Geo-Economic Competition and Trade Bloc Formation: United States, German, and
Japanese Exports, 1968-1992
Author(s): John O'Loughlin and Luc Anselin
Source: Economic Geography , Apr., 1996, Vol. 72, No. 2 (Apr., 1996), pp. 131-160
Published by: Taylor & Francis, Ltd.
Stable URL: https://www.jstor.org/stable/144263
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Taylor & Francis, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to
Economic Geography
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
Geo-Economic Competition and Trade Bloc Formation:
United States, German, and Japanese Exports,
1968-1992*
John O'Loughlin
Institute of Behavioral Science and Department of Geography,
University of Colorado, Boulder, CO 80309-0487
Luc Anselin
Regional Research Institute, West Virginia University, Morgantown, WV 26506
Abstract: In the post-cold war world, geo-economic competition is thought to be
replacing geopolitical competition as the focus of great power relations. The cold
war years corresponded to the period of U.S. hegemony in world trade and
relations in the Western bloc. With the shrinking of the power gap between the
United States and the other two great trading states, Japan and West Germany,
as well as increased competition for trade shares, a division of the world economy
into trade blocs has been anticipated. An examination of export shares for the
three great powers with 114 partners in the past quarter century, 1968 to 1992,
indicates there is not much evidence for the hypothesis of a world devolving into
trade blocs. While regional links have intensified somewhat between the United
States and its neighbors in the Americas and between West Germany and its
European Union partners, Japan is broadening and deepening its export linkages
with extraregional partners. Fears of the formation of blocs in the world trading
system are greatly exaggerated.
Key words: trade blocs, panregions, geo-economics, spatial analysis, geopolitical
order.
Two parallel developments have
changed the world as we knew it. The
post-World War II geopolitical order was
shattered with the collapse of the Soviet
*This research was supported, in part, by a
grant from the National Science Foundation to
John O'Loughlin (SES-9002699) and by National
Science Foundation grants to Luc
Anselin (SES-8810917 to the National Center
for Geographic Information and Analysis,
SES-8921385, and SBR-9410612). Thanks are
also due to Michael Shin of the Geography
Department of the University of Colorado for
his construction of the maps. A draft of this
paper was presented at a conference sponsored
by the IGU Commission on the World
Political Map and the United Nations' University
in Tokyo on "The Asia-Pacific and Global
Geopolitical Change," August 1993. Comments
from the conference participants are
appreciated.
Union in 1991, though it had been
tottering for about five years. The geoeconomic
order, built on a foundation of
United States hegemony, a liberal international
trade regime, and a Fordist mode
of production, is also challenged, not with
the finality of a military coup but through
the actions of multinational corporations
and the policies of governments. The
world of the GATT (General Agreement
on Tariffs and Trade), promoted and
dominated by the United States as a way
of linking the capitalist economies and
promoting international growth, is no
longer unquestioned by America's partners,
or even within the United States
itself.
In the past decade, a sharp reversal of
earlier trends toward tariff reductions has
occurred and many nontariff barriers have
been instituted to protect industries perceived
to be under threat from the free
131
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
ECONOMIC GEOGRAPHY
trade provisions of GATT (Meltzer 1991;
Grant 1993). One strategy to protect vital
trade interests is to delimit a trade zone
and attempt to exclude global competitors
from it. Whether the GATT world of
global commerce is gradually being replaced
by a tripartite division into trade
blocs, each dominated by the respective
superpowers, is the focus of this paper.
Economic geography, in the form of
regionalism in world trade, is said to be
reasserted as ideological, and nation-state
conflicts in the world's regions are mending
("Trade Winds Shift" 1993). While
many commentators have speculated on
this important development, empirical
examination of the evidence to date has
been sparse. We present evidence from
the past 25 years of trade developments
that indicate that the "end" of the GATT
is exaggerated because regional trade
blocs are still poorly developed.
Our method in this paper is exploratory
spatial data analysis (ESDA). We posit no
hypotheses on the development or lack of
development of trade blocs, since the
processes that would underlie any developments
are poorly understood. We determine
the accuracy of claims that the
global economic system is devolving into
geographically defined blocs. Our working
definition of the existence of a trade bloc
is statistical evidence that exports are
disproportionately sent to countries in a
specific region and that the pattern is
intensifying over time. As we review the
scattered literature on geo-economic developments
in the past 20 years, we
characterize it as poorly theorized and
unconvincing in its empirical details. We
prefer an inductive or data-driven approach
to the topic. By these means, we
gain insights into a question for which
theoretical and empirical insights are
currently underdeveloped. This approach,
of course, reflects the most common
geographic information system (GIS)
style, and it represents a phase in which
spatial patterns and structures are revealed,
hypotheses proposed, and models
suggested (Anselin and Getis 1992). In
ESDA, the emphasis is placed on structure
and association, an appropriate approach
for the analysis of trade patterns.
From Stable Geopolitics to
Mobile Geo-economics
The United States symbolizes all that
has changed in the past two decades.
From being an unabashed advocate of free
trade, Washington now sends a muted or
jumbled message, with various segments
of the state apparatus promoting different
agendas. "Managed trade" or "fair trade"
have become the slogans of the Clinton
administration as it pursues domestic and
international policies to try to reduce
America's chronic trade deficit and protect
vulnerable (textiles) or strategic (microchips)
industries from foreign imports
(Low 1993). International agencies, instituted
by the United States to manage the
integrated world economy, now decry the
retreat from an open global economy to
one in which individual state interests
take precedence over the fortunes of the
trade system (World Bank 1992). As trade
liberalization slows, the loudest complaints
have come from believers in
neoclassical trade theory, who hold that
all countries eventually benefit from
producing and exporting the products of
their comparative advantage (Crook 1990).
For Bhagwati (1991), the situation is
nothing short of a "world trading system
at risk."
How we arrived at a "new world order"
that seems increasingly disorderly is the
culmination of interacting economic and
political developments in the United
States and its relations to the other world
powers, as well as structural economic
forces that have shifted the locus of global
manufacturing and reduced the undisputed
U.S. lead in global production,
trade, and finance. The U.S. reaction to
this new world has been confused: it
seems uncertain whether the strategies
that sustained U.S. policy throughout the
cold war years should be abandoned,
modified, or maintained. What is undisputed
is that a new sphere has been
132
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
TRADE BLOC FORMATION
added to the traditional political-military
competition (Corbridge and Agnew 1991).
No longer can the United States be
perceived as one side of a unidimensional
geopolitical contest with a Soviet adversary;
its geopolitical allies are now its
economic competitors. The catch-word of
the 1990s summarizes the global shift:
"from geopolitics to geoeconomics" (Luttwak
1993). In a useful antidote to the
fixation on international competitiveness
and global market shares, however, Krugman
(1994) has blamed stagnant wage
levels and standards of living on domestic
causes, chiefly a failure to speed up
growth in domestic productivity and a
national spending spree that reduces
capital formation. Manufacturing losses to
imports account for only a small proportion
(0.07 percent) of national income; a
general demand shift to services accounts
for most deindustrialization in the United
States (Krugman and Lawrence 1994).
In the geo-economic world, trade is a
zero-sum contest. This perspective is
opposed to the economic view of Adam
Smith and David Ricardo, which holds
that gains from protected trade and
shifting investments are temporary and
that the world economy as a whole
prospers from free trade. The images,
models, language, and proposed policies
to examine and treat the competition from
both large established and new industrial
economies have been borrowed from the
geopolitical almanac (6 Tuathail 1993).
While Japan is clearly the object of most
attention in the United States, it has
recently been joined by the European
Community in a perceived (from the U.S.
perspective) triangular contest for world
leadership. From the perspective of the
United States, all has changed since the
hey-day of U.S. power in the 25 years
after World War II. Former enemies are
friends, while the motivations of allies are
questioned. As Taylor (1993) notes, Americans
have an inordinate respect for
rankings; evidence that the United States
is falling behind its competitors generates
further demands for action, though the
evidence for U.S. relative decline is quite
mixed (O'Loughlin 1993a). In the United
States, it is in the trade sphere that the
lines are clearest, the rhetoric is loudest,
and the clamor for action is strongest.
In the neoclassical view, trade is seen as
an engine of economic growth. Demands
for products overseas will generate more
jobs at home. In a time of deindustrialization,
the export of manufactured products
is seen as an essential part of global
competitiveness. For example, Duesterberg
(1993) notes that the boom in U.S.
exports since 1987, triggered by the 1985
Plaza agreement driving down the value
of the dollar to make U.S. exports more
competitive, has accounted for about 50
percent of U.S. growth and produced 2
million new jobs in manufacturing. World
trade grew faster than global economic
output between 1950 and 1990 (Summers
1991). The share of U.S. Gross Domestic
Product represented by exports was 7.5
percent in 1993, up from 5 percent just
five years earlier.
The close linkage of exports to economic
growth is questioned by critics of
neoclassical economics, such as the regulationist
school, and by empirical evidence
that booming exports do not always
reduce unemployment. (The Republic of
Ireland is a case in point.) Marxists have
traditionally noted the use of the state
apparatus to promote the interests of
domestic capitalists, and Drache and
Gertler (1991) believe that this advocacy
is strengthened in the new global policy
environment that pushes states to adopt
market-based and trade-centered policies.
Additionally, most capitalist states now try
to act as a buffer between their populations
and the negative effects of global
economic changes, whose impacts have
differential effects by industrial sector and
region. Instead of export-led growth, a
national industrial strategy that relies to a
large extent on domestic sources can be
an alternative to the dominant exportoriented
mode of development.
While belief in the merits of free trade
still generally holds sway in the major
world capitals, fears increase among free
trade supporters that basic principles,
133
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
ECONOMIC GEOGRAPHY
such as "rising trade benefiting all," are
no longer unquestioned. Whether directly
or inadvertently, GATT supporters believe
that the major exporters are targeting
regional export partners while working
actively to build free trade areas that
insulate exports from other competitors
and tie the partner countries into longterm
economic relationships. Trade blocs
are viewed as the first steps to deeper
economic linkages, following the example
of the European Community, and eventually
to political blocs (Lawrence 1991).
