strictly economic analysis, such political considerations will be decisive in determining whether monetary union is ultimately possible in the Western Hemisphere.

Anderson, Benedict (1991), Imagined Communities: Reflections on the Origins and Spread of Nationalism, revised edition (London: Verso).

Argy, Victor and Paul De Grauwe, eds. (1990), Choosing an Exchange-Rate Regime: The Challenge for Smaller Industrial Countries (Washington: International Monetary Fund).

Baliño, Tomás J.L., Adam Bennett, and Eduardo Bonrensztein (1999), Monetary Policy in Dollarized Economies (Washington: International Monetary Fund).

Cohen, Benjamin J. (2000), "Beyond EMU: The Problem of Sustainability," in Barry Eichengreen and Jeffry A. Frieden, eds., Political Economy of European Monetary Integration, second edition (Boulder, CO: Westview Press).

Courchene, Thomas J. and Richard G. Harris (2001), ""North American Monetary Union: Analytical Principles and Operational Guidelines," this volume.

De Grauwe, Paul (1993), "The Political Economy of Monetary Union in Europe," The World Economy 16: 6 (November): 653-661.

Fratianni, Michele, Jurgen von Hagen and Christopher Waller (1992), The Maastricht Way to EMU, Essays in International Finance No. 187 (Princeton: International Finance Section).

Fischer, Stanley (1982), "Seigniorage and the Case for National Money," Journal of Political Economy 90: 2 (April): 295-313.

Frieden, Jeffry A. (1993), "The Dynamics of International Monetary Systems: International and Domestic Factors in the Rise, Reign, and Demise of the Classical Gold Standard," in Jack Snyder and Robert Jervis, eds., Coping with Complexity in the International System (Boulder, CO: Westview Press): 137-162.

Goodhart, Charles A.E. (1995), "The Political Economy of Monetary Union," in Peter B. Kenen, ed., Understanding Interdependence: The Macroeconomics of the Open Economy (Princeton: Princeton University Press): ch. 12.

Gros, Daniel and Niels Thygesen (1992), European Monetary Integration: From the European Monetary System to European Monetary Union (London: Longman).

Grubel, Herbert (2001), "The Case for the Amero," this volume.

Honkapohja, Seppo and Pentti Pikkarainen (1994), "Country Characteristics and the Choice of the Exchange Rate Regime: Are Mini-Skirts Followed by Maxis?," in Johnny Akerholm and Alberto Giovannini, eds., Exchange Rate Policies in the Nordic Countries (London: Centre for Economic Policy Research): ch. 3.

Keynes, John Maynard (1924), Tract on Monetary Reform, reprinted in The Collected Writings of John Maynard Keynes, Vol. 4 (London: Macmillan, 1971).

Kirshner, Jonathan (1995), Currency and Coercion: The Political Economy of International Monetary Power (Princeton: Princeton University Press).

Klein, Lawrence R. (1993), "Some Second Thoughts on the European Monetary System, Greek Economic Review 15:1 (Autumn): 105-114.

Krugman, Paul R. (1993), What Do We Need to Know about the International Monetary System?, Essays in International Finance No. 190 (Princeton: International Finance Section).

Mundell, Robert A. (1961), "A Theory of Optimum Currency Areas," American Economic Review 51: 3 (September): 657-665.

North, Douglass C. (1990), "Institutions and a Transaction-Cost Theory of Exchange," in James E. Alt and Kenneth A. Shepsle (eds.), Perspectives on Positive Political Economy (New York: Cambridge University Press).

O'Mahony, David (1984), "Past Justifications for Public Interventions," in Pascal Salin, ed., Currency Competition and Monetary Union (The Hague: Martinus Nijhoff Publishers): 127-130.

Padoa-Schioppa, Tommaso (1993), Tripolarism: Regional and Global Economic Cooperation, Occasional Papers No. 42 (Washington: Group of Thirty).

Rose, Andrew K. (2000), "One Money, One Market: The Effect of Common Currencies on Trade," Economic Policy 30 (April): 7-45.

Schuldt, Jürgen (2001), "Latin American Official Dollarization: Political Economy Aspects," this volume.

Tavlas, George S. (1994), "The Theory of Monetary Integration," Open Economies Review 5: 211-230.

von Hagen, Jurgen and Michele Fratianni (1993), "The Transition to European Monetary Union and the European Monetary Institute," Economics and Politics 5: 2 (July): 167-186.

1. An earlier version of this chapter, in French translation, appeared under the title "Dollarisation: la dimension politique," in L'Économie Politique 5:1 (January 2000), 88-112.