Sei sulla pagina 1di 21

The Costs and Benefits of Monetary Union

Michael Grabner
June 2003

Abstract
With the issue of euro notes and coins in January 2002, the move towards a monetary union
between twelve European states has been completed. Around the globe, increased economic
integration has led to renewed interest in the costs and benefits of currency unions, a topic first
studied in Robert Mundell’s (1961) seminal paper on “Optimum Currency Areas.” Building on
De Grauwe (2000, 2003), this paper aims at providing a concise survey of the advantages and
disadvantages of such arrangements as developed in the economic literature. Furthermore, the
implications of such a comparison are studied for the case of the European Monetary Union.

Contents
1 Introduction 1

2 The Costs 1
2.1 Shifts in demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2.2 Insurance against asymmetric shocks . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.3 Differences between countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

3 A Critique of OCA Theory 5


3.1 Differences between countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.2 Exchange rate manipulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.3 Devaluation and time consistency . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.4 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

4 The Benefits 10
4.1 Transaction costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.2 Exchange rate uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.3 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

5 Costs and Benefits Compared 14


5.1 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.2 Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

6 Conclusions 17
1 Introduction
After the dearth of publications following Mundell’s (1961) path-breaking article on optimal cur-
rency areas recent years have seen a pronounced rise in interest in this topic. This is due to the fact
that as the number of countries increases and their economies become ever more integrated, ques-
tions arise once again on whether there are economic benefits from establishing currency unions,
and if so, which countries should form such unions. For instance, the surge in international trade
in goods and assets should raise the transaction benefits from a common currency, and unions are
a more credible instrument for inflation control than simply fixing exchange rates.
Indeed, many countries around the world have decided to set up monetary unions, most notably
twelve of the European Union’s 15 member states. Similar arrangements are under consideration
in Africa and the Middle East, and dollarization (forming a currency union with America) has
been implemented by several smaller Latin America nations (see Alesina et al. (2002)).
Following the discussion in De Grauwe (2000, 2003) this survey reflects on the arguments
provided in favor of and against the creation of a monetary union, similar in style to a “cost-
benefit analysis”, and presents related empirical evidence.

2 The Costs
The costs of a monetary union stem primarily from the loss of an independent monetary policy,
i. e. the ability to control the flow of money in the economy by setting interest rates and exchange
rates. The national central bank will either be abolished or devoid of any real power, and monetary
policy will be conducted on a union-wide level.
The ability to influence exchange rates may be important for adjusting economy activity if
certain differences between countries exist, as well as in other circumstances. While other policy
instruments often exist, they are likely to be more painful, thereby constituting the “cost side” of
a monetary union. This chapter rests on the theory of “Optimal Currency Areas” 1 (OCA theory)
introduced by Robert Mundell2 in 1961, and further developed by McKinnon (1963), Kenen (1969)
and Alesina and Barro (2002).

2.1 Shifts in demand


Suppose that two countries, say France and Germany, form a monetary union, with a common
currency (the euro) and a common central bank. Suppose further that the preferences of con-
sumers in both countries shift from French-made to German-made products, which is referred to
as an asymmetrical demand shock. According to standard aggregate demand/aggregate supply
models, higher demand for German products will lead to higher output and lower unemployment
in Germany, and lower output and higher unemployment in France. Both countries now experience
adjustment problems; France struggles with higher unemployment and Germany with inflationary
pressures due to its economic boom. There are two mechanisms which would be able to automat-
ically restore equilibrium:
• Wage flexibility: In the case of flexible wages, unemployed workers in France will offer
their labor services at lower wages, while there will be upward pressure on wages in Ger-
many due to excess demand for labor. This allows price reduction and output expansion
in France, and implies the opposite for Germany. These mechanisms restore equilibrium by
making French products more attractive, thus increasing demand for them. Second-order
effects reinforce this tendency: The wage and thus price increases in Germany render French
products even more competitive, further raising aggregate demand for French products (vice
versa for Germany.)
• Labor mobility: If labor is mobile enough, then unemployed workers in France would
relocate to Germany, with its excess demand for labor. Wages would not need to rise in
Germany or fall in France, and there is not going to be any inflationary pressure in Germany
or additional unemployment in France.
1 Originally
called the theory of “Optimum Currency Areas”.
2 For
his work on the economics of exchange rate regimes Mundell was awarded the Nobel prize for Economic
Sciences in 1999.

1
Either of these mechanisms would be able to solve the adjustment problems. But what happens
in case neither of the two conditions are met? With sticky wages in France and almost no labor
flows between the two countries upward pressure on wages in Germany will eventually lead to
higher inflation. This will make French goods more competitive, thus raising aggregate demand
and restoring equilibrium in both countries. The price would be a permanently higher price level
in Germany.
If the two countries had not been in a monetary union, they would have been free to use
individual monetary policy to counteract the asymmetric shock. In particular, France could have
devalued its exchange rate with Germany, thereby increasing the attractiveness of its products.
Aggregate demand in France would rise, while German output loses competitiveness due to a dearer
exchange rate, resulting in a fall in demand. There would be neither unemployment in France nor
inflation in Germany.
Exchange rate manipulation is not an option in a monetary union; the loss of this instrument
and the economic problems possibly resulting from a negative demand shock can be considered a
cost to France of joining the monetary union with Germany. Similarly, Germany incurs a cost by
having to accept a higher rate of inflation than it would have preferred.
In Mundell’s view, sufficiently flexible wages or sufficiently mobile labor markets are necessary
for a monetary union to be sensible, since they allow to deal with asymmetric shocks in the absence
of national monetary policies. These conditions assume the role of exchange rates in providing for
some flexibility.3

2.2 Insurance against asymmetric shocks


Besides flexible wages and mobile labor, a system of income transfers should be in place to help with
negative asymmetric shocks. Otherwise the costs of a monetary union could increase considerably.
In the example above, where Germany experiences a positive demand shock and France a negative
one, an income transfer from Germany to France would be highly desirable to soften the economic
consequences in France. To actually work, however, this transfer scheme must be in the interest of
Germany as well; there are also several ways how to organize such a system. One principle should
be reflected in any of them: The adjustment mechanism via wage and labor movements should
not be prevented from operating by the income transfers. Otherwise the affected countries will not
return to equilibrium and the transfers would become permanent. This increases in importance
the more permanent the shock turns out to be. In the above example, transfers from Germany to
France would reduce the latter’s incentive to adjust their wage structure to return to equilibrium.
The prospect of income transfers creates a situation of moral hazard, common to all insurance
problems. Temporary shocks, e. g. asynchronous business cycles between France and Germany,
suffer less from moral hazard.
Broadly speaking, insurance schemes can be distinguished by whether they are organized by
the public or the private sector.

• Public insurance systems: There are two subcases. First, large parts of the two coun-
tries’ budgets are centralized on a European level. This could take the form of a European
government taxing all citizens and redistributing the revenues to both countries. In case of a
change of preferences as outlined above, tax revenues of the supranational government from
France would fall, those from Germany surge, while government expenditure - in the form
of unemployment benefits - would soar in France and decrease in Germany. Such a scheme
would automatically transfer income from Germany, with its higher output, to France, where
output decreased following the shock. This system allows for consumption smoothing in
both countries: no fall in consumption in France and no rise in Germany. This implies that
3 The case of symmetric shocks is quite different: Assume both France and Germany are hit by a negative demand

shock. If they form a monetary union, the common monetary authority can lower union-wide interest rates (since
money markets are perfectly integrated), thereby stimulating demand in both countries without creating imbal-
ances. Monetary union entails no additional costs of adjustment. For asymmetric shocks, the common central bank
would have to choose between cutting rates (thereby raising demand in France but fuelling inflation in Germany)
or increasing them (restraining inflation in Germany but worsening the situation in France). Note that devaluation
by France if there is no monetary union would mean exporting the problem to Germany, which is likely to retal-
iate, thereby rendering this option less appealing. The cooperation required in dealing with symmetric shocks is
institutionalized in a monetary union. This advantage of a union disappears with asymmetric shocks.

