Exchange-Rate Regimes on the Road to EMU: Lessons from Greece’s Experience
28/02/2003 - Speeches
Exchange-Rate Regimes on the Road to EMU:
Lessons from Greece’s Experience*
Nicholas C. Garganas
Bank of Greece
*Luncheon Address, Seminar on “Monetary Strategies for Accession
Countries”, Hungarian Academy of Sciences, Budapest, February 28, 2003.
Ladies and Gentlemen,
I would like to thank Magyar National Bank, the Institute of World
Economics of the Hungarian Academy of Sciences, and the Center of European Integration
Studies of the University of Bonn for inviting me to speak to you today.
In contemplating the subject of this seminar, one is struck by how
dramatically things have changed in the course of a decade. In 1992 and 1993, when the
Exchange Rate Mechanism - - or ERM - - underwent a series of speculative attacks, the
prospects of a European monetary union were viewed, by many observers, with considerable
skepticism. This skepticism was not without foundation. After all, hadn’t there been
previous false starts on the road to EMU? Hadn’t the Werner Report prescribed a European
monetary union by the end of 1970s? Yet, here we are today, having fulfilled the dreams of
Pierre Werner and his colleagues and having celebrated the fourth birthday of EMU.
Today, I would like to address some lessons from the experience of
Greece, the newest member of the euro area, in its quest for EMU membership. I will
confront what appears to be a dilemma. As we have heard this morning, much of the
economics profession appears to have been converted to the “hypothesis of the vanishing
middle”; for economies well integrated into world capital markets, there is little, if
any, middle ground between floating exchange rates and monetary unification. Effectively,
this hypothesis rules out intermediate regimes. Yet, a requirement for entering the euro
area is participation in ERM II, which is, after all, an intermediate regime. How can this
dilemma be resolved?
The Retreat from Intermediate Regimes
Before I discuss Greece’s experience, let me address the reasons
underlying the retreat from intermediate regimes. What has caused this retreat? First, an
explosive increase in capital flows during the 1990s has made the operation of
intermediate regimes problematic. As has already been discussed at this seminar by Jorge
Braga de Macedo and Helmut Reisen, according to the thesis of the Impossible Trinity,
developed in the 1980s, under a system of pegged exchange rates and free capital mobility,
it is not possible to pursue an independent monetary policy on a sustained basis.
Eventually, current-account disequilibria and changes in reserves will provoke an attack
on the exchange rate.
The enormous increase in capital flows has been accompanied by abrupt
reversals of these flows. Whereas the logic of the thesis of the Impossible Trinity
suggests that exchange-rate attacks typically originate in response to current-account
disequilibria and build up gradually, in fact, recent speculative attacks have often
originated in the capital account, have been sudden and difficult to predict, and have
included the currencies of countries without substantial current-account imbalances.
Capital-flow reversals have involved a progression of speculative attacks, mostly against
pegged-exchange-rate arrangements. These reversals of capital inflows and resulting
exchange-rate devaluations or depreciations have often been accompanied by sharp
contractions in economic activity and have, at times, entailed “twin crises” - -
crises in both the foreign-exchange market and the banking system.
Finally, there has been a tendency of instability in foreign-exchange
markets to be transmitted from one pegged-exchange-rate regime to others in a process that
has come to be known as “contagion”. The victims of contagion have seemingly included
innocent bystanders - - economies with sound fundamentals whose currencies might not have
been attacked had they adopted one of the corner solutions.
This triad of factors - - (1) the difficulty of conducting an
independent monetary policy when the exchange rate is pegged and the capital account open,
(2) sudden capital-flow reversals, and (3) contagion - - have profoundly affected the way
we think about exchange-rate regimes.
The Experience of Greece: Pre-ERM
The foregoing factors significantly affected the Greek economy in the
1990s. Through 1994, the performance of the Greek economy was pretty dismal. Growth was
almost flat, and inflation and the fiscal deficit as a percentage of GDP, were in the
double-digit levels throughout the period. Other EU countries were moving forward in their
quests to become members of EMU while Greece was falling farther and farther behind.
Clearly, a regime change was called for.