Bloc building is anathema to GATT
supporters, since it violates the spirit and
purpose of Article XXIV of the treaty
(Jackson 1993). Already, the European
Community is deepening its internal
relationships and building barriers to
imports (Japan is the perceived main
target of these provisions), the North
American Free Trade Agreement has
extended south to Mexico, with expected
expansion to Latin America as a whole,
and Japan is reputed to be building a yen
bloc in East and Southeast Asia (Maidment
1989). In the worst fears of the free
traders, these blocs will eventually
broaden to encompass all countries in
three economically competitive and politically
coherent blocs. The trade-diverting
effects of bloc building are potentially
enormous, and clear answers to the
important question of whether blocs are
developing are important for the future
appearance of both the geo-economic and
geopolitical worlds. Economic relations
will undoubtedly affect political relations.
Trade Theory and
Panregional Visions
The trade alarmists worry that the
1990s will become increasingly like the
1930s. In that decade, the world trade
system, which had been gradually liberalizing
and opening, retreated into protectionism.
Five blocs became identifiable
(British empire, French empire, United
States and Latin America, Germany and
Eastern Europe, and Japan's "East Asian
Co-prosperity Sphere"), and each tried to
distance itself from the rest of the world.
In a decade of economic crisis, with
increased competition for shrinking markets,
a division of the spoils seemed like a
reasonable tactic to guard market share
against extraregional exporters.
In the context of the 1930s global
recession, various geo-economic strategies
were suggested to cope with the crisis.
One of the most debated and famous of
these strategies was the German panregional
model. In the view of the German
school of Geopolitik, dominated by Karl
Haushofer, a stable equilibrium could be
produced and maintained by the division
of the globe into three zones-Pan'America
(North and South America),
Pan-Europe (Europe, the Middle East,
and Africa), and Pan-Asia (Asia and
Australasia). Each would be comprised of
a core and periphery, and this complementary
trade relationship would reduce
the necessity to trade outside the blocs.
Obviously, in this world, the British and
French empires would be carved up, and
the United States, Germany, and Japan
would dominate their respective blocs
(O'Loughlin and van der Wusten 1990).
The German geopolitical writers argued
that, by the Monroe Doctrine, the United
States had effectively created an American
bloc, and in 1941, Japan actually
declared the establishment of the
"Greater East Asia Co-prosperity Sphere"
after its military successes on the Asian
mainland and on the offshore islands.
There is little evidence that the putative
panregions developed along the
expected "natural lines." After World War
II, the United States extended its core
trading area to encompass Western Europe
and Japan more firmrly (Nierop and
de Vos 1989). Africa would have been an
obvious case to remain closely attached to
Europe in a panregional world, but the
evidence shows that the links weakened
after African independence in the early
1960s (O'Loughlin and van der Wusten
1990). Cohen (1991, 566) calls the panregional
concept "outdated" because the
Third World is marginal to the central
134
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
TRADE BLOC FORMATION
economic pursuits of the rich core powers.
Certainly, this dismissal of Third World
trade importance cannot be true of East
Asia or parts of Latin America and the
Middle East.
The post-World War II reappearance
of the panregional concept indicates that
it retains appeal beyond the narrow
German geopolitical writings. Taylor
(1992) identifies its presence in British
left-wing writings of the 1930s and 1940s,
and the strategy of the "three Monroes"
was widely circulated in the immediate
postwar debates about Britain's role in
world affairs (Taylor 1990). In the British
view, the three remaining global powers
in 1945 (the United States, the United
Kingdom, and the USSR) each had
specific regions of dominance (akin to the
claimed United States dominance over
Latin America in the Monroe Doctrine)
and the postwar geopolitical order would
be determined by the alliance of two of
them against the other. Modern versions
of the theory rarely recognize its historical
antecedents, but the proposed contemporary
"natural" trading blocs closely follow
the lines of the German maps of the
1930s.
The highly debated view of the "end of
history" thesis (Fukuyama 1992), producing
a borderless and highly interdependent
world economy, is paralleled by
economists envisioning "the end of geography"
in the close integration of the
world trade and financial systems
(O'Brien 1992). In this view, state control
will ebb, and international companies will
be the main actors in the world economic
arenas. The "end-of-geography" thesis
also challenges the GATT concept, which
is based on nation-states bargaining for
their national interests. The thesis also
challenges the "reassertion of economic
geography," the integration of national
economies in regional blocs. The appearance
of a borderless world economy has
produced a backlash: "the end-of-geography
challenge presented by the liberalization
of trade was met by a strong
geographical counter-action (non tariff
barriers), ensuring that each old barrier
was replaced by a new one" (O'Brien
1992, 90). Further, the shift from nationstate
to trade bloc control is generated, in
part, by global structural forces pushing
from regional to more global arrangements
(O'Brien 1992). The end-of-geography
is really about the end-of-sovereignty
of nation-states.
The geo-economic world of the 1990s is
similar to that of the 1930s in another
respect. After a generation of U.S. economic
hegemony, the world economy has
acquired more balance. U.S. power and
influence over the other global economic
powers, Japan and the European Community,
is waning, while the set of rapidly
industrializing states in East Asia continues
to gain global economic prominence.
Economic power is more dispersed; any
discussion of the evolution of the world
trading system should be placed in the
context of the relative decline of the
United States. Except for a cadre of"Cold
Warriors," who anticipate the rise of
Russia to global status, there seems little
challenge to the notion that international
economic competition will become more
important for national security and welfare
(Gilpin 1992). The popularity of books
by Reich (1991), Garton (1992), and
Thurow (1992) pay testimony to the belief
that the United States has passed to a
different world stage, one of trading states
(Rosecrance 1986), in which U.S. leadership
is more questioned and its military
dominance counts for less. One strategy to
extend U.S. influence is to build a
partnership with one of the potentia
challengers, Japan being the most obvious
choice. There, U.S. military power would
be wedded to Japanese economic power
to build a trans-Pacific alliance to counter
a European Community-Eastern Europe
bloc (Inoguchi 1988; Wallerstein 1991).
The change in the U.S. position relative
to the world economy can be clearly seen
in the rhetoric of trade. From being the
prime mover behind the GATT world (the
institutions that controlled the world
economy after 1945, such as the International
Monetary Fund, the Bretton
Woods agreements on currencies, the
135
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
ECONOMIC GEOGRAPHY
World Bank for development, and the
United Nations for world order), the
United States increasingly has turned to
the unilateralist notions of "managed
trade." Its enormous and consistent trade
deficits ($118 billion in 1993) have eroded
public confidence in notions of free trade.
Bhagwati (1991, 1993) has been the most
vocal critic of the changing U.S. policies
in GATT, arguing that unilateral American
policies such as sanctions in the
Omnibus Trade Bill of 1988 ("Super 301")
against perceived trade violators run
fundamentally counter to the GATT
agreement, which has the force of treaty
for the United States. Suffering from
"diminished giant syndrome," the United
States, in his view, is demonstrating
"trade panic and attendant petulance"
(Bhagwati 1993, 155). Dhar (1992) and
Grant (1993) discuss the dramatic shift in
U.S. trade policies and Washington's
special attraction to nontariff barriers.
Calculations by Huffbauer, Berliner, and
Elliot (1986) show that special protection
on imports to the United States rose from
5 percent in 1955 to 21 percent in 1984.
In 1984, $44.4 billion of imports were not
allowed into the United States, equal to
13 percent of total imports. The equivalent
tariff percentage of the nontariff
barriers on steel, car, and textiles imports
is equal to a tariff of 24 percent (de Melo
and Panagariya 1992).
U.S. policy is increasingly challenging
the central principles of GATT. A perception
of trade based in power and politics is
taking hold, as recent U.S. administrations
have increasingly viewed the international
economy as a political mine field
that must be negotiated carefully and a
place where enemies and allies can be
identified. Authors such as Reich (1991),
Thurow (1992), and Garton (1992) support
this view and call for national industrial
and trade policies as part of an overall
restructuring of the U.S. political economy
in the post-cold war world. GATT is
not incompatible, in principle, with a
world of regional free-trading areas (Bhagwati
1993; Preeg 1993; Jackson 1993), but
free trade advocates believe that trade
blocs are bad in practice. If the level of
import restrictions is not higher to nonmembers
than before bloc formation,
GATT can tolerate such trade blocs. Since
50 percent of world trade is now contained
within the three global regions (up
from 40 percent in 1980), Preeg (1992,
1993) believes that future trade talks
should recognize these realities and use
them to make a base for more open
multilateral trade. The United States
initiated the Uruguay Round of GATT
trade negotiations in 1986 and pushed
hard (successfully, in the end) for liberalization
in services, intellectual property,
and agriculture, areas in which the United
States has a comparative advantage. Despite
a drop in the average tariff from 20
percent in 1950 to 4 percent in 1992 and
predicted gains of $80 billion for the
European Union, $25 billion for Japan,
and $20 billion for the United States with
the implementation of the Uruguay
Round ("GATT" 1993), continued trade
disputes between the big three (the
United States, Japan, and the European
Union) will likely promote even more bloc
formation. Gilpin (1992) thinks that the
processes of both trade liberalization and
bloc formation will occur simultaneously,
in a complementary manner.
The two biggest economies, the United
States and Japan, have joint and parallel
interests in promoting free trade and an
open world economy. They both trade
widely beyond their home regions, and
their recent economic growth has been a
result of booming exports. The U.S.organized
unipolar world of the immediate
postwar period has shifted to a world
of approximately three equal great powers.
The "spatial fix" for the negative
impacts of their competition, a geographic
division of the spoils along panregional
lines, indicates that "less competition
from outsiders will be traded off against
more competition from insiders" (Thurow
1990, 38). The rhetoric of free trade has
been replaced in the United States by
that of managed trade, including trade
rules negotiated for the free trade areas.
136
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
TRADE BLOC FORMATION
Blocs and Politics
The debates about bloc formation must
be seen within the context of the new
trade theory. Traditional models of trade
emphasized factors of endowment in a
Hecksher-Ohlin framework in which a
country's exports incorporated its relatively
abundant factors and its imports
were composed of scarce goods. Leontief
(1956) found this model to be questionable
because, in contrast to expected
patterns, the United States exported
labor-intensive goods and imported capital-intensive
products. A Linder (1961)
model, where similarity in demand preferences
between two countries is the
basis for trade in manufactures, with per
capita income serving as an indicator of
domestic market preferences, offers a
better explanation of trade between large
rich countries, as well as for the (ex)
Soviet Union (Hanink 1988, 1994; Hanink
and Kier 1993).