2
the costs of forming a monetary union are reduced for both countries. Germany benefits
because it can expect similar payments when it is hit by a negative shock itself. The trouble
with this scheme is the moral hazard it creates. On a national level, central budgets often
redistribute income from well-off regions to poor ones, which reduces incentives to adjust
and leads to permanent disequilibrium (and permanent payments). On the other hand, the
existing EU budget is tiny compared to EU output, and a centralization of national budgets
is not expected to happen any time soon. This insurance scheme is thus not available for the
European Monetary Union.
The second way to organize public insurance systems is by relying on automatic stabilizers in
national budgets: A reduction in output in France causes tax revenues to fall and unemploy-
ment spending to rise, which increases the budget deficit and public debt. But it also allows
for consumption smoothing through intergenerational redistribution - the consumption level
remains constant in the face of an output decline at the cost of not being able to raise con-
sumption in future when output increases. The opposite happens in Germany - the output
increase is saved for bad times instead of being consumed on the spot. Automatic stabilizers
in the budget allow negative shocks to be dealt with at lower cost, thus reducing the costs
of the monetary union. This mechanism also reduces the scope for moral hazard, since ever-
rising budget deficits create unsustainable debt levels, which will force national governments
to adjust. However, this creates a problem for countries with high levels of public debt, since
they will not be able to let their automatic stabilizers work in case of negative shocks for fear
of increases in the budget deficit. This insurance mechanism will thus often not be available
when it would be most needed.
• Private insurance systems: Insurance can also be provided through financial markets.
Suppose that French and German markets are completely integrated - a single market for
bonds, equities and banking. A negative shock in France lowers French firms’ profits and thus
their share prices. Since French shares are also held by German residents, they assume part
of the burden. Conversely, rising share prices of German firms also benefit French investors
holding German shares. Both countries share the risk of negative shocks, so losses from
asymmetric shocks are spread over both economies. The same is true for bond markets.
Unfortunately, while this system reduces moral hazard, it provides little relief for the poor
unemployed in France who are unlikely to hold German financial assets. In the absence of
any public scheme a majority of people would lack coverage in this system.
Rigidities in the labor markets combined with insufficient insurance schemes are thus likely to raise
the costs of monetary union.

2.3 Differences between countries


Countries generally differ in their preference towards the inflation-unemployment trade-off.
Some prefer low rates of inflation at the cost of higher unemployment, others strive to increase
employment regardless of higher inflation. This affects the costs of introducing a monetary union,
as analyzed by Corden (1972) and Giersch (1973).
Consider two countries, Germany and Italy, say, and their respective Phillips curves, describing
the relation between wage changes and unemployment in the two economies. Changes in the wage
rate and price inflation are connected via labor productivity: If wages rise by 10% and labor
productivity by 5%, prices have to increase by the difference of the two, 5%, for the share of
profits in output to remain constant. The two countries are linked via purchasing power parity
- price levels for a given consumption basket, normalized by the exchange rate, must be equal.
This implies that Italy has to devalue its exchange rate if its inflation rate exceeds the German
one, in order to stay competitive. In a monetary union this exchange rate is fixed; inflation rates
must be equal. Otherwise, the products of the country with a higher inflation rate increasingly
lose attractiveness.
Suppose now that Italy is soft in inflation and Germany soft on unemployment. Italy will have
a higher rate of inflation than Germany, and a fixed exchange rate becomes unsustainable. The
costs of monetary union stem from the fact that both countries now have to choose less preferred
points on their Phillips curves to achieve equal inflation rates. Italy will have to accept lower
inflation and higher unemployment, while Germany must be content with higher inflation.

3
This analysis has not taken into account the importance of expectations of future inflation
rates for the stability of the Phillips curve. If future inflation is expected to rise, the curve shifts
upwards, leaving ever less free choice between inflation and unemployment. In the long run, the
trade-off may even disappear. This has important implications for the costs of monetary union, to
be studied below.

Institutional differences across labor markets in the EU abound. For instance, labor markets
are far less centralized in Britain than in Germany. These differences affect wage and price reac-
tions, such that even similar shocks can result in divergent developments. This can imply serious
consequences for the cost side of a monetary union.
One of the more popular macroeconomic theories dealing with the effects of labor market
institutions is due to Bruno and Sachs (1985). In their model the centralization of wage bargaining
is crucial to determine the reaction of the economy to supply shocks, such as an increase in oil
prices. If wage bargaining is strongly centralized (the level of “corporatism” is high), labor unions
internalize the inflationary effects of wage increases. Excessive wage claims entail higher inflation
and no rise in real wages, so there is no point is striving for excessively higher wages. In the face
of a supply shock, it is understood that the negative impact on real wages of the shock cannot be
compensated by a rise in nominal wages.
The effects are different if the level of corporatism is low. Since it is only a small player in
the market, each union knows that possibly excessive wage claims for its members will not affect
the aggregate price level much. On the other hand, if it were to restrain itself, its members’ real
wage would decline given that all other unions will push for higher nominal wages for their own
members. In equilibrium, everyone claims higher nominal wages, so the aggregate nominal wage
level rises faster than in a strongly centralized bargaining framework. Establishing wage moder-
ation after an inflationary supply shock is more difficult in such a non-cooperative, decentralized
environment. Calmfors and Driffill (1988) note that this is not a linear process: A second external-
ity enters as bargaining becomes ever more decentralized. If it happens at firm level, wage claims
influence the competitiveness of the firm and thus the employment fortunes of union members
themselves. Excessive wages entail a strong reduction of employment. Supply shocks may thus be
met with wage restraint. Countries where wage bargaining is either highly centralized or highly
decentralized are thus better situated to deal with supply shocks than those with intermediate
structures. Inflation and unemployment performance will be superior in those ‘extreme’ countries;
empirical investigations seem to support this. Following the 1979/80 oil price shock, for instance,
intermediate countries appear to have had problems reigning in inflation without reducing output.
Even for symmetric shocks, wage and price developments may differ across countries, and a
monetary union will make it more difficult to reduce those imbalances.

Even after decades of integration legal systems continue to be highly diverse across the European
Union. Some of these differences seriously effect the way markets work. Mortgage markets, for
instance, function differently due to legal reasons: Some countries offer more protection to banks
than others, which affects the mortgages’ risk, and the required amount of collateral. Interest
rates may also be floating or fixed for a given mortgage depending on legal circumstances. Similar
shocks, e. g. a change in interest rates by the European Central Bank, are transmitted differently
throughout the member states of the monetary union.
As another example, the way companies finance their operations varies widely across the EU.
Countries with an Anglo-Saxon tradition prefer the capital markets (equities and bonds) for financ-
ing investment projects, and those markets are well developed and liquid as a result. Countries
with a Continental legal history rely more heavily on the banking system, such that capital markets
are less sophisticated. Again, changes in interest rates affect the economy in different ways. A
hike in rates will induce large negative wealth effects in countries with widespread share and bond
holdings, since increases in interest rates lower stock and bond prices, while the impact in bank-
heavy countries will be less severe (bank lending will suffer from credit rationing). Consumption
and investment patterns across the union may thus react differently to the same shock.

Across the EU, countries in the periphery experienced higher GDP growth rates in recent times
than those at the core, which may entail troubles when all those countries belong to the same

4
monetary union. Different GDP growth rates can translate into different growth rates of imports
and exports. This can create trade balance problems - ever-growing surpluses or deficits. To avoid
chronic deficits, for instance, a country must increase the attractiveness of its exports. This can
be achieved by either devaluing the exchange rate or lower rates of domestic inflation. Monetary
union does not allow for the first option, so the country would have to follow deflationary policies;
but these tend to restrict output growth as well. To avoid these constraints a fast-growing country
may find it better not to enter the monetary union and keep a national currency, to allow for
depreciation as an adjustment mechanism.

Finally, the structures of countries fiscal systems may differ. Government budgets can be financed
in different ways: taxes, debt and money creation (i. e. seigniorage). If countries joining a mone-
tary union differ significantly in their financing methods, they may find their options constrained
after entry.
According to public finance theory, optimal financing requires that the marginal costs of raising
revenue from different sources of funding are equalized (otherwise revenue could be expanded
costlessly by switching from the channel with higher marginal cost to one with lower marginal
cost). Since the seigniorage channel is directly linked to inflation, this also implies that there will
be an optimal level of inflation for each country. In general, the more underdeveloped a fiscal
system, the more costs are created when tax collection is performed. Such countries may find it
optimal to rely heavily on money financing.
If more and less developed countries establish a monetary union, inflation rates will have to
converge, and most probably downwards. For a given level of government spending, this implies
that less developed, high-inflation countries will have to increase taxes, which negatively affects
welfare. For Europe, this is mostly an issue for southern countries, where empirical studies found
seigniorage revenue to be 2 to 3% of GNP up to the middle of the 1980s. Since then, however,
the decline in inflation in those countries has led to a significant decrease in seigniorage revenue as
well. The additional costs from joining the European Monetary Union in the 1990s may thus have
been quite small. But they could yet be an issue for new members such as countries from Central
Europe.

