Cointegration and the Monetary Exchange Rate Model Revisited*
| DOI | http://doi.org/10.1111/1468-0084.00024 |
| Author | Jan J. J. Groen |
| Date | 01 September 2002 |
| Published date | 01 September 2002 |
Cointegration and the Monetary Exchange
Rate Model Revisited*
Jan J. J. Groen
Monetary Assessment & Strategy Division, Bank of England, Threadneedle Street,
London EC2R 8AH, United Kingdom; E-mail: jan.groen@bankofengland.co.uk
I. Introduction
Efficient exchange rate pricing based on the monetary exchange rate
model implies that the logarithm of the exchange rate and the logarithms
of the relative money supply and relative real income of two countries
are cointegrated with one cointegrating vector. Cointegration tests in
Sarantis (1994) and Groen (2000) for individual bilateral exchange rates
separately, however, do not provide evidence for cointegration based on
the monetary model. On the other hand, when we use for a large
number of countries cross-section regressions or test for cointegration
based on a fixed individual effect multi-country panel data model we do
find evidence pro the monetary exchange rate model, see also Husted
and MacDonald (1998) and Groen (2000). As such, the analysis of the
monetary exchange rate model across a multiple of countries simulta-
neously seems to provide more positive results relative to the monetary
model.
The panel cointegration frameworks described in Husted and MacDonald
(1998) and Groen (2000) are essentially pooled versions of the two-step
Engle and Granger (1987) procedure, and they only allow for a limited
*This paper has benefited from comments by an anonymous referee and seminar participants at the
CPB Netherlands Bureau for Economic Policy Analysis and the European Meeting of the Econo-
metric Society in Lausanne. The research for this paper was conducted while the author was affiliated
with the Tinbergen Institute at the Erasmus University Rotterdam. The views expressed in this paper
are those of the author and do not necessarily reflect the views of the Monetary Policy Committee or
the Bank of England.
OXFORD BULLETIN OF ECONOMICS AND STATISTICS, 64, 4 (2002) 0305-9049
361
Blackwell Publishers Ltd, 2002. Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350
Main Street, Malden, MA 02148, USA.
amount of cross-country heterogeneity. From the power analysis in
Groen (2000) it becomes clear that significant cross-country differences
in short-run dynamics can impede the finding of cointegration within the
panel framework which is used in that paper. This effect can especially be
severe within panels with a small number of countries as the scope for
‘averaging out’ the cross-country differences is limited. Mark and Sul
(2001) tried to deal with heterogeneous short-run dynamics within the fixed
individual effect panel framework through averaging the results of
individual residual ADF unit root tests into one statistic. An alternative
panel cointegration framework based on a panel of the individual vector
error correction [VEC] models across the constituting countries was
devised by Groen and Kleibergen (2002). Within this framework we can
test for a common number of cointegrating vectors per country where the
error correction coefficients are assumed to be heterogeneous, and given
this we can test whether or not the long-run parameters are homogeneous.
As such this framework should be able to deal with the aforementioned
cross-country differences in short-run dynamics. Also, in contrast to the
panel framework in Husted and MacDonald (1998), Groen (2000) and
Mark and Sul (2001) the framework from Groen and Kleibergen (2002)
enables us to test the parameter restrictions of the monetary exchange rate
model, i.e. a relative money elasticity of 1 and a negative relative real
income elasticity.
Together with the aforementioned panel Engle-Granger procedure and the
purely time series-based Johansen (1991) cointegrated VEC framework, the
panel cointegration framework from Groen and Kleibergen (2002) is applied
on two four-country panels of VEC models in order to test for the validity of
the monetary exchange rate model. We utilize quarterly time series data for the
1973–1994 period on the log exchange rate, log relative money supply and log
relative real income with respect to both the United Kingdom [U.K.] and the
United States [U.S.] for Germany, Japan, Switzerland and either the U.S. or
the U.K.
The remainder of this paper satisfies the following outline. Section II
summarizes the cointegration restrictions implied by the monetary
exchange rate model. Section III discusses the three tests utilized in this
paper to test for cointegration in a multi-country setting: the Johansen
(1991) approach for the individual countries separately, the panel Engle-
Granger approach from Groen (2000) and the panel VEC model framework
from Groen and Kleibergen (2002). This section also provides a Monte
Carlo comparison between these three methods. Cointegration test results
within our two four-country panels based on the three above mentioned
methods are reported in Section IV. Concluding remarks can be found in
Section V.
362 Bulletin
Blackwell Publishers 2002
To continue reading
Request your trial