As the economic predictors of trade
weaken, political relations between states
offer a significant additional explanation
for the patterns of international trade in
the postwar period. Though there is
strong evidence that trade and politics are
related, debate continues about the direction
of the relationship. In Pollins's (1989)
view, trade follows the flag, but Gasiorowski
(1986) finds that political relations
precede the development of trade flows.
The imports and exports of the United
States seem much more responsive to its
political relations than is the case for
Japan (Anselin and O'Loughlin 1990;
O'Loughlin 1993b). In the geo-economic
world of the late twentieth century,
political relations may not be as apparent
as they were in the days of the cold war,
but the extent of trade relations between
states is modified by politics, in addition
to the usual predictors of distance and
economic size.
Trade blocs are not expected to develop
ex nihilo. According to Krugman (1991a),
trade blocs in the Americas, Europe, and
East Asia follow "natural trade zones."
These natural regions are areas of intensive
intrazonal interaction resulting from
small political and geographic distances
between trading partners and similar
levels of economic development. A traditional
notion of "neighborhood" prevails
in building free trade areas (Weintraub
1991). If free trade zones follow these
"natural trade" lines, there will be little
appreciable trade-diverting effects (Krugman
1991b, 1991c). Natural and legal
trade zones will mostly overlap. This view
is challenged by Bergsten (1991), who
worries that trade blocs would be deepened
to encompass monetary and other
economic linkages and would eventually
be transformed to exclusionary regions.
This development would have important
implications for the world economy, and
in Bergsten's scenario the end result
would be a world of exclusionary and
competitive regions among which little
interaction would take place. Bergsten
also dismisses the idea of "natural trade
zones" in a world in which distance is less
and less a useful predictor of trade as the
world's regions become increasingly di-
versified.
There is broad agreement that fear of
Japanese imports has generated some of
the impetus for free trade areas. Japan has
prospered in the GATT-world of the past
30 years, and it is clearly in Japan's
interests that the current regime should
change little. By contrast, Western Europe
and, to a lesser extent, the United
States are in fear of losing local and
regional markets to Japanese exporters.
By tightening regional bonds while raising
barriers to outside exporters, it is hoped
that Japanese market penetration can be
halted. In contrast to the formal steps of
building free trade zones in Europe and
North America, no such effort has yet
occurred in Japan's home area of East and
Southeast Asia, though Maidment (1989)
believes that a "yen bloc" may eventually
result from Japanese trade, aid, financial,
industrial, and development linkages.
Suggestions of an "Asian-Pacific economic"
region are also premature, even
though 'the region has 50 percent of the
world's trade, up from 40 percent in the
137
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
ECONOMIC GEOGRAPHY
1960s. The region is unlikely to form a
free trade area because of fear of Japanese
domination of the smaller Asian economies,
domestic pressure in the United
States to exclude Japanese imports, East
Asian reluctance to remove barriers to
U.S. exports, especially farm products,
and the continued presence of Communist
regimes in the Asian half of the
proposed free trade area (Anderson 1991).
Japan resists the notion of an East Asian
trading bloc, though its influence is
growing in the region (Gilpin 1992, 33).
Because of the export orientation of the
Japanese economy, Japan's trade relations
are intimately linked to its domestic
economic and regional geopolitical position.
Japan is regionally isolated, is
strongly dependent on imports of raw
materials, remains strategically and militarily
dependent on the United States,
and carries a historical legacy of imperialism
that makes putative partners leery of
Japanese foreign economic policy (Grant
1993). Japan's role is quite specialized, as
a source of capital goods (75 percent of
total exports) and as intermediary between
the United States and the East
Asian Newly Industrializing Countries
(NIEs) (Muraoka 1991). Japanese export
growth slowed dramatically between 1990
and 1994, its economy entered recession,
and its trade surplus was strongly related
to the high value of the yen rather than to
export volume. The NIEs are chipping
away at Japan's traditional trade leadership
in capital goods (Corker 1991).
Debate about the evidence for the
formation of a "yen bloc," an American
bloc, and a European bloc has been
hindered by sparse data, contrasting
research methods, and inconsistent political
goals. Frankel (1991) has argued that,
after controlling for the rapid growth of
the East Asian economies, the impression
of Japanese regional orientation vanishes.
Schott (1991) and Lloyd (1992) maintain
that, while Europe is becoming a coherent
trade bloc and the Americas also show
tendencies in that direction, given the
weight of evidence, East Asia will not
meet the four conditions that Frankel
detects in successful free trade areas
(similar levels of GDP, geographic proximity,
similar or compatible trading regimes,
and political commitments to
regional organizations). Fieleke (1992)
finds that most of the preferential trading
arrangements that have appeared recently
account for about two-thirds of world
trade. Over time, this preferential trade
has come to resemble trading blocs
because trade is becoming more "inward"
among members of the agreement and
less "outward" with the rest of the world.
Anderson (1991) and Anderson and
Norheim (1993a, 1993b) seem to draw
similar conclusions, namely that the case
for bloc building is premature and that
the evidence for the existence of free
trade areas is currently mixed.
One clear agreement among the "unholy
trinity" (Western Europe, the
United States, and Japan) (Van Agt 1991)
is that they will not deliberately undermine
the GATT world, as the frenzied
final negotiations of the Uruguay Round
in late 1993 showed. However, their
protectionism and decreasing reluctance
to use political means to pursue economic
ends jeopardizes the world trading system,
in the eyes of free traders (Crook
1990). Three options are possible in a
world of three great powers (O'Loughlin
1992): the United States against an
antihegemonic alliance of Japan and Europe;
Japan against Europe and the
United States in a trade protectionist
alliance; and Europe against a Pacific
bigemony of Japan and the United States
(Inoguchi 1988; Aho 1993). The bargaining
and posturing in the Uruguay Round
of the GATT talks, 1986 to 1994, provide
some initial indications of possible future
conflicts. One thing remains clear: none of
the three great economic powers (Germany,
Japan, or the United States) will be
able to organize its home region without
considering the opposition emanating
from regional traditions and historical
memories. Regional partners are understandably
suspicious of plans for the
integration of their economies with their
large neighbors.
138
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
TRADE BLOC FORMATION
Methodology and Data
In his challenge to Krugman (1991b),
Bergsten (1991, 48) states that Krugman's
assertion that blocs will follow "natural trading
lines" is "an empirical question on
which Krugman offers little supportive evidence."
The same conclusion could be
made for the wider issue of the development
of trade blocs. Previous attempts to
measure either the trade-diverting effects
of blocs or the movement toward blocs have
relied on simple proportions. Typically, the
ratio of trade to countries within the (predefined)
bloc to trade outside the bloc is
used, and there is a tendency, criticized by
Frankel (1991) and Anderson and Norheim
(1993a, 1993b), to accept these proportions
at face value. These ratios take no account
of the overall trends in trade, so that recently
more attention has been given to
standardized proportions-that is, controlling
for the increasing size of trade volumes
in the respective regions. Based on
such controls, Frankel (1991) and Anderson
and Norheim (1993a, 1993b) find little
support for the hypothesis that the world
is devolving into regional blocs.
While recent studies have clarified the
debate on the nature of the empirical base
of blocs to some extent, they have not
examined the position and role of individual
states in the putative blocs. Rather,
economists have taken the blocs as given
and calculated trade ratios on the basis of
predetermined geographic bloc-lines.
This approach hides intrabloc country
differences and involvement with regional
partners as well as possibly hiding extrabloc
attachments. In this study, in addition
to replicating traditional aspatial
analyses based on indices of intensity of
trade and of the propensity to trade
extraregionally, we offer a local perspective
to the study of bloc formation by
evaluating the spatial relations of 114
countries to the big three exporters-the
United States, Germany, and Japan-over
the period 1968 to 1992. As well as
calculating export trade shares by panregion
(our geographic delimitation of a
bloc), we offer a more nuanced and
geographically specific view of how the
trade of the three great trading states has
developed in the past quarter century.
We use an exploratory spatial data analysis
(ESDA) approach, which allows us to
determine whether the concern about the
development of blocs is warranted. The
ESDA perspective is well suited to our
purposes in this paper, given its inductive
or data-driven approach to gain insights
into a question for which theoretical and
empirical insights are currently underdeveloped.
The revelation of patterns, proposition
of hypotheses, and suggestion of
models is effectively carried out in combination
with spatial data manipulation,
storage, and display in a geographic
information system (Anselin and Getis
1992; Anselin 1993a).
After reexamining the trade proportions
evidence, using the two growth/proportions
indices suggested in Anderson and
Norheim (1993a, 1993b), we shift to a
spatial analysis perspective and carry out
an investigation of both global and local
measures of spatial association. Our summary
global measure is Moran's I (e.g.,
Cliff and Ord 1973), which indicates the
extent of significant spatial clustering of
export and import shares for each of the
three main trading nations with the 114
countries in the sample and examines its
temporal trends in the period considered.
This measure is illustrated by means of a
Moran scatterplot (Anselin 1993b, 1994)
to show the decomposition of the association
into spatial clusters of similar high or
low values and clusters of dissimilar
values and to assess the existence of
outliers. This focus on local indicators of
spatial association (Anselin 1995) is further
explored by means of maps of significant
Gi* statistics (Getis and Ord 1992), which
illustrate the existence of separate spatial
clusters and show which countries deviate
from the regional patterns. Consideration
of these special cases offers insights into
the reasons for the anomalies and provides
further details on the trade blocbuilding
process.
All of our analyses are conducted at
five-year intervals (1968, 1973, 1978,
139
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
ECONOMIC GEOGRAPHY
1983, 1988, and 1992). Since we are
primarily interested in the longer-term
developments over the quarter century,
these temporal intervals were deemed
sufficient. Furthermore, year-to-year
changes in trade proportions are minuscule.
The data are taken from the
International Monetary Fund's yearly
report Direction of Trade Statistics
(1968-88), with some additional data from
United Nations and Organization for
Economic Cooperation and Development
sources. The 114 countries in the study
include all states with a population over 1
million who were members of the United
Nations during the entire time interval.
The country set, therefore, excludes states
such as Switzerland, North and South
Korea, Hong Kong, Taiwan, and former
East Germany. Furthermore, Lesotho
and Swaziland (two enclaves) were not
considered because their sole neighbor,
South Africa, was not included (and thus
contiguity analysis is meaningless). For
consistency's sake and to reduce bias
introduced by trade between the three
great powers in the study, they were not
included as trading partners of each other.