3 A Critique of OCA Theory


The analysis of costs associated with a monetary union discussed in the last chapter was based on
the theory of Optimal Currency Areas. This theory has attracted some criticism on different levels.
First, differences between countries may not be particularly critical. Second, the effectiveness of
exchange rate adjustment in correcting differences may be rather limited. Third, exchange rate
manipulation may do more harm than good.

3.1 Differences between countries


The first issue concerns the likelihood of asymmetric demand shocks, i. e. shocks concentrated
in a single country. Two opinions on the frequency of such shocks have been formulated, one
by the European Commission in its ‘One Market, One Money’ report of 1990, the other by the
distinguished trade economist Paul Krugman.
The European Commission believes that such asymmetric shocks are less likely in a monetary
union, due to the fact that trade between EU countries is largely intra-industry. Being based on
economies of scale and imperfect competition (via product differentiation), the same categories of
products are traded across countries. Therefore, demand shocks will affect different countries in
a similar manner. The abolition of trade barriers as a consequence of the single market policy
reinforces this dynamic. Asymmetric shocks will be transformed into symmetric ones, thereby
decreasing the costs of a monetary union.4
Krugman (1991), on the other hand, points to a different feature of trade based on economies
of scale which raises the importance of Mundell’s analysis: Trade of this kind leads to a regional
concentration of industry. Whenever barriers to trade are reduced, two opposing effects influence
the location decision of industries: Production can take place closer to final demand, but it also
4 The importance of trade structures was also highlighted by Kenen (1969).

5
gets more attractive to cluster activities to profit from economies of scale. Trade integration
may thus become more regionally concentrated instead of less, and asymmetric shocks will occur
more frequently. This phenomenon can be observed in the automobile industry, for instance.
Construction of automobiles is much more concentrated in the US than in Europe, and markets
in the US display a much higher degree of integration than Europe’s. If a similar development
prevails in European integration, sector-specific shocks may then turn into country-specific ones,
making exchange rate manipulation more attractive.
De Grauwe favors the Commission’s point of view, with the following line of thought: While
concentration of industrial activities as a consequence of trade integration cannot be disputed, in-
tegration also reduces the importance of national borders in a firm’s location decision. This implies
that clusters of specific industries may settle in regions where borders are crossed. Automobile
production may be concentrated not in, say, Germany, but an area covering southern Germany
all the way to northern Italy. In this case shocks to this sector could not be absorbed by the two
countries’ exchange rate. Industry-specific shocks will be transformed into region-specific ones,
with regions increasingly overlapping national borders as integration progresses. Integration will
reduce the capability of exchange rates based on nation-states to deal with asymmetric shocks.
Empirically, two issues turn up. First, is it true that monetary union furthers economic inte-
gration? Recent research, e. g. by Rose (2002), suggest that currency unions approximately double
trade flows between members, probably also as a consequence of deeper financial markets integra-
tion. Second, how does increased integration influence the (a)symmetry of shocks? Various papers,
for example by Frankel and Rose (1998), conclude that higher trade integration entails increased
correlation in economic activity - business cycles converge. Integration is hence likely to render
shocks more symmetric and less asymmetric. This is strengthened by the increasing importance of
services (about 70% of GDP in many EU countries), which are not as much affected by economies
of scale as industrial production is. A rise in trade integration will not imply regional clustering
of services, such that regional concentration of economic activity as a whole may subside. This
development might already be at work in the US.
In short, creation of a currency union might in itself initiate processes that favor its well-
functioning.

Another source of asymmetric shocks are nation-states, whose replacement by a central gov-
ernment would be a huge political undertaking. In its absence, while monetary policy would be
handled on a union-wide level by a common central bank - thereby unable to create asymmetric
shocks -, each country continues to formulate its own fiscal policy. While taxing and spending
powers of EU authorities amount to about 1.5% of aggregate GDP, individual states command al-
most 50% of output in spending and taxation. Changes in fiscal policy can create large asymmetric
shocks, which are by definition country-specific, and may lead to divergent economic developments.
Monetary union would require special arrangements to coordinate fiscal policies.
Besides budgetary means, other institutional differences can create asymmetric shocks, like
wage bargaining processes and legal markets (to be discussed below). One might argue that po-
litical unification would be highly desirable in such circumstances, to reduce the risks of painful
adjustment phases resulting from idiosyncratic disturbances. It is also possible, however, that the
existence of a monetary union will exert enough pressure on its members to increase fiscal central-
ization, or even move towards political union.

Given the significant differences in labor market institutions, it is interesting to ask whether
monetary integration might induce adjustments towards harmonization. Consider the usual trade-
off between higher wages and higher employment. Suppose that the government places more weight
on employment than the labor union. When the union selects a wage/employment combination
with less employment than favored by the authorities, the latter will react, for instance with expan-
sionary monetary policies to lift employment. Anticipating this behavior changes the constraint
of the union; in particular, an increase in real wages will reduce employment by less than if the
government did not intervene.
The willingness of governments to accommodate the union’s choices will differ across coun-
tries. Monetary union, however, reduces the ability of authorities to do so by eliminating national
monetary policy. This will tend to approximate the constraints unions face in different countries.

6
Furthermore, if unions attach similar weights to real wages and employment in their decision, wage
rates and employment levels may equalize across all member states. Of course, monetary policy is
just one among several tools governments have at their disposal to affect employment. Differences
in labor markets should be reduced, but they will not cease to exist. Moreover, the preceding anal-
ysis assumes highly centralized wage bargaining, but as mentioned above, the level of corporatism
also differs across countries. It is not clear how monetary union might affect these institutions. In
any event, differences in labor market institutions will persist, presumably resulting in divergent
wage and employment developments and the danger of serious adjustment problems.

Financial markets still differ across the EU, such that monetary shocks will be transmitted in
different ways; this is mostly, but not completely, due to different legal systems. Different mon-
etary policies in the past also played a role. For instance, investors in countries with a legacy of
high inflation are unwilling to hold long-term bonds, and markets for this type of asset are under-
developed. Government debt was primarily financed with short-term bonds, such that a change
in interest rates immediately affected the current budget situation. Due to the common monetary
policy in a currency union, such differences can be expected to vanish over time. Differences due
to varied legal systems, however, require further political integration.

It was claimed above that different growth rates of output may create trade imbalances, which
cannot be counteracted by exchange rate manipulation in a monetary union. Other policies will
restrict growth, such that monetary union imposes a cost on fast-growing countries by reducing
growth opportunities. This idea has little empirical support, though. For EU countries, data show
that slow-growth countries experienced both real depreciations and appreciations, while high-
growth countries were exclusively faced with real appreciations. Krugman (1989) provided this
explanation: Countries experiencing high growth rates will typically see their exports rise more
than their imports (their income elasticity of exports are higher than that for imports), so faster
growth need not entail chronic trade balance deficits. Exports will also increase without resorting
to real depreciations. Growth prospects when joining a monetary union are also positively affected
by capital flows: High-growth countries usually possess higher rates of capital productivity, encour-
aging investment inflows than can be used to finance current account deficits without a need for
devaluations. The elimination of exchange-rate risk following entry into a monetary union should
reinforce capital inflows, thus benefiting fast-growing countries as well.

3.2 Exchange rate manipulation


Joining a currency union implies the loss of exchange rate adjustments to counteract shocks to
the economy. The question arises whether such manipulation is actually an effective tool for
implementing corrections; or more precisely, whether changes to the nominal exchange rate are
passed on to the real exchange rate which governs trade patterns. If not, relinquishing exchange
rate flexibility on entry into a monetary union should not be counted as a cost.

Consider again the case of devaluations to correct for asymmetric demand shocks, in the
two-country model with France experiencing a negative shock. A devaluation increases demand for
French products, but it also increases the costs of imports. Production costs, and probably wages
as well will rise. Prices will increase and output fall, possibly inducing a new rise in nominal wages
(to restore purchasing power to workers). It is not possible to discern whether the devaluation’s
positive effects on the economy will disappear completely, this depends on various factors like
openness and wage formation; but there is empirical evidence for EU countries suggesting that few
beneficial effects will remain.
In other words, changes to the nominal exchange rate have just temporary effects on a country’s
competitiveness. In the long run the real exchange rate remains unaffected. However, short-term
effects also have to be taken into consideration. Suppose the negative demand shock produces a
trade balance deficit. Combining a devaluation and expenditure-reducing policies can succeed in
restoring the original output level as well as equilibrium in the trade account, but at the cost of
inflation. Without the devaluation, deflation will ensue before output and trade account are at
their initial levels; this will take longer if prices and wages are less flexible. Note that devaluation

7
on its own cannot achieve this - real imbalances cannot be corrected permanently with nominal
adjustments (the “neutrality of money” principle of classical economics). Nevertheless, the loss of
exchange rate manipulation as a policy instrument constitutes a cost of joining the currency union.
Empirical studies of Belgium’s 1982 or France’s 1982/83 devaluations suggest that the return to
equilibrium could have been more costly if alternative policies had to be implemented.

