://www.lafollette.wisc.edu/publications/workingpapers

A Faith-based Initiative: Do We Really Know
that a Flexible Exchange Rate Regime
Facilitates Current Account Adjustment?
Menzie D. Chinn
Professor, La Follette School of Public Affairs and Department of Economics
at the University of Wisconsin-Madison, and the National Bureau of Economic Research

mchinn@lafollette.wisc.edu

Shang-Jin Wei
Columbia University and the National Bureau of Economic Research














A Faith-based Initiative: Do We Really Know that a Flexible Exchange
Rate Regime Facilitates Current Account Adjustment?


Menzie D. Chinn*

University of Wisconsin, Madison, and NBER


Shang-Jin Wei**

Columbia University and NBER



February 24, 2008


Abstract:

The assertion that a flexible exchange rate regime would facilitate current account
adjustment is often repeated in policy circles. In this paper, we show that the assertion is
not supported in the data.


Keywords: floating exchange rate, fixed exchange rate, current account imbalances

JEL Classification Nos.: F3

Acknowledgments: We thank Graciela Kaminsky for very helpful comments.

* Robert M. La Follette School of Public Affairs; and Department of Economics, University of Wisconsin,
1180 Observatory Drive, Madison, WI 53706-1393. Email: mchinn [at] lafollette.wisc.edu

** Graduate School of Business, Columbia University, 619 Uris Hall, 3022 Broadway, New York, NY
10027, USA. Tel: 212/854-9139, Email: shangjin.wei [at] columbia.edu
1
The third part of the strategy [to address global current account imbalances] was to
increase exchange rate flexibility in order to facilitate the adjustment of the current
account over time.
John Taylor, Professor of Economics at Stanford University and former Under Secretary of
Treasury for International Affairs, speech at the IMF conference on April 21, 2006.

We also agreed that an orderly unwinding of global imbalances, while sustaining global
growth, is a shared responsibility involving ... greater exchange rate flexibility in a
number of surplus countries ...
G20 Communiqué, Meeting of Finance Ministers and Central Bank Governors, Cape Town, South
Africa, November 17-18, 2007.

From a global perspective, exchange rate flexibility ... would also help contribute to an
orderly process for resolving global current account imbalances.

IMF Staff, Peoples Republic of China: Staff Report for the 2006 Article IV Consultation.

1. Introduction

The assertion that a more flexible exchange rate regime would promote current account
adjustment has been repeated so often that policy makers and economic analysts take it as
self-evident that this must be true. There is in fact no systematic evidence supporting this
supposition. Until one finds persuasive empirical evidence, the policy recommendation
for a more flexible exchange rate regime in pursuit of current account adjustment is a
faith-based initiative based on something widely assumed to be true, actively peddled to
countries as a truth, but with little solid empirical support.

In this paper, we seek to address this deficiency by systematically investigating any
relationship in the data between exchange rate regimes and speed of current account
adjustment. In addition to using officially announced exchange rate regimes, we appeal to
de facto regimes in place. To buttress our case, we utilize two well-established and
familiar approaches to classifying a countrys exchange rate regime on a de facto basis,
2
by Levy-Yeyati and Sturzenegger (2003a,b), and by Reinhart and Rogoff (2004),
respectively.

To anticipate the results, after experimenting with a large number of statistical
specifications, we find no support in the data for the notion that countries on a de facto
flexible exchange rate regime exhibit a faster convergence of their current account (as a
percentage of their GDP) to the long run equilibrium, regardless of which de facto
exchange rate regime classification scheme we employ. This is true when we control for
trade and financial openness; and this is true when we separate large and small countries.

To be sure, the current account balance does have a tendency to revert to its long run
steady state; it does not wander off or stay away from the long run equilibrium forever.
This is clearly reflected in our empirical work. However, the speed of adjustment to the
steady state is not systematically related to the degree of flexibility of a countrys
nominal exchange rate regime.