A total of 114 countries, classified by
panregion, were included in the study.'
1 The regions and their members are (1)
Pan-Europe-United Kingdom, Austria, Belgium-Luxembourg,
Denmark, France, Italy,
Netherlands, Norway, Sweden, Canada, Finland,
Greece, Ireland, Portugal, Spain, Turkey,
Yugoslavia, Iraq, Israel, Jordan, Kuwait,
Lebanon, Oman, Saudi Arabia, South Yemen,
Syria, United Arab Emirates, Egypt, North
Yemen, Algeria, Botswana, Burundi,
Cameroon, Central African Republic, Chad,
Congo, Zaire, Benin, Ethiopia, Gabon, Gambia,
Ghana, Guinea, Ivory Coast, Kenya,
Liberia, Libya, Madagascar, Malawi, Mali,
Mauritania, Mauritius, Morocco, Niger, Nigeria,
Rwanda, Senegal, Sierra Leone, Somalia,
Sudan, Tanzania, Togo, Tunesia, Uganda,
Burkina Faso, Zambia; (2) Pan-AmericaCanada,
Argentina, Bolivia, Brazil, Chile,
Colombia, Costa Rica, Dominican 'Republic,
Ecuador, El Salvador, Guatemala, Haiti, Honduras,
Mexico, Nicaragua, Panama, Paraguay,
Spatial contiguity between countries was
defined as simple contiguity-that is,
having a land border in common and/or
having a border within 200 miles over
water. Two "islands," Canada and New
Zealand, were defined as contiguous to
Mexico and Australia, respectively. Since
our concern was only with trade blocs, we
did not include other indicators of bloc
building, such as development aid and
foreign direct investments.
In reexamining the trade shares, we
decided to use a priori regions, or regions
determined by panregional theory and
not by current trade patterns. This avoids
tautology. We used the four regions
identified by the geopolitikers of the
interwar period as promoting a global
equilibrium with each of the four great
powers (Germany, Russia/Soviet Union,
Japan, and the United States) with their
own region of influence. In this quadripartite
world, reciprocal trade relationships
within each panregion between an industrial
North and a, resource-rich South
would preclude the need to pursue trade
beyond the boundaries of the panregion.
The regions and their country members
offer a useful starting point to the
reexamination of bloc developments and
make a more defensible starting point for
regional divisions, since they include
countries in all regions. In previous
studies, countries beyond the immediate
regions of the major powers were simply
excluded from study.
Bloc Building? Evidence from
Intensity and Propensity of Trade
Though central to the issue of bloc
building and the correspondence of blocs
and "natural trading areas," the question
Peru, Uruguay, Venezuela, Jamaica, Cuba; (3)
Pan-Asia-Australia, New Zealand, Bhutan,
Myanmar, Kampuchea, Sri Lanka, India, Indonesia,
Laos, Malaysia, Nepal, Pakistan, Philippines,
Singapore, Thailand, Vietnam; (4) PanRussia-Iran,
Albania, Bulgaria, P.R. China,
Czechoslovakia, Hungary, Mongolia, Poland,
Romania, USSR/Russia.
140
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
TRADE BLOC FORMATION
of whether the world trading system is
becoming aligned along regional lines is
still unresolved. In a longitudinal analysis
of total European trade since 1928,
Anderson and Norheim (1993b) conclude
that extra-European trade has grown
since the 1930s and that the perception of
a "Fortress Europe" is inaccurate. Though
intra-Europe trade has also been increasing,
this trend is due more to the growth
of the individual national economies than
to a turning inward by Europe. In another
paper, Anderson and Norheim (1993a)
extend the series in time and reach
similar conclusions: world trade is not
regionalizing if one controls for economic
growth in the respective world regions
and uses two complementary concentration
indices, intensity of trade and regional
propensity to export. Because
Anderson and Norheim consider only
European trade, however, their studies
and conclusions are limited. To offer a
more globally complete picture of regionalization
trends, we used the same two
indices as Anderson and Norheim to
examine the exports of Germany, the
United States, and Japan in the past
quarter century.
The first index, measuring intensity of
trade, is computed as
Ii = x,/mi, (1)
where xi = the share of country i's
exports going to region j, and m = the
share of region j in world imports (net of
country i's exports, since i cannot export
to itself) (Drysdale and Garnaut 1982).
This index computes the extraregional
trade share defined by the region's share
with the rest of the world divided by the
rest of the world share of global trade.
This controls for overall importance in the
world trade system, and over time the
index takes account of interregional gains
and losses. If trade is not geographically
biased (whereby the ratio of i's trade
going to region j is equal to region j's
share of the global total), it will have a
weighted value of 1.0 for all regions.
The changing trade intensities are
shown in Table 1. For all three exporters,
the figures for Pan-Russia (the former
Eastern Europe) are volatile, reflecting
the ebb and flow of the cold war and the
signing and abrogation of trade agreements.
Since exports by all three Western
powers to the region are so small, the
indices cannot be regarded as reliable or
consistent. For U.S. exports, trade intensifies
over time in the home territory
(Pan-America), especially since 1973 (after
an initial drop from 1968), with a corresponding
decrease after 1983 in PanEurope
(Western Europe, the Middle
East, and Africa) and in Pan-Asia. Table 1
can be read such that U.S. exports to
Latin America are three and a half times
as much as would be expected from Latin
America's ratio of world imports, while
U.S. exports to Pan-Europe are only half
of the expected total.
German and Japanese exports also
indicate a home region bias, but not to the
extent of the United States. In contrast to
what might be expected from popular
opinion, Germany has only 30 percent
more exports than expected to the importers
of its home region, and this figure is
stable for the past quarter century.
German exporters have lost ground in
Latin America while gaining a little in
East Asia. Of the three exporters, Germany
seems to follow global trends most
closely. Japan's patterns show a decrease
in Asian concentration over time, a
significant contrast to the stability of
Germany and the growth of the United
States in their respective home regions.
Small intensity ratio growth in the other
two major importing regions balances this
regional decline, but the overall picture
for Japan is a matching of its exports to
world import trends.
Calculations of trade intensity indices
support arguments that bloc building in
exports is exaggerated. Only the United
States shows signs of such behavior in its
immediate area, and even there the
change over 25 years is fairly small. All
three exporters show growth in at least
one region outside the home area, and an
overall picture of balance and stability in
141
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
ECONOMIC GEOGRAPHY
Table 1
Index of Trade Intensity
Pan-America Pan-Europe Pan-Russia Pan-Asia
United States
1968 2.980 .598 .175 1.126
1973 2.014 .358 5.500 3.562
1978 2.563 .494 .630 .382
1983 2.260 .535 .356 1.894
1988 3.145 .542 2.126 1.329
1992 3.475 .493 3.386 1.323
West Germany
1968 .409 1.296 .772 .267
1973 .357 1.074 14.500 .382
1978 .341 1.181 1.630 .250
1983 .241 1.238 1.333 .455
1988 .258 1.457 .586 .282
1992 .261 1.293 .589 .578
Japan
1968 .745 .548 .596 3.407
1973 .992 .625 5.833 4.281
1978 .928 .625 1.021 3.455
1983 .956 .584 .911 3.242
1988 .719 .756 .310 2.226
1992 .846 .481 .323 2.922
Note: The index of trade
export intensities c
scenarios of a tradin
into its component
also reinforces earli
ratios must be con
shares of global tot
changing patterns pr
A second index, pro
and Norheim (199
the propensity of
extraregionally by
regional output, m
mestic Product (GDP). The index is
computed as
pi = ti * ii, (2)
where ti is the ratio of i's total exports to
i's GDP, and I. is calculated in Equation
(1) above as tle trade intensity index.
Unlike the ratios in Table 1, the values in
Table 2 should not be compared directly
between differently sized regions at a
point in time because of widely varying
GDPs. The index is strongly influenced
by the size of the regional GDPs, and,
other things being equal, ti is dependent
on the size of the economy. Using
index, Anderson and Norheim (19
determined that, although Western
rope's propensity to trade intraregio
has risen consistently since the 1930
propensity to trade with the rest o
world has at least been maintained. The
share of Europe's GDP traded with
non-Europeans has been stable for the
past 60 years.
Most of the indices in Table 2 increase
over time, as might be expected from the
general global trend of increasing ratios of
trade to GDP. It is the differences in the
increases that make the results interesting
and noteworthy. Over the period as a
whole (1968-92), both the United States
and West Germany show sizable increases
in the propensity to export to their home
regions (Pan-America and Pan-Europe,
respectively). In contrast, the home region
orientation for Japan decreased over
the 25-year period, though with some
increases in the intervening years up to
1983. All three exporters show an upward
trend in the propensity to export extra-
regionally.
142
This content downloaded from
147.251.68.36 on Wed, ffff on Thu, 01 Jan 1976 12:34:56 UTC
All use subject to https://about.jstor.org/terms
TRADE BLOC FORMATION
Table 2
Index of Propensity to Export Extraregionally
Pan-America Pan-Europe Pan-Russia Pan-Asia
United States
1968 .101 .019 .005 .037
1973 .102 .018 .028 .187
1978 .163 .032 .041 .165
1983 .136 .027 .017 .098
1988 .146 .025 .010 .062
1992 .216 .031 .020 .082
West Germany
1968 .061 .195 .005 .037
1973 .063 .190 .435 .066
1978 .069 .241 .332 .051
1983 .052 .271 .291 .099
1988 .066 .372 .149 .072
1992 .054 .270 .123 .121
Japan
1968 .044 .032 .015 .199
1973 .063 .040 .037 .271
1978 .067 .044 .072 .246
1983 .075 .046 .071 .256
1988 .044 .046 .019 .136
1992 .050 .029 .018 .175
Note: The index of propen
Though the U.S. prop
Latin America (a reg
low GDP and grow
sizably in the past q
only a brief decreas
1983), it lost groun
period in both PanPan-Asia
(1973-88).
respectively, the we
est-growing econom
this decade-long rever
trends understand
concern to U.S. polic
sal is partly due to
U.S. dollar toward t
(1985 to 1988); after
value, the ratios inc
contrast to the United States, West
Germany saw an overall growth trend to
1988 in its propensity to export to all
regions except East Asia. Over this
quarter century of dramatic global economic
change, evidence of the strength of
German exporters is seen in the sizable
increases in the propensity to export to
the wealthiest areas of the world economy,
except for Pan-America, where the
ratios were stable.