Now consider the use of devaluations to correct for different policy preferences, as high-
lighted above with the inflation/unemployment trade-off described by the Phillips curve. The loss
of exchange rate flexibility may imply that countries have to accept inflation and unemployment
levels other than those they prefer. However, as mentioned above, the monetarist critique claimed
that changing expectations of inflation will subject the Phillips relationship to instability. In the
long run, there is no trade-off; unemployment is given by its natural rate which is independent of
inflation. Countries wishing to form a currency union could thus equalize their inflation rates and
fix the exchange rate at no cost to their unemployment levels. Independent monetary policy is not
capable of dealing with uncomfortably high unemployment levels in the long term.
Nonetheless, the short-term trade-off still exists. Deflationary policies across the world in the
early 1980s resulted in higher unemployment; while it soon subsided in the US and the UK, it
continued to stay in continental Europe. Possibly, a high-inflation country joining a monetary
union has to incur considerable costs in terms of higher unemployment, even if temporarily. On
the other hand, these costs of deflationary policies will arise in any event, they are not specific to
becoming a member of a currency union. The question is now whether the adjustment costs differ
if a country is part of such a union or not. This will be discussed below.
Another factor that varies between countries is labor productivity. If two countries with differ-
ent rates of productivity join a currency union, nominal wages in the country with lower growth
must be smaller. Should wage bargaining be centralized, such that wage levels converge, the coun-
try with lower productivity growth will lose competitiveness. Thus wage bargaining should not
take place on a union-wide level so long as productivity growth rates differ.

Up to now it has been assumed that monetary union will equalize inflation rates. Empirically, this
need not necessarily be true - inflation rates in the EMU differ quite a bit even four years after
introduction of the common currency. One explanation brought forward is called the Balassa-
Samuelson effect, where labor productivity differences transmit into inflation differentials. 5
There are traded and non-traded goods; traded goods follow the law of one price (equal prices
adjusted for the exchange rate), while non-traded goods (like services) don’t. Total inflation re-
flects prices of both goods. Productivity differentials in the traded-goods sector translate into
differences in aggregate wage levels for the countries, so total inflation rates can differ.
Inflation rate differentials in a monetary union due to heterogeneous productivity growth are
in fact needed to retain competitiveness of the traded-goods sector. Across the EU, countries with
lower income levels tend to witness higher productivity growth and hence also higher inflation.
Of course, inflation differentials may also be caused by different demand situations, affecting a
country’s or region’s competitive position and possibly inducing adjustment conflicts.

3.3 Devaluation and time consistency


The idea of credible policies and time consistency, where policymakers and the private sector are
conceived as players in a strategic situation is due to Kydland and Prescott (1977) and Barro
and Gordon (1983). Agents react optimally to the strategies announced by the policymakers, and
the credibility of those announcements critically affect the strategies’ effectiveness and eventual
impact on the economy. In the context of a monetary union, this line of thought leads to criticism
concerning the discretionary use of exchange rate intervention.
The Barro-Gordon model combines the inflation-unemployment trade-off described by the
(expectations-augmented) Phillips curve with the preferences of authorities. Governments car-
ing more about unemployment are referred to as “wet”, while “hard-nosed” policymakers are more
concerned about inflation. The short-run inflation-unemployment trade-off allows a one-time gain
in welfare: Authorities announce they will follow a zero-inflation policy. Private agents believe
5 Balassa (1964) contains the basic insights.

8
this, and set their expectations accordingly. By unexpectedly raising inflation, the government
can now create a situation more to its liking, with some inflation but also lower unemployment.
Agents will now increase their inflation expectations, dislocating the short-run Phillips curve and
worsening the trade-off. Depending on the planning horizon of the policymakers, both the short-
term gains and these future losses will be taken into account when deciding about a move towards
“surprise inflation”. If short-sightedness prevails, agents’ expectations will change until a point
on the long-run vertical Phillips curve is reached again, with the same unemployment level as
before the unexpected rise in inflation, but with a much higher inflation rate. This point consti-
tutes an equilibrium with rational expectations and discretionary policy, where authorities select
the optimal rate of inflation given people’s current expectations. With discretionary policy and a
short-enough planning horizon, this is the only sustainable equilibrium. In any other configuration
authorities would have an incentive to “cheat”; the zero-inflation policy is time inconsistent (or not
incentive-compatible). Gaining reputation will only become a motive for policymakers if this game
is repeated often. Electoral realities, however, seem to favor short-term objectives for authorities.
The game can be extended to open economies. Imagine one ‘wet’ and one ‘hard-nosed’ gov-
ernment (Italy and Germany, say), linked via purchasing power parity. Italy, with its higher
equilibrium rate of inflation, will have to depreciate continuously. Italy could achieve much lower
inflation if private agents could be convinced that their government will not exploit the short-term
trade-off. Italy could announce that its exchange rate with Germany will be fixed from now on,
forcing Italy’s inflation down to Germany’s level. A surprise devaluation, however, would again
allow it to benefit once from the Phillips relationship, rendering the fixed exchange-rate policy not
credible.
If Italy was to join a monetary union with Germany, on the other hand, and leaving this
arrangement would be virtually impossible, Italy could in fact achieve as low an inflation rate as
Germany. Without an independent monetary policy, the authorities’ “wet” preferences cannot be
realized. Italy has “borrowed credibility” from Germany. It seems that this policy is a win-win
situation - large gains for Italy and no losses for Germany -, which made it popular especially in
Latin American countries. There are two caveats, however. First, only full monetary union will
suffice, the national currency has to be abolished (currency boards, for instance, fall short of this
requirement).6 Second, the authority in control of the union’s monetary policy must possess enough
“hard-nosed” credibility. This would certainly have been the case with Germany’s Bundesbank,
but a newly created institution, where both German and Italian interests are represented, may
lack this authority. The new equilibrium inflation rate may be higher than that of Germany
before union, such that Italy stills benefits while Germany would lose. This will affect Germany’s
incentives to form a currency union in the first place.
Regarding the cost side of monetary union, the previous analysis stresses the fragility of ex-
change rate manipulation as a policy tool for repeated intervention. Using it once generates expec-
tations which strongly influence its future effectiveness. Unlike in Mundell’s analysis, exchange rate
intervention cannot be used freely and frequently. Possible gains and losses have to be evaluated
every time policymakers wish to use it as an instrument. Suggestions for their complete abolition
may go too far though. As mentioned above, in some circumstances devaluations yielded success.
If coupled with drastic measures in other areas, devaluations may be perceived as singular events
and negative reputation effects could be kept under control. Forfeiting this policy instrument does
impose costs.

3.4 Notes
Mundell’s (1961) first contribution to the analysis of monetary union was rather skeptical about
the chances of such an arrangement in Europe. An article of 1973 sounds much more optimistic. He
argues that, first, monetary union is more efficient in dealing with asymmetric shocks than a system
of national currencies with exchange rate uncertainty. In case of a temporary asymmetric shock, if
capital markets are sufficiently integrated, citizens should find it easy to smooth consumption by
borrowing from neighboring countries. Exchange rate risk in the absence of currency union might
make such automatic flows less likely and insurance schemes more costly. Second, exchange rate
movements will be as much the source of shocks as an instrument to deal with them, as analyzed
6 Experiments with “dollarization” in some Latin American countries can be rationalized this way.

9
above. Empirical studies show that exchange rate movements are often not in accordance with
economic “fundamentals” such as interest rate spreads or output growth differentials. Psychological
factors such as herding behavior seem to be quite influential, and increased volatility in turn affects
the real side of the economy. Mundell came to regard monetary union as a means to reduce and
insure against asymmetric shocks.
But caution is necessary. Permanent, as opposed to temporary shocks will still require wage
and price adjustments, since permanent credit flows are unlikely to happen. And exchange rate
manipulation can be useful to deal with major shocks from other sources, even if their volatility
produces shocks itself.