It came in 1995. The signing of the Maastricht Treaty in 1992 and the
government’s publicly-stated objective of joining the euro area provided powerful
incentives for mobilising broad public support for policy adjustment. Among the policy
measures undertaken were the following:
- Fiscal policy was progressively tightened. The fiscal deficit, as a percentage of GDP,
fell from over 10 per cent in 1995 to around to 4 per cent in 1997.
- Financial deregulation, which had begun in the late 1980s, was completed and in 1994 the
capital account was opened. The deregulated financial system facilitated the use of
indirect instruments of monetary policy so that small, frequent changes in the instruments
became feasible, enabling rapid policy responses.
- The Greek Parliament approved independence of the Bank of Greece and provided the Bank
with a mandate to achieve price stability.
For its part, beginning in 1995, the Bank of Greece adopted what became
known as a “hard drachma policy”, under which the exchange rate was used as a nominal
anchor. For the first time, the Bank announced a specific exchange-rate target. Underlying
this policy, both in Greece and elsewhere during the 1990s, was the belief that the
adoption of a visible exchange-rate anchor could enhance the credibility of the
dissinflationary effort because (1) the traded-goods component of the price level could be
stabilised and (2) wage-settling and price-setting behaviour were restrained.
I will not go into details of the operation of the hard-drachma policy.
Suffice it to say that, during the years 1995-97, real interest rates at the short-end
were kept in the vicinity of 5 per cent. Fairly specific targets for the exchange rate
were announced in each of the three years and were achieved. Importantly, inflation fell
from about 11 per cent in 1994 to under 5 per cent at the end of 1997 while, in the first
three years of the policy, real growth almost tripled compared with the rate of 1991-94.
Yet, as is typically the case with all nominal-anchor exchange-rate
pegs, this regime produced difficulties. As predicted under the thesis of the Impossible
Trinity, the ability to conduct an independent monetary policy under an exchange-rate peg
and open capital account became increasingly demanding. The wide interest-rate
differentials in favour of drachma-denominated financial instruments led to a
capital-inflows problem. The Bank of Greece responded by sterilising these inflows,
limiting the appreciation of the nominal exchange rate, reducing the impact of the inflows
on the monetary base, and buying time for other policies to adjust. Still, the
sterilisation entailed quasi-fiscal costs and, by preventing domestic interest rates from
falling, tended to maintain the yield differential that had given rise to the inflows.
Additionally, the Greek economy experienced a fundamental problem associated with all
exchange-rate nominal-anchor pegs during the move to lower inflation; the real
exchange-rate appreciated significantly contributing, along with strong domestic demand,
to a widening current-account deficit.
This circumstance brings me to the second and third of the triad of
factors - - sudden capital-flow reversals and contagion. The widening current-account
deficit, combined with rapid wage growth, fed market expectations that the drachma was
overvalued and provided the basis for contagion from Asia, which commenced with the
devaluation of the Thai baht in July 1997. The sharp rise in interest rates required to
support the drachma increasingly undermined growth and fiscal targets. A further regime
shift was needed.
That regime shift was provided by the ERM. Effective March 16, 1998,
the drachma joined the ERM at a central rate that implied a 12.3 per cent devaluation
against the ECU. As I will explain, entry into the ERM allowed Greece to orient its
policies to stability, fostering convergence. Because participation in the ERM, without
severe tensions, plays a role in the convergence criteria for joining the euro area, it
acts as testing phase for the central rate as well as for the sustainability of
convergence in general.
Lessons from the ERM
Before I discuss the drachma’s experience in the ERM, let me posit a
question: What kind of an intermediate exchange-rate regime do we have in mind when we
refer to the ERM? In my view, the present ERM system is the result of an evolutionary
process. Like all Darwinian evolutions, several distinct versions of the species can be
distinguished. To provide analytic focus, I think it useful to distinguish among the
ERM Mark 1. This species lasted from the inception of the EMS, in 1979,
to 1987. Compared with the 2 per cent bands under the Bretton-Woods system, the ERM bands
of fluctuation, at 4.5 per cent, were fairly wide and the system was supported by capital
controls. For some currencies, the bands were even wider, at 12 per cent. Small
realignments occurred frequently. Between March 1979, when the system started, and January
1987, realignments occurred on eleven occasions. The foregoing features of the system
allowed France, for example, to devalue the franc by shifting the upper and lower limits
of its band without affecting the market exchange rate, helping to forestall the
possibility of the large profit opportunities that give rise to speculative attacks.