The effect of currency fluctuations may
be seen in the 1992 Japanese index for
Pan-Europe as the yen reached its apogee:
the Japanese propensity to export to
this region declined after a quarter
century of growth. The indices indicate
stable Japanese export propensities for
the other two areas, Pan-America and
Pan-Russia, but indicate a sizable drop in
East Asia in the past few years. Since the
mid-1980s, Japanese propensity to export
to its neighbors has dropped sharply, not
in the total volume of exports, but in
exports as a ratio of Japanese GDP,
related to the share of the world's imports
accounted for by East Asia. The result will
appear somewhat surprising to those
impressed by the ratios and raw numbers
of Japanese exports. Based on the index of
propensity to export extraregionally, it
appears that, since 1968, the United
States is a very successful intraregional
exporter; it has managed to keep pace
143
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
ECONOMIC GEOGRAPHY
with West Germany and Japan outside its
small home region.
The overall impression from trends in
the indicies of intensity and propensity to
trade extraregionally since 1968 is the
relative stability of the regional exporting
patterns. There is little evidence of bloc
building in the exporting patterns, though
the United States and West Germany
seem to be developing closer trading ties
to their Latin American and Pan-European
neighbors, respectively, while Japan
is (relatively) leaving its home region. The
United States has four times more trade
with its neighbors than the respective
ratios for Japan and West Germany (the
value was twice as much in 1968); West
Germany has 9 times more trade with its
Pan-European neighbors than Japan and
the United States do (the figures were 10
and 6 times as much in 1968); and Japan
has 2 and 1.5 times more trade with its
East Asian neighbors than the figures for
West Germany (about 6 times as much in
1968). Japan's exports are diffusing more
evenly around the world, while its economic
rivals continue to concentrate on
their home territories. It is little wonder
that the strongest recent exponents of
regional free trade areas have been the
United States and West Germany.
Temporal and Spatial Trends in
Trade Shares, 1968-1992
While we were able to replicate and
extend the conclusions of the aspatial
studies of Anderson and Norheim that
bloc building in exports is not very
advanced, it is yet unclear to what extent
the individual importing countries contribute
to the regional patterns. Attention
to spatial association and to countryspecific
trade patterns (both import and
export) has been lacking in the existing
examinations of the geography of trade. In
this section, we consider both global and
local indicators of spatial association
(LISA) to measure trends in spatial
clustering of export shares and to examine
the contribution of individual countries.
These shares are computed as
Ei = x i/mj, (3)
where x0 are the total exports of i (United
States, Germany, or Japan) toj, mj are the
total imports of j, and
M = xjiej (4)
where xji are the total imports by i
(United States, Germany, or Japan) fromj,
and ej are the total exports ofj. Although
these figures are fractions, they are not
proportions in the usual statistical sense,
in that they do not add up to one for each
i (the ratios would add up to one for each
j, but such a perspective is not considered
here).
The recently suggested LISA approach
(Getis and Ord 1992; Anselin 1995) allows
us to identify regional anomalies, represent
them cartographically, and suggest
possible reasons for local trends and
deviations. Our point of departure is a
global measure of spatial autocorrelation,
the familiar Moran's I, which is expressed
as
I = (N/So) EiEj Wi XiX/i Xi2, (5)
where w0 is an element of a spatial
weights matrix W that shows whether or
not i and j are contiguous (typically, and
also in our study, this weights matrix is
row-standardized such that its row elements
sum to one); xi and xj are observations
at locations i and j, respectively
(measured as deviations from the mean);
N is the number of observations; and So is
a normalizing factor, equal to the sum of
all weights (EiEj wij). Since the data (the
export and import shares defined in
Equations (3) and (4)) did not conform to
the normal distribution, we based the
significance of Moran's I on the randomization
assumption and computed the
associated standardized z-value accordingly
(for details, see Cliff and Ord 1973,
1981; Upton and Fingleton 1985). All
computations were carried out by means
of the SpaceStat spatial data analysis
144
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
TRADE BLOC FORMATION
software (Anselin 1992). The temporal
trends in the Moran's I indices for export
and import shares of the three countries
with respect to their 114 partner countries
can be assessed from the results reported
in Table 3.
The indices of spatial association for
both export and import shares of the
United States are strongly significant and
positive for every year in the sample,
indicating strong spatial clustering of both
the destination countries of U.S. exports
and the origin countries of U.S. imports.
Of the three main traders considered, the
Moran's I for the United States is the
most stable, ranging between 0.3 and 0.5
for exports and between 0.3 and 0.4 for
imports, with no detectable trend in
either direction. This absence of any trend
over a quarter century period confirms
earlier caution about the significance of
bloc formation in U.S. trading patterns.
The indices for West Germany show a
higher degree of fluctuation (between
0.15 and 0.54 for export shares, and
between 0.17 and 0.35 for import shares),
and they are consistently positive and
highly significant (at p < 0.01) throughout
the period. In contrast to the United
States, an increasing trend in spatial
association is indicated for West Germany,
with higher significance in every
year after 1973 for export shares and after
1978 for import shares, though with a
decrease in 1992. This decrease may be
due to the disruption caused by the
consolidation of the economies of the two
Germanies in the previous year.
Of the three great economic powers,
the trade pattern of Japan shows the
greatest variation in spatial clustering.
Export shares do not show significant
spatial autocorrelation in both 1968 and
1992, though they are highly significant
and positively associated in the other
years; imports show significant and positive
spatial association in all years. The
magnitude of the coefficients and their
significant move up and down from year
to year does not reveal any trend toward
Table 3
Morans I and Associated z-scores for Export and Import Shares
Exports Imports
I z I z
United States
1968 0.321 4.56 0.392 5.53
1973 0.435 6.16 0.404 5.75
1978 0.418 6.00 0.309 4.40
1983 0.303 4.44 0.362 5.14
1988 0.484 6.98 0.407 5.76
1992 0.401 5.72 0.369 5.22
West Germany
1968 0.283 4.13 0.172 2.81
1973 0.149 2.67 0.329 4.68
1978 0.271 4.00 0.226 3.27
1983 0.391 5.69 0.319 4.63
1988 0.535 7.74 0.321 4.61
1992 0.454 6.43 0.349 4.90
Japan
1968 0.079 1.61* 0.094 3.31
1973 0.407 5.94 0.061 3.53
1978 0.356 5.61 0.302 4.35
1983 0.633 8.83 0.455 6.57
1988 0.502 7.10 0.168 2.55
1992 -0.004 0.16* 0.213 3.16
* Values not significant at
145
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
ECONOMIC GEOGRAPHY
greater geographic concentration of Japan's
trade patterns. This oscillating pattern
instead indicates a dynamic and
globally changing export position, which
contrasts with the more static and entrenched
U.S. and German trade rela-
tions.
While Moran's I provides an indication
of spatial clustering of similar values, it
does not necessarily imply (spatial) trade
bloc formation. In fact, the positive spatial
autocorrelation measured by Moran's I in
Table 3 includes both similarity of small
values between a location and its neighbors
as well as similarity of large values.
In the context of trade bloc formation,
only the latter is relevant. The Moran
scatterplot suggested in Anselin (1993b,
1994) provides a means to disentangle the
different country contributors to the
spatial autocorrelation index. It consists of
a scatterplot of the spatially lagged values
for each observation (Wx)i against the
observation xi (with the x in standardized
form, i.e., with the mean subtracted and
divided by the standard deviation). The
(Wx)i, xi pairs give an indication of the
degree of association between the average
of the neighboring values and the value at
each location. Perfect spatial association
would result in (Wx)i, xi pairs located
around the 45 degree line. Poor spatial
association would be reflected in a random
circular cloud around the origin.
Since the x values are standardized, the
scatterplots are comparable between variables.
Also, values of either (Wx)i or xi
larger than two can be considered as
"extreme," in the sense that they are
more than two standard deviations away
from the mean (the so-called two sigma
rule).
The Moran scatterplots for the export
shares of the United States, Germany, and
Japan are presented in Figures 1 to 3 for
the six years in the sample (the plots for
import shares are not shown but follow a
similar pattern). The four quadrants in
each plot distinguish the four categories of
spatial association: two kinds of positive
association (similarity between the value
at a location and the values at neighboring
locations) and two kinds of negative
association (dissimilarity between the
value at a location and the values at its
neighboring locations). Positive spatial
association between small values (i.e.,
below the mean) is shown by the points in
the lower left quadrant; positive association
between large values (i.e., above the
mean) are shown in the upper right
quadrant. It is the latter that is of
particular interest for trade bloc formation.
Negative association is shown in the
upper left quadrant (between small values
at a location and high values at the
neighboring locations, or islands of low
values) and in the lower right quadrant
(between high values at locations and low
values at neighboring locations, or islands
of high values). As demonstrated in
Anselin (1993b, 1994), the slope of the
linear smoother in this scatterplot-that
is, the slope of a regression line of Wx on
x-is Moran's I. In this respect, the
stability of the slope for U.S. exports is
striking (Fig. 1) and contrasts sharply with
the oscillating line for Japan (Fig. 3). For
Germany (Fig. 2), the slope increases
steadily from 1973 on, with a slight
decline in 1992. It is clear from all figures
that the bulk of the positive association is
low-low-that is, between a country's
imports and that of its neighbors in the
lower left quadrant of the graphs-and is
not of particular interest in the context of
bloc formation. The distribution of the
points in the three country plots by
quadrant is detailed in Table 4.
Countries that are extreme with respect
to the linear regression-that is, countries
that do not fit the overall (global) pattern
of spatial association-are labeled with
their International Monetary Fund code
in the figures. For illustrative purposes,
the six most extreme observations are
labeled, although strictly speaking only
one or two of these should be considered
"outliers." For all three main trading
countries, the number of high-high spatial
associations seems remarkably stable, only
indicating some decline for Japan (in 1992,
the Moran's I was not significant for
Japan). The makeup of more than 20
146
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
US Export Share 1968
Moran's I = 0.321
243
*928
258 .
* 522 . - .
!.' .. . . . . . ... . . . . . . . ..
... :-.: 582
r"' :' .'* 156 US
Export Share 1973
Moran's I = 0.435
. 243
. 928
. 258
238 .
