NBER WORKING PAPER SERIES
THE MODERN HISTORY OF EXCHANGE RATE ARRANGEMENTS:
A REINTERPRETATION
Carmen M. Reinhart
Kenneth S. Rogoff
Working Paper 8963
http://www.nber.org/papers/w8963
NATIONAL BUREAU OF ECONOMIC RESEARCH
1050 Massachusetts Avenue
Cambridge, MA 02138
June 2002
The authors wish to thank Vincent Reinhart and Miguel Savastano and participants at Harvard’s Canada-US
Economic and Monetary Integration Conference for useful comments and suggestions and Kenichiro
Kashiwase, Daouda Sembene, and Ioannis Tokatlidis for excellent research assistance. The views presented
in this paper are those of the authors and are not necessarily those of the International Monetary Fund. This
work was largely undertaken while the authors were professors at the University of Maryland and Harvard
University, respectively. The views expressed herein are those of the authors and not necessarily those of
the National Bureau of Economic Research or the International Monetary Fund.
© 2002 by Carmen M. Reinhart and Kenneth S. Rogoff. All rights reserved. Short sections of text, not to
exceed two paragraphs, may be quoted without explicit permission provided that full credit, including ©
notice, is given to the source.
The Modern History of Exchange Rate Arrangements: A Reinterpretation
Carmen M. Reinhart and Kenneth S. Rogoff
NBER Working Paper No. 8963
June 2002
JEL No. F30, F31, F41
ABSTRACT
We develop a novel system of re-classifying historical exchange rate regimes. One difference
between our study and previous classification efforts is that we employ an extensive data base on
market-determined parallel exchange rates. Our “natural” classification algorithm leads to a stark
reassessment of the post-war history of exchange rate arrangements. When the official categorization is
a form of peg, roughly half the time our classification reveals the true underlying monetary regime to be
something radically different, often a variant of a float. Conversely, when official classification is
floating, our scheme routinely suggests that the reality was a form of de facto peg. Our new classification
scheme points to a complete rethinking of economic performance under alternative exchange rate
regimes. Indeed, the breakup of Bretton Woods had a far less dramatic impact on most exchange rate
regimes than is popularly believed. Also, contrary to an influential empirical literature, our evidence
suggests that exchange rate arraignments may be quite important for growth, trade and inflation. Our
newly compiled monthly data set on market-determined exchange rates goes back to 1946 for 153
countries.
Carmen M. Reinhart Kenneth S. Rogoff
Research Department Research Department
Room 10-700H International Monetary Fund
International Monetary Fund 700 19th Street, NW
700 19th Street, NW Washington, DC 20431
Washington, DC 20431 and NBER
and NBER krogoff@imf.org
creinhart@imf.org
- 1 -
I. Introduction
Contemporary thinking on the costs and benefits of exchange rate flexibility is
heavily influenced by the large number of studies on the empirical differences in growth,
trade, inflation, business cycles, and commodity price behavior across different exchange rate
arrangements. These include the influential work of Baxter and Stockman (1989) and Flood
and Rose (1995), both of whom failed to detect any significant difference across fixed and
floating rates, except for the volatility of real exchange rates that Mussa (1986) had already
noted. Nearly all such studies have relied on the official or “standard” classification
published in the IMF’s Annual Report on Exchange Rate Arrangements and Exchange
Restrictions which, until recently, asked member states to self-declare their arrangement as
belonging to one of four categories (fixed, limited flexibility, managed floating, and
independently floating). Yet, a closer reading of the experience suggests that these official
classifications often fail to describe actual country practice, implying that the gap between de
facto and de jure can be vast. A few previous studies have attempted to either extend the
four-way official classification into a more informative meaningful taxonomy (see Gosh et.
al., 1997), or have relied on purely statistical methods to regroup country practices (see,
Levy-Yeyati and Sturzenegger, 2002).
1
The Fund, recognizing the limitations of its former
strategy, significantly revised and upgraded the standard official approach toward classifying
exchange rate arrangements in 1997, though it did not re-evaluate its historical classification
after the fact. Notably, all previous approaches to exchange rate regime classification,
1
See also Masson (2000) and Glick (2001).
- 2 -
whether or not they accept the country’s declared regime, have been based almost solely on
official exchange rate series.
One problem with the pre-1997 official classification that has received substantial
attention in the recent literature is the frequency of episodes where the regime is classified as
floating (independently or managed) when, in effect, the country had a de facto peg or
crawling peg.
2
However, a less well known but equally serious problem, shared even by the
small number of earlier efforts to reclassify regimes, is that the officially-reported exchange
rate itself is often profoundly misleading. Over the course of post-World War II history,
virtually every country has relied, at one time or another, on capital controls and/or multiple
exchange rate systems.
3
By failing to look at market-determined exchange rates, one often
gets a false picture of the underlying monetary policy and the ability of the economy to adjust
imbalances. When there are dual or parallel markets, a regime that is officially labeled a peg
might easily turn out to be a de facto float or a crawling band. As an illustration suppose, for
example, that the parallel market rate undergoes sustained and significant depreciation but
the official rate remains fixed. (Over the broad sweep of post-war history, we find this a
common occurrence.) Then the underlying monetary policy is inflationary, but the effects on
the official nominal exchange rate are effectively masked–at least in the short run--by an
ever-increasing tariff. By far the most common outcome in these circumstances is that the
2
See Calvo and Reinhart (2002), Levy-Yeyati and Sturzenegger (2002), and Wickham (2002).
3
Edwards (1989) stresses the inadequacy of official exchange rate data for the many Latin American economies
with multiple exchange rates.
- 3 -
official rate is ultimately devalued, thus validating what had already transpired in the free
market.
Thus when there are dual markets, it is not sufficient to just discard the official
exchange rate classification. One very often also has to discard the official rate, which can
be meaningless and far removed from the rate at which transactions take place. Another
difference from other studies that have tackled the regime classification issue is that our
approach relies on extensive historical chronologies that allow us to identify when dual or
multiple rates are in place or when parallel markets are active. In the vast majority of these
cases, our data set contains monthly data on dual/parallel market exchange rates from 1946
through 1998, and we use these data to assess whether, from a monetary perspective, the true
de facto regime corresponds to the stated de jure regime. Very often—roughly half the time
for official pegs over the post-1970 period—we find that dual/parallel rates have been used
as a form of back-door floating.
4
Lastly, we rely on a large battery of descriptive statistics to
classify exchange rate arrangements into as many as 15 categories. We argue that both pieces
of information—historical chronologies and data on market-determined exchange rates--are
essential to any meaningful classification algorithm.
We note also that the IMF’s monthly data on exchange rates—which only includes
the official rate—currently begin in 1957, whereas ours begin in 1946. Our analysis is
extremely comprehensive, covering 153 countries. Since exchange rate regimes are often
4
The IMF data for official exchange rates begins on January 1957 for most countries. We also compiled data
on official exchange rates from 1946 through 1957; see appendix.
