ON THE CAUSES OF INSTABILITY IN EXPORT EARNINGS
| Published date | 01 August 1981 |
| Author | Peter Svedberg,Peter Brundell,Henrik Horn |
| Date | 01 August 1981 |
| DOI | http://doi.org/10.1111/j.1468-0084.1981.mp43003005.x |
ON THE CAUSES OF INSTABILITY IN
EXPORT EARNINGS
Peter Brundell, Henrik Horn and Peter Svedberg
I. INTRODUCTION
The most frequently cited reference on the causes of instability in
export earnings is the article, now ten years old, by Benton F. Massell
(1970). Massell was the first to provide a thorough theoretical dis-
cussion of most of the links between instability and those explanatory
variables he included, and his results have not been seriously challenged
by later studies.1 In this article, we have extended Massell's model
to include a few additional variables, aimed at picking up the influence
of trade policies on export earnings instability. Further, we provide
a theoretical rationale for testing two of the variables included ad
hoc in previous studies, and we improve on some variable definitions
and measurements. We also extend the data basis considerably; we
have a complete set of observations of twelve independent variables
for 139 countries.
II. STRUCTURAL vs. POLICY-INDUCED CAUSES OF INSTABILITY
Massell set out to test 'the relationship between instability in the
value of exports and a set of variables that help characterize a country's
economic structure'. The term 'economic structure' seems to imply
characteristics of countries that are given except in the long run. We
are inclined to believe that export-earnings instability is also explained
by factors subject to policy manipulation, if not in the short, at least
in the medium term. Specifically, we think that a country's trade
*The authors are indebted to Lars E. 0. Svensson, Peter J. Lloyd and ke G. Blomqvist for
helpful comments on an earlier draft. The third author gratefully acknowledges financial
support from the Bank of Sweden Tercentenary Foundation and the Swedish Agency for
Research Cooperation with Developing Countries. Thanks are due to Anita Oxenham for
editorial assistance.
'In the studies prior to Massell (1970), most notably Coppock (1962), Massell (1964),
MacBean (1966) and Erb and Schiavo-Campo (1969), little theoretical reason was given for
testing the different variables, and few statistically significant results were obtained (see Maizels
(1968) and Ady (1969)). Later studies have contributed mainly to highlighting the importance
of using different instability indices (Leith (1970); Lawson (1974)) and, in some instances, they
have been based on more comprehensive data bases (Sheehey (1977); Landen (1978)).
301
To continue reading
Request your trial