.. . 273
^ ^ ' ;156
-1 0 1 2 3 4 5
EXSH68
I { I r I
-1 0 1 2 3 4 5
EXSH73
US Export Share 1983
Moran's I = 0.303
US Export Share 1978
Moran's I = 0.418
. 243
. 928
. 258
273 .
" : '-- - '*""- 156
646
-1 0 1 2 .3 4 5
EXSH78
* 243
. 928
. 258
. 343
278.
.;.......... ?'.. ? ?....... ..
* 646
-1 0 1 2 3 4 5
EXSH83
US Export Share 1988
Moran's I = 0.484
. 243
. 928:
. 258
278.
? . ?
\*.\^ ~ . ~273
.... .......................
-^^t : * 156
I I I r I
1 0 1 2 3 4 5
EXSH88
US Export Share 1992
Moran's I = 0.401
. 258
928 . 278
253
243
. I4
P ?? * *443
-1 0 1 2 3 4 5
EXSH92
Figure 1. Moran scatterplots of the export shares of the United States, 1968-1992.
Go
r
uLJ
Go
co
Ix C -
U4
?:
-
r
XN
3Li
cg
I
x"
S:
a~)a:>
X) cm -
u
U"
-
aU
X-x
ui
This content downloaded from
1:ffff:ffff:ffff:ffff:ffff on Thu, 01 Jan 1976 12:34:56 UTC
All use subject to https://about.jstor.org/terms
German Export Share 1968
Moran's I = 0.283
German Export Share 1973
Moran's I = 0.149
112 . . 131
918 t..914
466. 944
.:-';.'.... ............
$::I-i~? ??? ?
-2 0 2 4 6
EXSH68
German Export Share 1978
Moran's I = 0.271
. 944
136 . 934
. 968
'. . . . , ?
?' 668
128
-2 0 2 4 6
EXSH73
German Export Share 1983
Moran's I = 0.391
o
112 136
914.:-944. .'
.....: ' .......................
. 518 '213
-2 0 2
EXSH78
German Export Share 1988
Moran's I = 0.535
6 -2 0 2 4 6
EXSH83
German Export Share 1992
Moran's I = 0.454
934
914 918 . ..82
156 t . .
I.I: * *443
II ....
-2 0 2 4 6 -2 0 2 4 6
EXSH88 EXSH92
Figure 2. Moran
co
x
cn
5
w
t,
3i
0
298 ?
228 288288
218 - . 944
.:.. ' . . ...
... ........
'r
r,
n-
x
u.
:t
O -
C\
914 *' 934 . . /
968.- I. . 122
.. ................648
~':liii: ' 648
do
co
x
x
Ul
3.
U)
x
3:
o This
content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
Japan Export Share 1968
Moran's I = 0.079
576 .' 193
. 548
*578
918 ? * 522
-2-2 0 2 4 6
EXSH68
Japan Export Share 1978
Moran's I = 0.356
cr,
0 Japan
Export Sh
Moran's I = 0.407
* 548
459 44 .576 .
522 *
. P . . 726
I I I
0 2 4 6 8
EXSH73
Japan Export Share 1983
Moran's I = 0.633
544 .
5 . 548.534
522 . .- . 5
514 -:. I 518
,-s. . . ....
-2 0 2 4 6 8
EXSH78
Japan Export Share 1988
Moran's I = 0.502
..544
238 196 :.
514 ...
. . . . II . .
.668 283
-2 0 2 4 6 8
EXSH83
Japan Export Share 1992
Moran's I =-0.004
238 .
233 .
! o 6 .. .
_ I
I I I I I T
-2 0 2 .1 6 8 -2 O 2 4 6 8
EXSH88 EXSH92
Figure 3. Moran
co
x
ui
co
r
Cf)
x
Uu
3: j
0
Cc
co
ID
co
co
'r
(f)
x
LI)
3r C4
C>
Inj
8
I
X)x
LL
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
ECONOMIC GEOGRAPHY
Table 4
Relative Presence of Four Types of Spatial Association
Positive Spatial Association
High-High Low-Low
United States
1968
1973
1978
1983
1988
1992
West Germany
1968
1973
1978
1983
1988
1992
Japan
1968
1973
1978
1983
1988
1992
23
23
22
20
18
21
31
23
31
29
26
29
24
32
33
27
27
13
59
63
64
70
78
71
55
63
43
53
67
58
49
56
48
65
60
74
countries in the upper right quadrant for
the United States is very stable as well,
with 14 countries (Argentina, Chile, Colombia,
Costa Rica, Dominican Republic,
Ecuador, El Salvador, Guatemala, Haiti,
Mexico, Panama, Peru, Venezuela, and
Jamaica) present in every year. These
countries are all in South and Central
America and the Caribbean, confirming
the continuing importance of the PanAmerican
bloc for U.S. exports.
For the United States, there is again a
remarkable consistency in outliers. The
two most extreme values in every year are
Cuba (928), because of the U.S. trade
embargo (low Cuban imports contrast
with very high U.S. imports by its
neighbors), and the Dominican Republic
(243), which is extreme in terms of its own
imports from the United States. Another
outlier that can easily be explained by
changes in political relations is Vietnam
(582), as an island of high imports (in the
lower right quadrant) in 1968 but not
later. The presence of Canada (156) in the
lower right quadrant is partly due to its
Negative Spatial Association
High-Low Low-High
17
12
13
12
12
10
15
16
15
12
6
12
13
12
12
15
14
7
15
16
28
17
7
20
28
9
16
10
12
13
13
17
17
12
15
14
lack of "neighbors," since the United
States is not included in this set.
For West Germany, there is a similar
stability in the set of high-high associations
(Fig. 2), with some indication of an
increasing reach, as well as a shift from a
primarily Southern and Western European
focus (six countries are present in
every year of the sample-Austria,
Greece, Portugal, Spain, Turkey, and
Yugoslavia; while the Western European
and Scandinavian ones are present except
in 1978-the United Kingdom, Belgium,
Denmark, France, Italy, the Netherlands,
Norway, Sweden, and Finland) to a
broader one that includes Eastern Europe
as well after 1973 (with Bulgaria, Czechoslovakia,
Hungary, Poland, Romania, and
the former Soviet Union present in every
year). There is no clear pattern for the
outliers, except that Albania (914) and
Hungary (944) are present in more than
half the years, Albania as an island of low
imports and Hungary because of higher
values than its East European neighbors.
Interestingly, in 1973, most outliers for
150
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
TRADE BLOC FORMATION
West Germany are in South America,
with indications of negative spatial association
(Uruguay (298), Chile (228), Paraguay
(288), and Bolivia (218)).
For Japanese exports (Fig. 3), the
dominance of the Pacific Hemisphere in
general and Southeast Asia in particular is
clear; eight countries are present in the
high-high quadrant in every year of the
sample (Australia and New Zealand, and
Myanmar, Indonesia, Malaysia, the Philippines,
Singapore, and Thailand). Since
1973, three Persian Gulf countries are
consistently present as well-Iran, Oman,
and the United Arab Emirates (Kuwait
until 1992)-whereas the southern Asian
cluster of Afghanistan, India, Nepal, Pakistan,
and Vietnam, as well as China, is
present for the first five sample years.
Clearly, here again, there is no indication
of a Japanese bloc forming outside its
traditional sphere of influence.
The indication given by the Moran
scatterplot is only exploratory, in the
sense that no measure of statistical significance
is associated with the individual
patterns. This is obtained by means of the
G,* statistic of Getis and Ord (1992),
which is one of a general class of local
indicators of spatial association or LISA
statistics (Anselin 1995). The Gj* statistic
is computed as
G* = Sj wy(d) ' y/j Yj, (6)
where wij(d) is an element in a binary
contiguity matrix, and yj is an observation
at location j. The Gi* statistics are
computed for each country and can be
interpreted as a measure of clustering of
similar values around that country (it
includes the value for the country, i, in
the computation). In contrast to the values
used for the computation of Moran's I and
the Moran scatterplot, the Gi* statistics
are based on the unstandardized observations.
The G* index only measures
positive spatial association in the usual
sense, but distinguishes between autocorrelation
of small values (the low-low
quadrant in the Moran scatterplot) and
that of high values (the high-high quadrant
in the Moran scatterplot). These two
cases are referred to as, respectively,
negative and positive spatial association
by Getis and Ord (1992). As for Moran's I,
significance of the Gi* statistic is assessed
by means of a standardized z-value, which
is asymptotically distributed as a normal
variate. As illustrated in Anselin, Dodson,
and Hudak (1993), the G,* statistics can be
easily visualized by means of a GIS or
mapping system as triangles pointing up
(for high-high association) or down (for
low-low association) and with varying
sizes corresponding to different significance
levels.
Maps of the Gi* statistics for export
shares in 1968 and 1992 for the three main
traders are given in Figures 4 to 9. These
maps confirm and clarify the impressions
gained from the Moran scatterplots. For
the United States (Fig. 4), the strong
positive association for the Caribbean,
Latin America, and South America is
highly significant and contrasts with the
significant clustering of small export
shares in an arc from Scandinavia through
Eastern Europe, the Middle East, and
East Africa south to Zambia. The clear
regionalization effect implied by the map
strengthens the notion that the evidence
for bloc building must be sought inside
the large world regions. Distance from the
United States and political-ideological
effects would seem to be strong predictors
of U.S. export shares (see also Anselin and
O'Loughlin 1990). The level of concentration
of U.S. exports in the Caribbean is
remarkably stable over the quarter century
covered by our data, as illustrated for
1992 in Figure 5. The maps for U.S.
imports (not shown here) are quite
similar, which leads to the conclusion that
U.S. trade is highly concentrated in
sub-panregions. Bloc building is not occurring,
as there have been no significant
changes since the late 1960s.
In 1968, a significant spatial cluster of
high German export shares can be found
in North and West Africa and the
Mediterranean region (Fig. 6). Marginal
significant association is shown for Centr
America (positive) and Southeast Asi
151
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
ECONOMIC GEOGRAPHY
Figure 4. Location of significant Gi* statistics for United States export shares, 1968.