Should we be surprised by this finding? Perhaps not. The current account responds to real
exchange rate, not the nominal exchange rate. If the real exchange rate adjustment does
not depend very much on the nominal exchange rate regime, then the current account
adjustment would not depend very much on nominal exchange rate regime either. That is
why another key part of our analysis examines whether the nature of a countrys nominal
exchange rate regime significantly affects the pace of real exchange rate adjustment.

We conclude that the answer is no: the real exchange rate adjustment is not
systematically related to how flexible a countrys nominal exchange rate regime is. Again,
this is true regardless of which de facto exchange rate regime classification we use. If
anything, there is slight, but not very robust evidence that less flexible nominal exchange
rate regimes sometimes exhibit faster real exchange rate adjustment.

3
The rest of the paper is organized as follows. Section 2 lays out the empirical
methodology, data, and benchmark results. Section 3 conducts a series of extensions and
robustness checks. Finally, Section 4 concludes.

2. An empirical examination of exchange rate flexibility and current account
adjustment

2.1 Empirical Methodology

We estimate the rate at which current account balances (expressed as a share of GDP)
revert to their mean values, using variations on this basic autoregression:

it
it
it
v
ca
ca
+
+
=
1
1
0








(1)

Where ca is the current account to GDP ratio for country i.
1
One can determine how the
autoregressive coefficient varies with the exchange rate regime in a variety of ways. The
simplest would be to order the exchange rate regimes by degree of flexibility, and then
interact with a single variable. An alternative would be to stratify the sample by exchange
rate regime and run separate regressions per regime. The third (nearly equivalent)
approach would be use dummy variables for each regime but in a single regression.

The first approach imposes the condition that there is a monotonic and linear relationship
between flexibility and current account reversion. The second approach imposes the
fewest assumptions, but might yield imprecise estimates due to substantially decreased
number of observations for each regression. The third approach will yield the same point
estimates as obtained in the second approach but will yield different estimated standard
errors. The validity of this approach for making inference is dependent upon the
condition that the error term is distributed in a similar fashion across exchange rate
regimes.

1
Note that this is a fairly simple manner with which to characterize the degree of persistence in the current
account. While it is possible there are higher order autoregressive terms, only an AR(1) seems necessary
for the annual data used in this study.
4

We will focus on the second and third approaches, although we will discuss the results
from the first briefly. The second approach relies upon estimating equation (1) for each
category of exchange rate regime. The third approach involves estimating (2):

it
it
it
it
it
it
v
regime
ca
regime
ca
ca
+
×
+
+
+
=
)
(
1
1
0
1
1
0

(2)

The variable regime is the de facto exchange rate regime.
2


In all instances, we would like to control for other more structural variables that might
also affect the rate of reversion. In the case of equation (2), we augment the equation with
level and interaction effects.

it
it
it
it
it
it
v
controls
regime
ca
regime
ca
ca
+
+
×
+
+
+
=
)
(
1
1
0
1
1
0
(3)

Where controls includes different measures of economic openness, including trade and
financial openness, described in greater depth below.

2.2. Data
The current account and trade openness data are from the World Banks World
Development Indicators. The trade openness variable is the standard measure (the sum of
imports and exports divided by GDP). Over 170 countries are included, over the 1971-
2005 period. The sample encompasses both developed and developing countries (as
classified by the IMF).

The de facto exchange rate regime variables come from two sources: the Levy-Yeyati
and Sturzenegger (2003a,b) and the Reinhart and Rogoff (2004) measures.


2
We have used the de jure index based upon the IMFs Annual Report on Exchange Arrangements and
Exchange Restrictions (AREAER) instead of the de facto measures, to little effect. The results indicate no
systematic relationship. Note that the de jure indices extend only up to 1999.
5
The Levy-Yeyati and Sturzenegger index ranges from 1 to 5, with 1 indicating
inconclusive determination, 2 free float, 3 dirty float, 4 dirty float/crawling peg, and 5
fix. In this study, we drop 1s, and subtract 2 off the index, so that the revised index
ranges from 0 to 3 (hereafter the LYS index).

The Rei