ERM Mark 2. What has been dubbed the “new EMS” began to take shape
in 1987. Increasingly, the deutsche mark was used as a nominal anchor, with the
Bundesbank’s reputation as a stalwart inflation fighter supporting the monetary
authorities of some participating countries in their efforts to attain anti-inflation
credibility. Except for an implicit devaluation of the Italian lira in January 1990, when
that currency moved from the wide band to the narrow band, realignments ceased to be a
characteristic of the system. In 1990, the remaining capital controls were eliminated by
participating countries. German reunification required a tight monetary policy to
counterbalance the large fiscal deficits generated by reunification. With other ERM
countries in recession and requiring a loosening of monetary policy, the ERM was
confronted with a classic (n-1) problem.
ERM Mark 3. Beginning in September 1992, many of the currencies
participating in the ERM were subjected to attacks, eventuating in a series of
devaluations and the suspension of the pound sterling and the Italian lira from the
system. These attacks were sudden and massive. Thus, they were different from the currency
realignments of the 1980s, which were mainly the result of pressures that had build up
gradually in response to current-account disequilibria. With the lifting of capital
controls, the attacks in 1992-93 arose on the capital account and sometimes infected the
currencies of countries with seemingly-sound fundamentals, such as France. Several new
terms made their way into the lexicon of economists: (1) capital-account-driven crises,
(2) sudden-stops, and (3) contagion.
ERM Mark 4. In a last-ditch effort to rescue the system, in August 1993
the ERM bands were widened to ± 15 per cent. As things turned out, this move helped
salvage the system. It provided necessary breathing space for nominal convergence to
ERM Mark 5, or formally, ERM II. Under the present arrangement,
exchange-rate stability is explicitly subordinated to the primary objective of price
stability for all participating currencies and obligations under the system are
deliberately more asymmetric than under the previous ERM. The notion of asymmetry is
particularly important in underlining the principle that it is the country whose currency
comes under pressure that has to undertake the necessary policy adjustments.
Let me now return to the case of the drachma and the ERM. By the time
that the drachma entered the ERM, in March 1998, the participating countries had
demonstrated their determination to form a monetary union by attaining considerable
nominal convergence, as specified in the Maastricht criteria. This convergence took place
under a backdrop of market-based economies that could compete effectively in an open
economic and financial system. In these conditions, the system built up a considerable
amount of trust. As Otmar Issing has pointed out in several recent papers, the trust
evoked by governments, including in their ability to deliver credible, non-inflationary
policies, is a prerequisite for the existence of a stable currency, both internally and
When the drachma entered the ERM, it became a beneficiary of the
credibility established in the system over the previous years. It also benefited from the
market’s knowledge of the availability of the system’s mutual support facilities, such
as the Very Short-Term Financing Facility. Yet, Greece’s participation was not a free
ticket. The other participants in the system did not wish to endanger the credibility that
had taken so long to achieve. They rightly asked for an entry fee. This fee included the
- The new central rate - - which as I have noted, involved a 12.3 per cent devaluation - -
had been agreed by all members.
- The devaluation of the drachma was both backward looking and foreward looking. The
magnitude of the devaluation took account of both past inflation differentials between
Greece and other EU countries and prospective differentials in the period leading up to
Greece’s expected entry into the euro area. Thus, the new central rate was meant to be
- A package of supportive fiscal and structural measure was announced. Efforts to
restructure public enterprises were stepped up. The aim of the measures was to ensure the
sustainability of the drachma’s new central rate. In other words, the ERM was meant to
be a testing phase for EMU participation, not a free pass into the euro area.
Unlike many other devaluations of the mid-1990s and late-1990s, the
drachma’s devaluation was not followed by further rounds of speculative attacks, nor by
a financial crisis. Unlike other devaluations, it was not followed by contraction in
economic activity, but by an acceleration. Moreover, and again in contrast with the other
devaluations of this period, the impact of the drachma’s devaluation on inflation was
strictly contained. What accounts for the drachma’s successful exit from one central
rate to another? In my view, the key ingredients of the successful devaluation were the
- Unlike other currencies that were devalued during the mid-1990s and late 1990s, the
drachma exited a unilateral peg and entered a systems’ arrangement, benefiting from the
credibility of the ERM.