(negative). Interestingly, these clusters do
not qualify as high-high associations in the
Moran scatterplot, since the values for the
standardized (Wx)i, xi pairs are slightly
below average (perhaps because the Gi*
statistics are based on unstandardized
values). This dispersed pattern counters
any notion of a home territory effect, and
most of the world experiences a random
distribution of German exports. By 1992,
the pattern of significant German high
and low export shares had changed
substantially (Fig. 7). High significant
values are found in both Western and
Eastern Europe (confirming the indication
given by the stable high-high associations
in Figure 2) and to a lesser extent
in central Africa, while spatial clusters of
low export shares are found in the Middle
East and to a lesser extent in South and
Southeast Asia. Hence, combining the
evidence of Figure 2 with Figures 6 and 7,
there is a strong impression of bloc
building in terms of the European neighbors
of Germany, but the notion of a
wider panregional bloc is not substantiated.
From the maps for 1973-88 (not
shown here), it follows that the major shift
toward Europe happened in 1973 and has
been stable since then (confirming the
interpretation given for Figure 2). Maps of
German import shares show stability over
time and positive association throughout
Europe, through the Balkans to the
oil-exporting states of the Persian Gulf.
Unlike the impression given by the
maps for the United States and Germany,
the clusters of Japanese export shares are
small and much more fragmented. In
1968, only a Southeast Asian cluster is
evident (Fig. 8), while in 1992 this is
extended by a pattern of positive association
in the Persian Gulf (Fig. 9). This
pattern was essentially in place in 1973, as
152
i,
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
TRADE BLOC FORMATION 153
Figure 5. Location of significant Gj statistics for
is also borne out by the interpretation of
Figure 3. Exports from Japan seem to be
more evenly spread around the globe than
those of its two global trade competitors,
and there is certainly no evidence of a
"yen bloc" in East Asia. Indeed, the
existence of such a bloc would be inimical
to Japan's long-term economic interests, a
fact not lost on Japanese policymakers. In
contrast, Japanese import shares show
strong signs of regional concentration. In
1968, they were heavily clustered in
South and East Asia, with no significant
high or low concentrations elsewhere. By
1992, there were three additional clusters
of positive association, including the oil
states of the Middle East, Central America,
and West Africa and a cluster of
low-import shares from former Eastern
Europe. The mid-1970s, when the OPEC
oil cartel was at its peak influence, marks
the consolidation of these clusters, demonstrating
the Japanese dependence on
resource imports emanating from a few
sources; these are critical to Japanese
economic prosperity (Friedman and
Lebow 1991).
The summary picture of this exercise in
exploratory spatial data analysis is that the
geographic evidence, the most important
for the support of the bloc-building
hypothesis, is weak and mixed. Global
indicators of spatial autocorrelation (the
Moran's I values) have not changed much
over time, except for Japan. The local
evidence of trade clustering provided by
the Moran scatterplot and the G,* statistics
supports hypotheses that the United
States and West Germany continue to
develop strong trading links with their
immediate neighbors, but these links
were well established before the mid-
1970s and have been essentially stable
since then. The Japanese experience is
TRADE BLOC FORMATION 153
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
ECONOMIC GEOGRAPHY
SkSh,
Figure 6. Location of significant Gj* statistics for West German export shares, 1968.
the reverse; Japanese exports are globalizing,
while imports are increasingly clustered.
The notion of a so-called yen bloc
seems premature.
Conclusions
Like rumors of Mark Twain's death, the
formation of blocs in the world trading
system has been greatly exaggerated. The
fears of free traders like Bhagwati (1991),
that the system is "at risk" because of the
formation of large regional blocs, are
unwarranted. There is some intensification
of long-established trade links between
the United States and its neighbors
in the Americas, as well as between West
Germany and its European Community
partners. Japan, on the other hand, is
showing the reverse trend, toward broadening
and deepening its export and
import linkages with extraregional partners
to complement its well-established
connections to its East Asian partners.
For the purposes of this paper, world
regions were equated with the panregions
identified by the German geopolitical
writers of the 1930s. In previous research,
blocs have been narrowly defined to
exclude areas far from the country of
interest. Interest in Germany's bloc building
was thus confined to GermanEuropean
Community trade relations;
German relations with the rest of the
world were ignored. As Anderson and
Norheim (1993a, 1993b) noted, a country
could be intensifying its trade relations
with its neighbors and with the rest of the
world simultaneously; a narrow regional
focus will likely miss the propensity to
export extraregionally. Within the large
panregions that we used, there is a great
deal of diversity in the intensity of trade
relations with the regional superpower.
154
.I,
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
TRADE BLOC FORMATION
Figure 7. Location of significant Gi* statistics for West German export shares, 1992.
For both the United States and West Germany,
the linkages to other countries are
highly varied and countries within their
"blocs" exhibit both very high and low trade
links with their large economic neighbor.
Japan, by contrast, is a global nonregionalizing
trader. Significant changes in the
GATT world, such as more management of
trade, raise the possibility of negative consequences
for the seemingly inexorable Japanese
trade expansion.
This exploratory spatial analysis indicates
that the geographic evidence, the
most important for the support of the
bloc-building hypothesis, is weak and
mixed. Global indicators of spatial association
provide evidence of clustering but
the Moran's I values have not changed
much over time and even diminished in
some years. The local evidence of trade
clustering in the G,* statistics supports
hypotheses that the United States and
West Germany continue to develop
strong trading links with their immediate
neighbors, but these links were established
before the mid-1970s; recently,
these clusters have stabilized rather than
spread. The Japanese experience is the
reverse: Japanese exports are globalizing
while imports are increasingly originating
in the East and Southeast Asian region.
The scatterplots showed a large amount of
deviation from the best-fit line, another
indicator of lack of regional clustering.
The picture that emerges in this paper
of the analysis of trade ties is one of
regional complexity and local circumstances.
These local disjunctures, so evident
in the scatterplots, suggest that a
more contextual and local-specific view of
trade is warranted. These local effects, so
often the result of cold war divisions
translated into the trading world, provide
further support for the emerging view in
155
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
ECONOMIC GEOGRAPHY
Figure 8. Location of significant G* statistics for Japan export shares, 1968.
political geography that the approaches of
international relations must be more
nuanced to account adequately for sizable
deviations from global trends (O'Loughlin
and Anselin 1991; Ward 1992). Or, stated
in the language of Most and Starr (1989),
"domain-specific" models need to be
considered, since general laws miss important
effects in the various contexts that
constitute the world system. There has
been, thus far, little attempt to define and
measure these "domains," which, in the
language of geographers, can be equated
with regional contexts. It seems certain on
the basis of this exploratory spatial analysis
that the contexts are much smaller
than the panregional, or large bloc, view
adopted in this kind of trade analysis.
GDP trends and political relations are
driving export patterns, while the "reassertion
of economic geography" in the
form of stronger ties to regional neighbors
is only beginning to be seen after the end
of the ideological divide between capitalism
and Communism in 1990. Dilemmas
arise as a state's involvement with the
international economy grows (Strange
1988). Though the interests of the state
and its exporting companies, especially
multinationals, are not always coincident,
a trend toward "managed trade" will
likely bring the political and economic
sectors closer together. It is in no
country's interests unilaterally to confine
its trade energies to one region. A better
strategy is to try to hold market share in
the region of dominance while expanding
the direction of trade to other, especially
growing, regions of the world. Regional
partners tend to chafe under the leadership
of a regional hegemon, making
regional trade and economic pacts hard to
solidify.
Trade is only one, though probably the
156
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
TRADE BLOC FORMATION
Figure 9. Location of significant Gi* statistics for Japan export shares, 1992.
most visible, of the many international
interactions defining the new geo-economic
world. However, unlike the visions
of the 1930s geopolitical writers, trade is
not controlled by state policy except at
the margins. Transnational corporations
are increasingly the engines of the world
economy. For most capitalist countries,
imports and exports flow unfettered by
government regulation and control. The
Uruguay Round agreement will further
solidify control of the international regime
at the expense of individual state preferences.
The geo-economic world is different
than the recent geopolitical world
because state control of important international
interactions is more circumscribed.
The battle lines between competing
states are not so neatly drawn, nor are
the interests of the exporters always the
same. State policy must balance the
demands of uncompetitive industries, like
textiles in the United States, against the
wishes of successful exporters, like aircraft
and grain interests in the United States,
for a global free trade regime. As Ruigrok
(1991, 83) notes, "national security and
preservation of internal order have been,
and will remain, more important government
concerns than maximizing effi-
ciency."
The shifting international division of
labor consequent on the changing patterns
of industry will reorganize the trade
patterns of the world economy. The past
quarter century has seen few shifts in the
export patterns of the major countries, as
this paper shows. Trade follows global
trends in output; regional integration,
though more developed than in the late
1960s, is not yet a global trend. Defensive
strategies to protect market shares and
national interests in the form of sponsorship
and support of regional trade pacts
157
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
ECONOMIC GEOGRAPHY
will come in conflict with the clamor for
open markets and access to expanding
economies by successful industries. State
policy must balance these interests. The
end product of a geo-economic world may
be a trading regime that is liberal in
rhetoric and law but illiberal in practice.
In such a world, real geo-economic
competition begins and regional trade
strategies loom ever more prominent.
References
Aho, C. 1993. America and the Pacific
Century: Trade conflict or cooperation.
International Affairs 69:19-38.
Anderson, K. 1991. Is an Asian-Pacific bloc
next? Journal of World Trade 25:27-40.
Anderson, K., and Norheim, H. 1993a. From
imperial to regional trade preferences: Its
effects on Europe's intra- and extra-regional
trade. Weltwirtschaftliches Archiv 129:78-
101.
. 1993b. Is world trade becoming more
regionalized? Review of International Economics
1:91-109.
Anselin, L. 1992. SpaceStat: A programfor the
statistical analysis of spatial data. Santa
Barbara, Calif.: National Center for Geographic
Information and Analysis.
. 1993a. Exploratory spatial data analysis
and geographic information systems.
Regional Research Institute Research Paper
9329. Morgantown: West Virginia Univer-
sity.
. 1993b. The Moran scatterplot as an
ESDA tool to assess local instability in
spatial association. Regional Research Institute
Research Paper 9330. Morgantown:
West Virginia University.
. 1995. Local indicators of spatial
association-LISA. Geographical Analysis
27:96-115.
Anselin, L.; Dodson, R; and Hudak, S. 1993.
Linking GIS and spatial data analysis in
practice. Geographical Systems 1:3-23.
