What Does Trade Openness Measure?
| DOI | http://doi.org/10.1111/obes.12275 |
| Date | 01 August 2019 |
| Published date | 01 August 2019 |
868
©2018 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd.
OXFORD BULLETIN OF ECONOMICSAND STATISTICS, 81, 4 (2019) 0305–9049
doi: 10.1111/obes.12275
What Does Trade Openness Measure?*
Eiji Fujii†,‡
†School of Economics, Kwansei Gakuin University, Nishinomiya, Japan (e-mail:
efujii@kwansei.ac.jp)
‡Center for Economic Studies & Ifo Institute, Munich, Germany
Abstract
An empirical measure of trade openness is defined as the ratio of total trade to GDP, and
represents a convenient variable routinely used for cross-country studies on a variety of
issues. However, the effects that the crude measure captures remain ambiguous, making it
difficult to interpret the empirical results. Drawing on several strands of the literature, this
study examines the informational content of the trade openness measure using intranational
and international data. We find that, even for fully integrated economies within a country,
trade openness is approximately half as variable as it is for segmented diverse countries
around the world. The information it conveys is better characterized as the extent of the
economic remoteness and idiosyncratic distribution of sectoral production. The cross-
country variation of trade openness derives more from the variability in GDP than trade.
I. Introduction
An empirical measure of trade openness, defined as the ratio of exports plus imports to GDP,
is a convenient variableroutinely used in a variety of international macroeconomic studies.
Use of the variableis per vasivein cross-country studies on a broad range of topics, including
growth analyses (Levine and Renelt, 1992), real exchange rate dynamics (Goldfajn and
Vald´ez, 1999), government size (Rodrik, 1998), output volatility (Fat´as and Mihov, 2001),
and sovereign debt defaults (Levi Yeyati and Panizza, 2011). Albeit an admittedly crude
measure of openness, the fact that this variable is widely used in the literature highlights
the importance of understanding the informational content that it conveys.
The prevalent use in the literature of trade openness seems partly a result of data
availability. Because data on total trade as a percentage of GDP are available for many
countries from standard international databases, researchers may find them useful in con-
ducting cross-country analyses. However, despite this pervasiveness, the effects that the
JEL Classification numbers: F40, F14.
*The author is grateful to an anonymous referee and Brian Bell (the editor) for helpful comments and suggestions.
He also thanks Menzie Chinn, Kenji Fujiwara and Mark Spiegel for comments and discussion.All remaining er rors
are solely the author’s.
What does trade openness measure? 869
trade openness variable captures remain ambiguous. Although it may well be an intended
proxy measure, the question is for what exactly is the variable a good proxy.
Conceptually, trade openness may be defined as the degree to which an economy main-
tains its outward orientation in trade. However, empirically, adopting this definition is
challenging because it requires detailed data for many countries on the extent of explicit
and implicit trade impediments that are product-, destination-, and origin-specific and
time-variant. Even if such data are available, an additional hurdle exists. Aggregating the
detailed data into an overall index that qualifies as a universal measure of trade openness
is difficult (Harrison, 1996). In fact, although various indicators are invented, Pritchett
(1996) showed that they are virtually uncorrelated with each other, casting doubt on their
consistency and reliability.1
Although the aforementioned issues may tempt researchers to seek a more practical
alternative, using trade-to-GDP ratios introduces significant ambiguity because this mea-
sure of openness depends on a variety of factors. In addition to the extent of the outward
orientation of trade policy, they include the sizes of the domestic and external markets, the
distances to consumers and from producers outside one’s own territory, and factor endow-
ments that induce specialization in production. As a composite of the numerous factors,
the empirical measure of trade openness, as convenient as it may be, inflicts difficulty on
the interpretation of its effects as documented in various contexts.
This discussion motivates us to examinethe infor mational content of the trade openness
variable by drawing on several strands of the literature. In doing so, we analyse both
intranational and international data. In particular, we use Japan’s prefecture-level data as
a case of reference in which assured free trade purges the ambiguity arising from failing
to measure various trade restrictions in international contexts. More specifically, the intra-
Japan sample consists of geographically clustered economies that are relatively similar in
size and fully open to trade with each other. The extent of variability the trade openness
still exhibits in these circumstances serves us as a useful benchmark for evaluating the
cross-country variations.
Furthermore, intra-Japan data on trade, sectoral output, product prices, and so on allow
us to evaluate several candidates for the effects captured by the trade openness variable,
which is difficult if not infeasibleto do in inter national contexts.Namely, they are economic
remoteness (Anderson and van Wincoop, 2003; Baldwin and Harrigan, 2011), market
integration (Engel and Rogers, 1996; Goldberg andVerboven,2005), uniqueness in sectoral
production distributions (Krugman, 1991; Kalemli-Ozcan, Sørensen and Yosha, 2003),
and inclination to pursue economies of scale effects (Alesina and Wacziarg, 1998; Alesina,
Spolaore and Wacziarg, 2000). Through the exercises, this study contributes to the literature
by shedding light on the informational content of the widely used albeit difficult-to-interpret
empirical variable.
To anticipate, our chief findings are as follows. Even for the fully integrated (and hence
perfectly open) regional economies in Japan, the measured trade openness is approxi-
mately half as variable as the trade openness of 171 segmented diverse countries around
the world. The intranational variation of trade openness embodies two primary informa-
1Fora discussion on the issue, see also Harrison and Rodr´
iguez-Clare (2010), Edwards(1993, 1998) and Rodriguez
and Rodrik (2000).
©2018 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd
To continue reading
Request your trial