How does openness affect the costs of monetary union? First, it influences the frequency of asym-
metric shocks, with the European Commission and Paul Krugman arriving at opposite results,
as mentioned above. Second, the effects of a devaluation differ depending on the openness of a
country. Both demand and supply effects are much more pronounced in a relatively open country
- exports as well as imported inflation increase to a greater extent. Combining the two effects,
it is unclear which country benefits more in terms of increased output following the devaluation.
But it is clear that the price level will react more strongly in the open economy. The greater price
variability resulting from frequent exchange rate manipulation can be seen as undesirable for the
economy. The loss of this instrument in a monetary union will hence impose lower costs on the
relatively open economies.
Considering both effects, it seems that the costs of monetary union tend to decrease as the
degree of openness rises. Note that this would not be true if the frequency of asymmetric shocks
surges as countries get more open, as suggested by Krugman. But this seems less likely to happen
(see the discussion above), and even if it were the case, the effect would be offset by the higher
costs of exchange rate manipulation in relatively open economies.

Concluding this chapter, it has been demonstrated that the inclusion of a few critical remarks on
OCA theory removes some of the pessimism regarding monetary unions inherent in this model.
Flexible exchange rates turn out to be less useful to deal with asymmetric shocks, and can cause
such disturbances themselves. The main insight of OCA theory remains, though: There are
political and institutional differences between countries which will not disappear or even be reduced
when they become members of a monetary union. These differences may induce divergent economic
developments, and eventually difficult adjustment processes. The loss of national currencies will
then be regarded as a considerable cost of the union. Thus EU countries took a “calculated risk”
when forming their monetary union, and OCA criteria remain relevant for future applicants or
other countries who contemplate the formation of such a union. Moves towards more flexibility
and political union could reduce the risks of high adjustments costs as well as the frequency of
asymmetric shocks.

4 The Benefits
While the costs of a monetary union accrue mostly at a macroeconomic level, the benefits are to
be found in microeconomic relations. Efficiency gains will arise due to reductions in transaction
costs and risk, both associated with a multitude of national currencies, now to be merged into a
single one.

4.1 Transaction costs


Elimination of the costs associated with currency exchange counts as a direct gain of monetary
union, which can be easily estimated as well as being highly visible. The European Commission
(1990) provides estimates of around 13 to 20 billion euros a year, about 0.25 to 0.5% of the combined
GDP of EU members (in 1990). The savings increase for small open and less developed countries
whose currencies are not a means of international payments. They are calculated by estimating the
resources used by the financial sector and firms in exchange-related activities. Using an underlying
theoretical model à la Baumol, a recent study by Mendizábal (2002) provides an upper bound
of 0.7% of Union GDP. These values might seem small, but this is just one of the gains from a

10
currency union. Note, however, that these gains also represent lost revenue to the banking sector.
But these values cannot offset each other; such transaction costs are a deadweight loss to the
consumers. Banks, on the other hand, have to look for other sources of revenue; the employees
who were involved in exchanging money can now presumably turn to some productive activity.
Note also that transactions across countries remain more expensive as long as payments systems
are not fully integrated.

An indirect gain from the elimination of transaction costs should turn up as increased price
transparency. Consumers should be able to compare prices across borders and shop around, thereby
increasing competition among producers. The questions is whether this effect will be strong enough
to be noticeable.
Currently, price discrimination abounds across the Union, with price differentials for some
brand-name products of about 30% of the cheapest price (EC Commission (2001, 2002)). Differ-
entials within countries are much lower by contrast. Similar results have been derived for the US
and Canada: Price differences between border towns are as high as between the east and the west
coast. Borders seem to be crucial in market segmentation and subsequent price discrimination.
This continues to be the case although there remain no explicit trade barriers (like import tariffs)
among US states or EU countries. Will monetary union have any effect on this in the EU?
Concerning retail shopping there are still huge impediments to exploiting price differences across
countries, namely transaction costs such as travelling. This prevents arbitrage. Retail business
within countries is often dominated by a few big chains (e. g. supermarket stores), serving the whole
market, which induces only small price differences. This national segmentation stems partly from
different regulations and cultures, partly from the fact that those stores are predominantly national
companies. Price differentials for electronics, on the other hand, are much higher both within and
across countries. One reason is the high degree of product differentiation which negatively affects
price comparisons. Monetary union is unlikely to affect this situation much. A common currency
may thus affect price convergence more by furthering economic integration, for instance of financial
markets, which in turn may lead to increased efforts towards legislative harmonization (e. g. for
regulation rules).

4.2 Exchange rate uncertainty


Uncertainty about future exchange rates implies uncertainty about a firm’s revenue. For risk-averse
individuals this represents a welfare loss, since the assumption of extra risk must be compensated
by extra returns. 7
One feature of the theory of the firm argues against this: Under uncertainty, the firm increases
output when the price turns out to be unexpectedly high, and reduces output if it is low. These
output adjustments increase profits for a high price and reduce losses for a low price, such that
profits will be higher on average if prices fluctuate symmetrically around a given price than when
this price is sure to occur. This model can be extended in a variety of ways, but the central message
remains: price uncertainty may increase average profits. For welfare consideration, this positive
effect must be balanced against the negative influence of uncertainty per se. The result remains
unclear, so the movement towards a single currency may raise or reduce firms’ welfare.
Alternatively, one can argue that greater exchange rate volatility also increases the probability
of huge profits. Exporting can then be considered as an option, to be implemented by the firm
depending on whether the exchange rate is favorable or not. As is known from option theory,
increased variability of the underlying asset (the exchange rate) increases the value of the option.
With high volatility the firm is better off. A similar argument can be constructed for consumers.
Consumer surplus is higher on average if prices fluctuate, because demand adjusts to different
prices. The desirability of such an arrangement for risk-averse consumers has to be compared to
the increased risk, and no unambiguous answer can be given. In general, gains from a reduction
in risk in a currency union presumably need to be found somewhere else.

Such gains may arrive via the price mechanism: Decisions about investment, production and
consumption in a market economy are based on information contained in prices. If exchange rate
7 Usually, however, firms are modeled as risk-neutral agents.

11
risk increases price uncertainty, the quality of such decisions will decline. Investment projects
abroad may turn unprofitable if the exchange rate differs from what was originally assumed; this
entails costs. Such errors can be expected to happen more frequently if exchange rate volatility
increases.
Note that this discussion revolves around real exchange rate uncertainty, i. e. uncertainty asso-
ciated with the inability of exchange rate changes to reflect price changes. The dollar appreciation
from 1980-85 surpassed inflation differentials, thus negating purchasing power parity. Due to this
‘misalignment’ firms’ profits became unpredictable; some had to close and aggregate output fell. A
few years later the dollar depreciated again, beyond its level before the appreciation. Such swings
inflict serious adjustment costs to the economy.
Introducing a common currency can reduce such uncertainty and hence adjustment costs. This
improves the price mechanism and therefore economic decision-making; but the efficiency gains are
hard to estimate. They are nonetheless quite important: note the unreasonable and highly costly
allocation decisions in countries with hyperinflation, another situation where the price system fails
to work.
Price and exchange rate uncertainty also affects the price mechanism in a different way: Higher
risk increases the real interest rate. This is because, for risk-averse investors, higher risk requires
higher returns, and because future returns will be discounted more heavily when uncertainty about
them increases. Higher interest rates reduce the efficiency of selecting between investment projects,
because they reinforce moral hazard and adverse selection problems. These two effects imply that
higher interest rates will entail riskier investment projects. Again, monetary union should reduce
price uncertainty and thus systemic risk.8

Concerning further benefits, the EC Commission report of 1990 used a neoclassical growth model,
allowing for dynamic economies of scale, to show that reduced exchange rate uncertainty will
increase economic growth. In the standard version, the long-run growth rate of output is
solely determined by population growth and (exogenous) technological innovation. A monetary
union, by eliminating exchange rate risk, is likely to reduce the real interest rate. This implies a
rise in the capital stock, and movement to a new equilibrium, with a higher stock of capital (and
output) per person. The growth rate increases during the adjustment period, but returns to its
old level once the new equilibrium has been reached. The productivity of capital, however, has
fallen, as indicated by its lower rate of return, i. e. the real interest rate.
Suppose now that there are dynamic economies of scale9 , for instance learning effects and
knowledge accumulation, such that the increased capital stock implies a rise in capital productivity
(more precisely, the productivity of capital per worker). Essentially, a lower interest rate would
in this case permanently expand production capabilities, thereby placing the economy on a higher
growth path. It is not surprising that official EU reports may want to highlight this possibility.
From an empirical point of view, these conclusions are less unambiguous (see below).