- Fiscal tightening continued following the devaluation. The fiscal deficit, as a per cent
of GDP, fell to about 1 per cent in 1999, from 4 per cent in 1997. Labour-market policy
gradually adjusted to the needs for fiscal discipline and for enhancing international
- Prudential regulation and supervision of the banking system had been strictly enforced
and there was no net foreign exposure of the banking system. Thus, there was no
currency-mismatching problem - - i.e., large, uncovered foreign-currency positions, which,
under conditions of currency devaluations, raise the debt burden of domestic
foreign-currency borrowers, resulting in bankruptcies and financial crises.
In addition to placing the Bank of Greece’s disinflation strategy
within a new institutional framework that gave it added credibility, the ERM provided
another important advantage. Entry at the standard fluctuation bands of ± 15 per cent
gave the Bank ample room for manoeuvre. Thus, when capital inflows resumed following ERM
entry, the Bank allowed the exchange rate to appreciate relative to its central rate,
helping to maintain the tight monetary-policy stance and to contain the inflationary
impact of the devaluation. The exchange rate remained appreciated relative to its central
rate throughout the rest of 1998 and for all of 1999. In 1999, for example, the drachma
traded (on average) 7.7 per cent above its central rate while, for most of that year, the
three-month interbank rate stood about 700 basis points above the corresponding German
rate. As a result of the tightened and consistent policy mix, inflation reached a low of 2
per cent during the second half of 1999. Then, in order to limit the degree of
depreciation that would be required for the market rate to reach its central rate and the
resulting inflationary pressures, the central rate was revalued by 3.5 per cent in January
2000. The rest, as they say, is history; having fulfilled all the Maastricht criteria, on
January 1, 2001 Greece became the 12th member of the euro area.
What are the key lessons that emerge from Greece’s experience? Let me
highlight the following.
First, an intermediate exchange rate regime can be viable in today’s
world of high capital mobility provided that (1) the participants adhere to sound and
sustainable policies, including those that ensure the existence of a well-functioning
market economy, and, (2) the intermediate regime is used as a transitional arrangement on
the road to the corner solution of a monetary union. There is no contradiction between the
existence of ERM II and the hypothesis of the vanishing middle regime.
Second, ERM II provides the credibility of a systems’ arrangement,
built up in a Darwinian evolution, and flexibility through the wide fluctuation bands that
may be necessary to achieve nominal convergence. The credibility derived by participating
in the ERM II should not, however, be endangered by use of frequent adjustments of the
central parity. As the ERM experiences of the 1980s and 1990s vividly demonstrated,
frequent adjustments of central parities within a pegged regime are feasible only in the
presence of capital controls. In today’s world of high capital mobility, exchange-rate
changes are not instruments that policy-makers can use flexibly and costlessly. The more
often they are used, the less can be the credibility of a pegged exchange-rate system.
Third, the need of a consistent policy mix is crucial and has important
implications in terms of how we need to view policy analysis. An implication of both the
Mundell-Fleming model and Mundell’s famous assignment problem is that monetary policy
and fiscal policy constitute two, separate policy instruments. Yet, the simple accounting
fact that government expenditure has to be financed by either taxation, borrowing, and/or
money creation, implies that any analysis of monetary policy must make consistent
assumptions about fiscal policy. Monetary policy and fiscal policy, in other words, are
not independent policy instruments. They must work in tandem on the road to EMU, and in
Fourth, ultimately, a credible exchange-rate regime depends upon the
trust evoked by governments. A governance structure that enforces the rule of law and
sanctity of contracts and a political system that delivers credible, non-inflationary
policies are prerequisites for the existence of a stable exchange-rate regime.
Ladies and Gentlemen, thank you for your attention.
Eichengreen, Barry (2002), “Lessons of the Euro for the Rest of the
World”, first annual Marshall Plan Lecture, Austrian Marshall Plan Foundation, Vienna, 4
Issing, Otmar (2001), “The Euro and the ECB: Successful
Start-Challenges Ahead”, Address at Market News Seminar, London, May 3.
Werner, Pierre (1970), “Report to the Council and the Commission on
the Realisation by Stages of Economic and Monetary Union in the Community”, Bulletin of
the European Community, Supplement, Vol. 3 (Werner Report).