Anselin, L., and Getis, A. 1992. Spatial
statistical analysis and geographical information
systems. Annals of Regional Science
26:19-33.
Anselin, L., and O'Loughlin, J. 1990. Once
more with space: An analysis of U.S trade
and conflict relations, 1966-88. Paper presented
at the Annual Meeting of the
American Political Science Association, 30
August-i September, San Francisco.
Bergsten, F. 1991. Commentary: The move
toward free trade zones. In Policy implications
of free trade zones, 43-57. Kansas
City, Mo.: Federal Reserve Bank of Kansas
City.
Bhagwati, J. 1991. The world trading system at
risk. Princeton, N.J.: Princeton University
Press.
. 1993. Beyond NAFTA: Clinton's trading
choices. Foreign Policy 91:155-62.
Cliff, A. D., and Ord, J. K. 1973. Spatial
autocorrelation. London: Pion.
. 1981. Spatial processes: Models and
applications. London: Pion.
Cohen, S. B. 1991. Global geo-political change
in the post-Cold War era. Annals of the
Association of American Geographers 81:
551-80.
Corbridge, S., and Agnew, J. 1991. The U.S.
trade and budget deficits in global perspective:
An essay in geopolitical economy.
Environment and Planning D: Society and
Space 9:71-90.
Corker, R. 1991. The changing nature of
Japanese trade. Finance and Development
28 (2):6-10.
Crook, C. 1990. Nothing to lose but its chains:
A survey of world trade. Economist, 22
September.
De Melo, J., and Panagariya, A. 1992. The new
regionalism in trade policy. Washington,
D.C.: World Bank.
Dhar, B. 1992. The decline of free trade and
United States policy today. Journal of World
Trade 26 (December):133-54.
Drache, D., and Gertler, M. 1991. The
world-economy and the nation-state: The
new international order. In The new era of
global competition, ed. D. Drache and M.
Gertler, 3-25. Montreal: McGill-Queen's
University Press.
Drysdale, P., and Garnaut, R. 1982. Trade
intensities and the analysis of bilateral trade
flows in a many country world: A survey.
Hitotsubashi Journal of Economics 22:62-
84.
Duesterberg, T. 1993. Global competitiveness
and United States-EC trade relations.
Washington Quarterly 16 (Summer):115-28.
Fieleke, N. 1992. One trading world or many?:
The issue of regional trading blocs. New
England Economic Review (May/June):
3-20.
Frankel, J. 1991. Is a Yen Bloc forming in
Pacific Asia? In Finance and the interna-
158
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
TRADE BLOC FORMATION
tional economy 5, ed. R. O'Brien, 4-21.
London: Oxford University Press.
Friedman, G., and Lebow, M. 1991. The
coming war with Japan. New York: St.
Martin's Press.
Fukuyama, F. 1992. The end of history and the
last man. New York: Free Press.
Garton, J. 1992. A cold peace: America, Japan,
Germany and the struggle for supremacy.
New York: Basic Books.
Gasiorowski, M. 1986. Economic interdependence
and international conflict: Some
cross-national evidence. International Studies
Quarterly 30:23-38.
GATT: The eleventh hour. 1993. Economist, 4
December, 23-26.
Getis, A., and Ord, J. K. 1992. The analysis of
spatial association by use of distance statistics.
Geographical Analysis 24:189-206.
Gilpin, R. 1992. The debate about the new
world economic order. In Japan's emerging
global role, ed. D. Unger and P. Blackburn,
21-36. Boulder, Colo.: Lynne Rienner Pub-
lishers.
Grant, R. 1993. Trading blocs or trading blows:
The macro-economic geography of U.S. and
Japanese trade policies. Environment and
Planning A 25:273-91.
Hanink, D. 1988. An extended Linder model
of international trade. Economic Geography
64:323-34.
. 1994. The international economy: A
geographical perspective. New York: John
Wiley.
Hanink, D., and Kier, V. 1993. Selected
characteristics of Soviet foreign trade and
implications for the foreign trade of the
Commonwealth of Independent States. Geographical
Analysis 25:234-41.
Hufbauer, G.; Berliner, D. T.; and Elliot, K. A.
1986. Trade protectionism in the United
States: 31 case studies. Washington, D.C.:
Institute of International Economics.
Inoguchi, T. 1988. Four Japanese scenarios for
the future. International Affairs 65:15-28.
International Monetary Fund. 1968-88. Direction
of trade statistics. Washington, D.C.:
IMF.
Jackson, J. 1993. Regional trading blocs and
the GATT. World Economy 16 (March):121-
58.
Krugman, P. 1991a. Geography and trade.
Cambridge: MIT Press.
. 1991b. Regional trade blocs: The good,
the bad and the ugly. International Economy
5 (6):54-58.
. 1991c. The movement toward free
trade zones. In Policy implications of free
trade zones, 7-42. Kansas City, Mo.: Federal
Reserve Bank of Kansas City.
. 1994. Peddling prosperity: Economic
sense and nonsense in the age of diminished
expectations. New York: W. W. Norton.
Krugman, P., and Lawrence, R. 1994. Trade,
jobs and wages. Scientific American 270
(4):44-49.
Lawrence, R. 1991. Emerging regional arrangements:
Building blocks or stumbling
blocks? In Finance and thi international
economy 5, ed. R. O'Brien, 22-35. New
York: Oxford University Press.
Leontief, W. 1956. Factor proportions and the
structure of American trade: Further theoretical
and empirical analysis. Review of
Economics and Statistics 38:386-402.
Linder, S. 1961. An essay on trade and
transformation. New York: John Wiley.
Lloyd, P. J. 1992. Regionalization and world
trade. OECD Economic Studies 18:7-43.
Low, P. 1993. Trading free: The GATT and US
trade policy. New York: Twentieth Century
Fund Press.
Luttwak, E. 1993. The endangered American
dream: How to stop the United States from
becoming a Third World country and how
to win the geo-economic struggle for industrial
supremacy. New York: Simon and
Schuster.
Maidment, P. 1989. The yen block. Economist,
15 July.
Meltzer, A. 1991. U.S. leadership and postwar
progress. In Policy implications offree trade
zones, 237-57. Kansas City, Mo.: Federal
Reserve Bank of Kansas City.
Most, B., and Starr, H. 1989. Inquiry, logic
and international politics. Columbia: University
of South Carolina Press.
Muraoka, T. 1991. Pursuing the new international
economic order: Overseas investment
and trade of Japan, the Asian NIEs and
ASEAN. In Japan and the global economy:
Issues and trends in the 1990s, ed. J. Morris,
154-71. New York: Routledge.
Nierop, T., and de Vos, S. 1989. Of shrinking
empires and changing roles: World trade
patterns in the postwar period. Tijdschrift
voor Economische en Sociale Geografie
79:343-64.
O'Brien, R. 1992. Global financial integration:
The end of geography. New York: Council
on Foreign Relations Press.
O'Loughlin, J. 1992. Ten scenarios for the
"New World Order." Professional Geographer
44:22-28.
159
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms
ECONOMIC GEOGRAPHY
. 1993a. Fact or fiction: The evidence
for the thesis of U.S. relative decline,
1966-1991. In The political geography of the
new world order, ed. C. Williams, 148-80.
London: Belhaven Press.
. 1993b. Geo-economic competition in
the Pacific Rim: The political geography of
U.S. and Japanese exports, 1966-1988.
Transactions of the Institute of British
Geographers 18:438-59.
O'Loughlin, J., and Anselin, L. 1991. Bringing
geography back to the study of international
relations: Spatial dependence and regional
context in Africa, 1966-1978. International
Interactions 17:29-61.
O'Loughlin, J., and van der Wusten, H. 1990.
The political geography of panregions. Geographical
Review 80:1-20.
6 Tuathail, G. 1993. Japan as threat: Geoeconomic
discourses on the USA-Japan
relationship in US civil society. In The
political geography of the new world order,
ed. C. Williams, 181-209. London: Belhaven
Press.
Pollins, B. 1989. Does trade still follow the
flag? American Political Science Review
83:465-80.
Preeg, E. 1992. U.S. leadership in world trade:
Past, present, future. Washington Quarterly
15 (Spring):81-91.
. 1993. The compatibility of regional
economic blocs and the GATT. Annals of the
American Academy of Political and Social
Science 526:164-71.
Reich, R. 1991. The work of nations. New
York: Alfred A. Knopf.
Rosecrance, R. 1986. The rise of the trading
state: Commerce and conquest in the modern
world. New York: Basic Books.
Ruigrok, W. 1991. Paradigm crisis in international
trade theory. Journal of World Trade
25:77-89.
Schott, J. 1991. Trading blocs and the world
trading system. World Economy 14:1-17.
Strange, S. 1988. States and markets. London:
Frances Pinter.
Summers, L. 1991. Regionalism and the world
trading system. In Policy implications offree
trade zones, 295-302. Kansas City, Mo.:
Federal Reserve Bank of Kansas City.
Taylor, P. 1990. Britain and the cold war: 1945
as geopolitical transition. London: Frances
Pinter.
. 1992. Nationalism, internationalism
and a "socialist geopolitics": A review essay.
Antipode 24:327-36.
. 1993. The last of the hegemons:
British impasse, American impasse, world
impasse. Southeastern Geographer 33:1-22.
Thurow, L. 1990. GATT is dead: Post World
War II economic world is over. Journal of
Accountancy 170 (3):36-39.
. 1992. Head to head: The coming
economic battle between Japan, Europe and
America. New York: William Morrow.
Trade winds shift, The. 1993. Economist, 20
November, 15-16.
Upton, G., and Fingleton, B. 1985. Spatial
data analysis by example. New York: John
Wiley.
Van Agt, D. 1991. United States-JapanEurope:
The unholy trinity. European Affairs
4:61-65.
Wallerstein, I. 1991. Geopolitics and geoculture:
Essays on the changing world-system.
New York: Cambridge University Press.
Ward, M., ed. 1992. The new geopolitics. New
York: Gordon and Breach.
Weintraub, S. 1991. Western Hemisphere free
trade: Probability or pipe-dream? Annals of
the American Academy of Political and
Social Science 526:9-24.
World Bank. 1992. Regionalism in trade is
back-and here to stay. Policy Research
Bulletin 3:2-7.
160
This content downloaded from
147.251.68.36 on Wed, 24 Feb 2021 13:06:39 UTC
All use subject to https://about.jstor.org/terms