4.3 Notes
The previous sections mentioned several ways in which monetary union could positively affect trade
and economic growth. First, consider trade, which could be stimulated by reduced transaction costs
and the elimination of exchange rate uncertainty. Would empirical evidence lend support to
those claims? The first generation of studies used time series techniques to link bilateral trade
flows and measures of exchange rate variation, and found only weak or insignificant connections.
The second generation of econometric studies, pioneered by Andrew Rose, has arrived at quite
opposite conclusions. Rose (2000) used cross-section data, controlled for several other variables
known to affect trade flows (like income, distance, language) and found that trade doubles on
average between pairs of countries when they are part of a currency union. An effect of this size
8 Starting with Poole (1970), some economists, on the other hand, claim that reduction of exchange rate risk
inevitably leads to increased risk somewhere else in the economic system. An OECD (1999) study, for instance,
found that variability of output tends to be higher among members of monetary unions. Such results are essentially
similar to the effects of asymmetric output shocks in OCA theory, based on the loss of exchange rate intervention
for adjustment. Random shocks to the money market, however, will only affect the output of countries outside a
currency union, due to equilibrating liquidity flows.
9 An old idea formalized for the first time by Romer (1986).

12
created some surprise and initiated a wave of research trying to replicate and extend his results.
As surveyed in Rose (2002), most studies confirm his findings.
Output growth, on the other hand, may profit from a currency union indirectly by enhanced
trade, or directly through lower interest rates (as described above). Frankel and Rose (2002)
investigate the first channel and find that a one-percent increase in trade between members of a
currency union raises per capita income by 0.3 percent. Given that monetary union seems to imply
a surge in trade, countries in such a union can also expect a significant increase in output. For
the second channel no comparable empirical evidence exists. This can be rationalized in light of
the dubious effects of reduced exchange rate uncertainty on firms’ profits mentioned earlier. Lower
risk reduces the real interest rate, but also the expected return of investment. The total effect
on investment, and thus output, remains unclear. Furthermore, reduced exchange rate risk may
increase the riskiness of other parts of the economy, such that systemic risk remains unchanged.
Firms operating in a currency union may be exposed to similar amounts of risk as those outside,
such that the link between reduced exchange rate risk and investment proves weak (see footnote 7).

After forming a monetary union, the new currency is likely to be more important in international
financial markets than any of the single national currencies before (or even their sum), turning into
an international currency. Usage of the currency outside the union will presumably increase,
which yields additional benefits to member countries.10
First, the authority issuing the currency receives higher revenues. The Federal Reserve in the
US more than doubled its profits in 1999 because less than half the dollars it issued were used
domestically. These additional revenues enter the government’s budget, and can be used, for in-
stance, to finance a higher level of spending for a given level of taxes. The euro, for its part, is
increasingly used in Central and Eastern Europe, thus creating benefits for EMU members. But
note that the Fed’s total profits are less than 1% of US GDP, so the benefits of an international
currency accruing from this source may not be large. Second, an international currency creates
additional income for domestic financial markets. Foreign investors may strive to invest in assets
and issue debt denominated in this currency, boosting the banking sector and capital markets.
This effect may be larger, but is also harder to quantify. On the other hand, an international
currency is probably not a prerequisite for the creation of dominant financial markets, as the City
of London has shown: Pound Sterling is not a major international currency any more, but the
UK’s financial markets thrive.

As with the cost side, a relationship can be established between a country’s openness and the
benefits from joining a monetary union. It is reasonable to assume that transaction costs and
exchange rate uncertainty weigh more heavily in countries with higher trade integration. Elimina-
tion of these effects will lead to greater welfare gains in small, open economies than in large and
closed ones. This establishes a positive relation between openness and the economic benefits from
a currency union.

On a concluding note, the notion that monetary union will significantly raise the long-run out-
put growth rate seems rather optimistic; the benefits are likely to appear elsewhere, especially
in increased trade and enhanced economic integration. For the trade effect, in particular, more
empirical studies would be desirable to confirm earlier results, and to uncover the mechanisms
behind this large effect.

10 Note an additional benefit of monetary unions: They rule out speculative attacks on the currencies of particular

member countries, which could happen in basic fixed exchange rate regimes (and they did occur, for instance in
1992, when Britain and Italy were forced to leave the European Monetary System, the EMU’s predecessor).

13
5 Costs and Benefits Compared
The previous chapters outlined the costs and benefits for a country when joining a monetary union.
This concluding chapter aims at combining the insights derived above by comparing the advantages
and disadvantages in a single model, with an empirical focus on the European Monetary Union.

5.1 Theory
Recall the relations derived earlier linking costs respectively benefits with the level of openness of
a candidate country. Combining these yields a critical level of trade which determines whether it
will be useful for the country under review to join a currency union with its trading partners. Some
qualitative conclusions can be gained this way. The cost relation, for instance, depends critically on
one’s beliefs about the exchange rate being an effective instrument to correct diverging economic
developments between the member countries.
One extreme would be the monetarist view, according to which the exchange rate is completely
ineffective in such situations. And even if it were an effective tool, it would probably make the
country worse off. Thus there is little loss when national currencies are eliminated, and the critical
level of trade rendering a monetary union beneficial is small. A lot of countries worldwide would
stand to gain by forming such unions. The other extreme is represented by the Keynesian view,
with a world riddled with rigidities - sticky wages and prices, immobile labor -, such that the
exchange rate becomes a very useful instrument in dealing with disequilibrium situations. This
position is actually embodied in Mundell’s OCA theory discussed in chapter two. The critical level
of trade is high, so few countries would find it a good idea to join a currency union; indeed, large
countries (like the US) with a single currency should contemplate splitting into several monetary
zones to reap economic benefits. Since the early 1980s the monetarist view gained a lot of influence;
this may well have been part of the reason why the EMU really came into existence in the 1990s.
What would this analysis have to say about the optimality of EMU? Statistics for intra-EU
trade show varying degrees of openness across EU countries, with the Benelux countries and Ireland
being the most open, and Italy, the UK and especially Greece at the other end of the spectrum.
This implies that different EU countries will get different results from a cost-benefit comparison;
relatively closed countries will find they are not likely to form an optimal currency area with the
remaining EU. Still it will be hard to draw an exact line between member and non-member countries
of a European OCA; first, because other parameters have to be included in the calculation, such
as flexibility of markets and asymmetry of shocks. Second, joining a monetary union can have
benefits even for closed economies, as the discussion of credibility stressed. This would rationalize
the decision to join for traditionally high-inflation countries like Italy and Greece. For them, the
loss of a national currency might not be very costly, such that the critical trade level favoring union
would be low. A monetarist view might induce a desire to join even for countries with weak trade
connections, since it might make overall economic sense if benefits outweigh costs.

As pointed out in chapter two, increased price and wage flexibility tends to lower the costs of
monetary union. Therefore, other things equal, the minimum trade level that would favor joining
a union drops if rigidities diminish. The same applies for flexibility in the labor market. If the
single market would increase labor mobility, EMU would become more attractive for other EU
countries. But note that other effects of economic integration, such as regional concentration of
industrial activities, can move the cost-benefit analysis in the other direction.

Besides flexibility, the degree of asymmetric shocks (size and frequency) is a critical determinant
of a country’s attitude towards the establishment of a monetary union. Countries facing sufficiently
different demand and supply shocks (due to different industrial patterns) will find it costly to join
such a union.
OCA theory allows to link labor market flexibility and asymmetric shocks. Shocks are de-
scribed by the degree of ‘real’ divergence between regions (or countries) - the extent to which the
growth rates of output and employment tend to diverge as a result of asymmetric shocks (those
independent of a given monetary regime). According to OCA theory, countries subject to diver-
gent developments need more flexible markets to profit from monetary union, and to avoid serious
adjustment problems. Thus, a higher degree of ‘real’ divergence requires more flexibility. For a

14
given level of divergence, if flexibility is high enough, adjustment costs will not exceed the benefits
of joining a monetary union. If it is too low, countries should strive to retain some exchange rate
flexibility, otherwise there will be a heavy economic price to pay.
How would EU countries fare in such a scheme? Studies trying to implement this theory on an
empirical level seem to agree that the EU-15 is not an optimal currency area. 11 A monetary union
including all 15 member countries of the EU seems not advisable. However, the same studies show
that there is also agreement that a subset of the 15 EU countries does form an optimal currency area.
This area would contain at least Germany, France and the Benelux countries. Recent studies tend
to increase the number of countries that stand to benefit from joining a monetary union. Business
cycles, including those of Southern European countries, correlate stronger since the 1980s; the
frequency of asymmetric shocks has decreased (possibly due to EMS participation, which reduced
the scope of independent monetary policies, a major source of shocks). This confirms earlier
thoughts about the effects of economic integration. Other results challenge this core-periphery
view of currency union. Monetary instruments seem to be weak in affecting real variables across
the EU, such that the loss of these tools does not seem very costly, as the use of national monetary
policy to handle asymmetric shocks is quite limited. Moreover, most asymmetric shocks seem
to occur at sectional instead of national levels, which can not be dealt with by exchange rate
manipulation. Recent studies, altogether, seem more optimistic about the number of countries in
the EU which could benefit from a monetary union. The exact number is controversial (but the
whole EU-15 is still not optimal). The EU-5 or even EU-10 are possible candidates - note that
this is not because (labor market) flexibility is higher than for the EU-15; rather, real divergence
is lower. Clear-cut decisions are not feasible, however; there always remains scope for subjective
judgment. Regarding the relative positions of the EU-15 and the US, the degree of real divergence
(between states in the case of the USA) seems to be roughly identical (see Krugman (1993)). But
flexibility is much more pronounced in the US than across the EU; for instance, real wages respond
more to unemployment, and labor mobility is much higher (whether the US really forms an OCA
is another question).
Including thoughts delineated in chapter three may change the picture, however. Monetary
union, by eliminating national monetary policies, reduces a potential source of asymmetric shocks.
Credibility issues may make participation recommendable in the absence of other justifications.
It thus cannot be claimed that even the EU-15 wouldn’t gain from a monetary union. The costs
of such a union could obviously be reduced by either mitigating real divergence or increasing
flexibility. Real divergence seems hard for policymakers to influence, for instance if it stems from
regional specialization patterns. Political unification, though, by coordinating economic policy and
harmonizing diverse institutional arrangements would certainly reduce asymmetric shocks and thus
real divergence.12 The other strategy aims at increasing the flexibility of labor markets (real wages
and labor mobility). Reforms in this area may be tough to implement, but they are necessary for
a large monetary union to work.

How are costs and benefits going to evolve in the long run? Recall the discussion in chapter
three dealing with the effect of economic integration on economic convergence. First, OCA theory
would imply that increased economic divergence must be “counteracted” with higher levels of
trade integration for a monetary union to make sense - the higher costs of divergence need to be
balanced by higher benefits due to deeper trade integration. Second, according to the European
Commission view, trade integration will reduce real divergence. Currently, the divergence of the
EU-15 countries does not make them an optimal currency area. Over time, however, continuing
trade integration will work to reduce differences and asymmetric shocks, so at some point these
countries may indeed constitute an OCA. Monetary union will seem attractive to all EU member
countries, and full union seems inevitable.
Paul Krugman’s view, on the other hand, presents a less optimistic scenario. Higher trade
integration may in fact lead to increased economic divergence between countries, as industrial
activities concentrate regionally. There are two cases to consider: If the benefits of a currency
11 See, for instance, Eichengreen (1990), Bayoumi and Eichengreen (1993), De Grauwe and Vanhaverbeke (1993).

Of course, the European Commission (1990) disagrees.


12 How should labor unions be organized in such a monetary union? Acknowledging that shocks will most probably

occur at a sectoral (or regional) level in the future, flexibility will be of high concern, so wage bargaining should
take place in a decentralized way.

15
union due to higher integration increase more rapidly than the costs in terms of more asymmetric
shocks across countries, the experiment may yet turn out to be worthwhile; the EU-15 may become
an OCA eventually. If the costs of higher integration outpace the benefits, however, forming a
monetary union would never be rational according to OCA criteria. It may be that in this case a
reduction in trade integration would imply that the benefits again outweigh the costs, a strange
result. But even in Krugman’s scenario, monetary union could be optimal in the long run (and
recall that this scenario seems subject to both theoretical and empirical critique).
Finally, the criteria for joining a monetary union are somewhat endogenous: Trade increases
due to a union tend to harmonize business cycles and thus lower the costs of joining (see Frankel
and Rose (1998)). Also, simply joining the union, even if it is not an OCA, will speed up the
integration process and thus the move towards an OCA: Market segmentation relies heavily on
national currencies, as documented in chapter four. Joining in the hope of reaping the benefits
from an OCA shows self-fulfilling characteristics. The decision to join per se alters the cost-benefit
calculation.

5.2 Application
As a case study of different adjustment processes after asymmetric shocks De Grauwe compares
the reaction of the state of Michigan in the US to that of Belgium in Europe to the recession in the
early 1980s. Both areas were characterized by an older industrial structure, and are of similar size.
Unemployment rose more in both areas than in the US and Europe, respectively, as a whole. Higher
economic integration in the US actually implied a larger unemployment differential - Michigan is
home to large parts of the automobile industry. Adjustment in Michigan mainly took the form
of migration to other states, quickly helping to alleviate the unemployment problem. Belgium
resorted to real exchange rate depreciation, closing the unemployment gap, with some delay, at
the end of the 1980s. Real exchange rates moved little in the US (i. e. price developments were
roughly similar across states), while Belgium experienced almost no outward migration. Fiscal
redistribution helped improve the situation in Michigan; a similar scheme was not in place in the
EU then (it is still not in place now). Instead, Belgium turned to international capital markets,
creating an astonishing level of government debt. It seems like even small open countries like
Belgium, where the benefits of joining a monetary union should outweigh the costs, took some
risk. For future shocks national monetary policy will not be available to help the adjustment
process.

To what extent do the current members of the European Monetary Union, the EU-12, form
an OCA? This question is not just of academic value now that the union has been established;
the answer determines whether future developments will be smooth or if individual members will
voice discontent with the common monetary policy.13 Asymmetric shocks will pose difficulties for
the European Central Bank, which is in charge of monetary policy across the union, and member
states might be less than satisfied with policies that do not (and cannot) respect different economic
conditions. The enlargement of the EMU may intensify those problems; equally well, it could move
the region further towards becoming an OCA.
Enlargement could raise the number of member countries to a maximum of 25; new members
might consist of Denmark, Sweden and the UK, which have preferred staying outside when the
union came into existence, as well as several Central and Eastern European states who will join the
EU in 2004. Regarding the economic openness of these CEE applicants, their trade connections
(exports to the EU-15) are at least as important as those of current EMU members themselves,
they even exceed those of the three EU outsiders.14 The openness criterium would thus favor
enlargement. The asymmetry of shocks has been evaluated by Fidrmuc and Korhonen (2001), using
similar methodology as Bayoumi and Eichengreen (1993). Most countries show little correlation
of individual demand and supply shocks with average shocks in the euro area. Some have weak
negative demand shock correlations (except Lithuania and Latvia, where the negative correlation is
about -0.5). Insofar as those asymmetries stem from independent monetary policy, they will vanish
13 Not to mention the fiscal implications as set out in the “Stability and Growth Pact”, which does not bear any

relation to OCA theory.


14 Note that CEE countries are quite small - with the exception of Poland - and such states tend to be quite open.

But they are even more open than small EMU countries.

16
in a currency union. The low correlation in supply shocks (except in Hungary), on the other hand,
represents asymmetries such a union will presumably not reduce. For the UK, correlations are
small (and negative for demand, possibly reflecting its idiosyncratic monetary policy) while its
exports to the EU-15 are low (about 10% of GDP), indicating that it may not be part of an OCA
with the rest of Europe. CEE countries are far more open towards the EU, but suffer from larger
asymmetric shocks, and thus may not gain from joining EMU (however, if they wish to import
monetary stability by joining this would, of course, alter the cost-benefit comparison).
Present members of the EMU might also experience changes to their cost-benefit calculation
after enlargement: The larger union will need more time to become an OCA, and the average
economic development in the union, on which the ECB bases its policies, will shift. Some current
members could become “outliers” feeling that the common monetary policy does not fit their
national situation. This effect gets even more pronounced if it is assumed that integration increases
economic divergence (the Krugman scenario). The perceived costs of monetary union will increase
relative to the perceived benefits, and the strains inside the common monetary authority (the
ECB) as well as outside it, i. e. between member states will become stronger. There isn’t much
the ECB can do about this, it is impossible to adjust the interest rate to suit particular national
circumstances. Instead, member countries need to exploit other instruments to absorb asymmetric
shocks. Flexible labor markets are likely to prove most crucial in this respect.

While the United Kingdom has recently (yet again) decided not to join EMU for the time being
(see, for instance, The Economist, June 14-20), the economic case for becoming a member seems
to have become more favorable over the past few years. Although the UK remains relatively closed
towards the EMU countries (even the CEE countries engage in more trade), there is some evidence
for economic convergence (interest rate spreads keep getting lower, for instance) and Britain’s
flexible labor markets should make it easier to deal with future asymmetric shocks once inside the
EMU. But as pointed out above, supply shock correlation with the EMU area is low, so asymmetric
shocks have been significant in the past. The benefits of joining might be lower due to Britain’s
less pronounced openness, but the City of London is likely to enhance its position as Europe’s
major financial center, bringing benefits of its own. Overall, the UK should benefit from joining,
though some uncertainty remains. Subjective elements are likely to prove important in dealing
with this uncertainty, which might be crucial give that the decision to join will be subject to a
referendum (probably sometime after the next elections, planned for mid-2005). Since the UK
will have a substantial impact on the ECB’s policy once inside the union, other members might
again find their cost-benefit calculations altered, possibly to the worse, given Britain’s asymmetric
shocks relative to the euro zone.

Might Latin America be an optimal currency area? The turmoil of past decades and the failure
of individual countries’ experiments at stabilization (by pegs to the dollar or currency boards)
raises the question of whether full monetary union might be more beneficial. Regarding openness,
countries in Latin America are relatively closed in general, and most of their trade involves the
US or Europe, so intra-regional trade is very low. The literature on other criteria is sparse; there
seems to be some evidence that asymmetric shocks are quite large, and labor markets segmented,
reducing the possibilities of adjustment. On these grounds Latin America would not seem to
constitute an optimal currency area. Establishing a monetary union to achieve price stability is
also unlikely to work given the region’s past problems in reigning in inflation. The new institutions
created by such a union would lack the credibility for such a task. A full-blown monetary union
among Latin American states is not likely to be created soon; some smaller countries have opted
for dollarization (monetary union with America) instead for the same reasons. Since this implies
losing all monetary sovereignty, larger countries are less inclined to dollarize their economies.

17
6 Conclusions
The aim of this paper has been to provide a consistent and complete discussion of the various costs
and benefits associated with the formation of a monetary union. It is worth mentioning, however,
that this discussion focused on an economic cost-benefit analysis. Countries might also choose
to establish a currency union for political reasons. Indeed, as seems to be the case in Europe,
monetary union is meant by some to be the first step towards political union. And yet, in the
words of De Grauwe (2000, p93), “[t]he economic cost-benefit analysis remains useful (...) because
it gives an idea of the price some countries will have to pay to achieve these political objectives.”
If the underlying economics points to deficiencies, frictions on the political level are likely.
For EMU, the preceding analysis suggests that not all countries currently in the EU will find it
beneficial joining a monetary union, although the number of countries forming an optimal currency
area is probably larger than previously thought. This number will increase further as integration
deepens over time, so countries should be allowed to wait before joining the union as this will
positively affect their cost-benefit calculation. In the long run, most, if not all EU countries
will find this currency union to their advantage. But joining does incur risks even for countries
expected to gain on average, since they will find it more difficult to adjust to large shocks without an
independent monetary policy. Being a true OCA would help in this respect. Finally, enlargement
of the EMU will pose some serious challenges. It is unclear if and to what extent the new applicants
will profit from joining, and current members will also find that their cost-benefit considerations
are being altered by enlargement.

18
References
[1] Alesina, A. and R. Barro (2002): “Currency Unions.” Quarterly Journal of Economics 117
No. 2 (May), 409-436.
[2] Alesina, A., R. Barro and S. Tenreyro (2002): “Optimal Currency Areas.” Harvard Institute
Research Working Paper No. 1958. NBER Macroeconomics Annual, forthcoming.
[3] Balassa, B. (1964): “The Purchasing Power Parity Doctrine: A Reappraisal.” Journal of
Political Economy 72, 584-596.
[4] Barro, R. and D. Gordon (1983): “Rules, Discretion and Reputation in a Model of Monetary
Policy.” Journal of Monetary Economics 12, 101-121.
[5] Bayoumi, T. and B. Eichengreen (1993): “Shocking aspects of European monetary integra-
tion.” In F. Torres and F. Giavazzi (eds.): Adjustment and growth in the European Monetary
Union. Cambridge: Cambridge University Press.
[6] Bruno, M. and J. Sachs (1985): Economics of Worldwide Stagflation. Oxford: Basil Blackwell.
[7] Calmfors, L. and J. Driffill (1988): “Bargaining Structure, Corporatism and Macroeconomic
Performance.” Economic Policy 6, 13-61.
[8] Corden, M. (1972): “Monetary Integration.” Essays in International Finance 93, Princeton.
[9] De Grauwe, P. (2000): Economics of Monetary Union. Fourth Edition. Oxford: Oxford Uni-
versity Press.
[10] De Grauwe, P. (2003): Economics of Monetary Union. Fifth Edition. Oxford: Oxford Univer-
sity Press.
[11] De Grauwe, P. and W. Vanhaverbeke (1993): “Is Europe an Optimum Currency Area? Evi-
dence from Regional Data.” In P. Masson and M. Taylor (eds.): Policy Issues in the Operation
of Currency Unions. Cambridge: Cambridge University Press.
[12] EC Commission (1990): “One Market, One Money.” European Economy 44.
[13] EC Commission (2001, 2002): “Price dispersion in the Internal Market.” May
2001. “Price differences for supermarket goods in Europe.” May 2002. Available at
http://europa.eu.int/comm/internal market/en/update/economicreform/.
[14] Eichengreen, B. (1990): “Is Europe an Optimum Currency Area?” CEPR Discussion Paper
478.
[15] Fidrmuc, J. and I. Korhonen (2001): “Similarity of Supply and Demand Shocks Between the
Euro Area and the Accession Countries.” Focus on Transition 2, 26-42.
[16] Frankel, J. and A. Rose (1998): “The Endogeneity of the Optimum Currency Area Criteria.”
Economic Journal, July, 1009-1025.
[17] Frankel, J. and A. Rose (2002): “An Estimate of the Effect of Currency Unions on Trade and
Growth.” Quartlery Journal of Economics 117 No. 2 (May), 437-466.
[18] Giersch, H. (1973): “On the Desirable Degree of Flexibility of Exchange Rates.”
Weltwirtschaftliches Archiv 109, 191-213.
[19] Kenen, P. (1969): “The Theory of Optimum Currency Areas: An Eclectic View.” In
R. Mundell and A. Swoboda (eds.): Monetary Problems of the International Economy.
Chicago: University of Chicago Press.
[20] Krugman, P. (1989): “Differences in Income Elasticities and Trends in Real Exchange Rates.”
European Economic Review 33, 1031-1047.
[21] Krugman, P. (1991): Geography and Trade. Cambridge, Massachusetts: MIT Press.

19
[22] Krugman, P. (1993): “Lessons of Massachusetts for EMU.” In F. Torres and F. Giavazzi (eds.):
Adjustment and growth in the European Monetary Union. Cambridge: Cambridge University
Press.

[23] Kydland, E. and E. Prescott (1977): “Rules rather than Discretion: The Inconsistency of
Optimal Plans.” Journal of Political Economy 85.

[24] McKinnon, R. (1963): “Optimum Currency Areas.” American Economic Review 53, 717-725.

[25] Mendizábal, H. (2002): “Monetary Union and the Transaction Cost Savings of a Single Cur-
rency.” Review of International Economics 10(2), 263-277.

[26] Mundell, R. (1961): “A Theory of Optimum Currency Areas.” American Economic Review
51 (September), 657-665.
[27] Mundell, R. (1973): “Uncommon Arguments for Common Currencies.” In H. Johnson and
A. Swoboda (eds.): The Economics of Common Currencies - Proceedings of the Madrid Con-
ference on Optimum Currency Areas. London: George Allen & Unwin.

[28] OECD (1999): EMU: Facts, Challenges and Policies. Paris.

[29] Poole, W. (1970): “Optimal Choice of Monetary Policy Instruments in a Simple Stochastic
Macro Model.” Quarterly Journal of Economics 85, 197-216.

[30] Romer, P. (1986): “Increasing Returns and Long Run Growth.” Journal of Political Economy
94 (October), 1002-1037.

[31] Rose, A. (2000): “One Money, One Market: Estimating the Effect of Common Currencies on
Trade.” Economic Policy 15, No. 30 (April), 7-46.

[32] Rose, A. (2002): “The Effect of Common Currencies on International Trade: A Meta-
Analysis.” Unpublished manuscript, UC Berkeley, April 2002.

20

Potrebbero piacerti anche