No. 9 — March 3, 2000

Feature Article


Arthur J. Alexander

Issue Index aaaa Subscriber Area aaaa Publications aaaa Home


Events data for the 1948-97 period provide insight into U.S.-Japan relations in the economic, political and military spheres across those 50 years. As defined by scholars, an "event" is a discrete occurrence with a beginning and an end and an actor and a target that has sufficient prominence to stand out from the background of normal developments. An analysis of more than 4,000 such transpacific records indicates that military events were the most numerous in the 1950s, political activities characterized the 1960s and economic issues have shaped the U.S.-Japan relationship since the 1970s. In fact, in the 1980s, economic interactions came to dominate bilateral affairs. The expansion of two-way trade between the United States and Japan was the main force making economic events preeminent.

Regardless of the sphere, most U.S.-Japan events are routine, reflecting the many activities that require governmental attention in daily dealings between the world's two largest economies. Although the number of conflictual events has jumped, especially since the 1980s, the volume and the intensity of cooperative developments have increased even more.

A major finding of the analysis is that neither positive nor negative transpacific economic incidents have much effect on bilateral political or military relations. However, a higher level of political conflict, stimulated largely by random events, appears to produce more economic and military cooperation. One interpretation of this result is that diplomats react to political conflict by working harder to maintain harmony in other parts of the complex U.S.-Japan relationship.


Evolving Use Of Events Data

How are relations between countries best characterized? How, for example, can the following questions concerning dealings between the United States and Japan be answered?

  • Are U.S.-Japan relations strengthening, declining or suffering from trade frictions, or are they entering an era of cooperation on the security and diplomatic fronts?
  • Are bilateral political and military relations at the mercy of trade disputes, or are American economic interests sacrificed to advance military and political cooperation?

This report attempts to address such questions through the use of so-called events data on U.S.-Japan relations that span the 50 years from 1948 through 1997. Before proceeding, though, it may be useful to discuss the origins of events data and the problems associated with their use, especially since until now, they have not been applied to analyses of U.S.-Japan affairs.

The use of events data in the study of political science, international relations and economics began in the 1950s, as social scientists in universities and policy analysts working for government agencies and think tanks attempted to flesh out and apply theories coming from the research community. Case studies of particular events had been the main method for addressing these subjects. However, critics of this approach noted that cases usually were selected for their outstanding or unusual properties and, therefore, conclusions based on a limited number of cases were likely to be highly biased by the selection process itself. As political scientists sought better tests of their theories of intercountry relations, especially those concerning the sources of conflict and war, they turned to collections of comprehensive, unbiased data on international relations.

One type of information that appeared to satisfy research requirements was the "event." The concept appealed both to diplomats and to historians, who viewed international relations as a series of occurrences — negotiations, protests, agreements, crises, conferences and wars. As scholars tried to distill the data that described these acts into a manageable form, they defined an event as a distinct action with a start and a finish, an initiator and a quarry, and sufficient prominence to differentiate it from everyday happenings.1

The Conflict and Peace Data Bank, a collection used in this report, was compiled in the late 1960s by Edward Azar in a dissertation on conflict in the Middle East. Mr. Azar defined an event as an occurrence between nations that stands out enough from the constant flow of transactions to have been reported by a news source.

Building on his doctoral work, Mr. Azar expanded his data set from its original regional focus to global international events. To enlarge COPDAB, a small army of graduate students at the University of Michigan scanned newspapers and journals from around the world for evidence of intercountry events involving 135 nations and then assigned each event to one of nine subject areas — for example, politics or economics. The researchers also coded each event into one of 15 weighted cooperation or conflict categories. TheCOPDAB data bank incorporated more than 500,000 event records for 135 countries from 70 periodicals worldwide.

As political science and the study of international relations became more mathematical and data-oriented in the 1960s and the 1970s, the use of events data surged. A survey of 15 major U.S. international politics journals for the years 1974 through 1986 found events data employed in 61 articles.2 By the early 1980s, more than 40 sets of events data were stored at an international depository for political and social research.3

As the usage of events data grew, however, the basic concepts behind this technique attracted criticism from its creators and users as well as from skeptics. Charles McClelland, the originator of one widely used set of data, the World Event Interaction Survey, summarized many of the doubts raised by compilers and critics alike. Events based on public reporting of the news, he said, are likely to be biased by the selection and the suppression of stories by reporters, editors and media owners. Moreover, since much of the work of international relations takes place in secret, publicly reported events may not represent the real circumstances of world affairs. Even if the reporting is unbiased and complete, interpretation can be distorted by the partisan interests of reporters, editors and data set compilers. Mr. McClelland accepted the contention that "it is impossible to claim that reports of happenings in international relations are never fabricated, never manipulated or never concealed."4

Nonetheless, he went on to note that the gathering and the diffusion of world news have become highly organized and are taking place on an ever-increasing scale. He reasoned that reports appearing in major newspapers and carried by wire services constitute the first version of international history. In addition, participants in secret, closed-door negotiations often have reasons to leak information about the discussions; for instance, they may be motivated by attempts to influence outcomes or to establish their own places in history. Mr. McClelland cited the finding that even intelligence officers derive an estimated 80 percent to 90 percent of their reporting from unclassified sources. Although later revelations and historical scholarship may modify the initial story, "often the current selections, descriptions and interpretations survive as originally published, or they are modified only slightly to become the conventional 'truth' of history."5

Despite this positive commentary, the critics had made several valid points, and funding for and use of events data fell sharply in the 1980s. Work on many data sets was abandoned. COPDAB, for one, ceased to be updated after 1978. However, the growth of computer capabilities and the development of such massive on-line news sources as Lexis-Nexis led researchers to rethink their strategies and to explore the deployment of the new powers.

A group of international scholars in the early 1990s developed a coordinated research agenda that tried to address the critics while figuring out how to take advantage of the available computer and on-line facilities. The Data Development for International Research consortium received funding from the National Science Foundation to improve existing collections of high-quality events data and to design software to facilitate the generation of such information.6

The University of Maryland's Global Events Data System, a DDIR consortium participant, aimed to create a core data set that could be used by various researchers for their own particular requirements. Each record includes sufficient information to code an event and interpret it according to individual interests and questions. The GEDS events framework is based on the older COPDAB approach.

The GEDS project addressed one of the main criticisms of the older methods. Most previous data-gathering efforts used major newspapers like The New York Times as their main sources of information. It was recognized early on that editors' customary selection of each day's stories from a mountain of potential articles imparted a bias to the published news and, therefore, to a data base created from that record.

To get around this problem, GEDS used the Reuters World Service, which has news gatherers around the world, usually local reporters, who put stories on the wire for use by editors everywhere. Reuters is one of the most extensive and widely used news wires publicly available. Tapping this source to develop events data removes most of the bias that worried the earlier generation of scholars, although some residual bias may persist in terms of what local reporters find newsworthy.

The Japan Economic Institute acquired all of the 1,479 COPDAB records on U.S.-Japan relations from 1948 to 1978. Through funds provided by an Abe Fellowship grant from the Center for Global Partnership administered by the Social Science Research Council, it arranged for the GEDS project to extract the 2,838 events on U.S.-Japan relations occurring from 1979 through 1997. In all, more than 4,300 observations covering 50 years of postwar experience are available for analysis.7

Each COPDAB event was punched into an 80-column card to create a record that included date, actor and target, a verb indicating the action and a 45-character description of the event. As noted, the COPDAB teams coded events into one of nine areas and scored each according to its intensity; this gauge ranged from 1 for the most cooperative to 15 for the most conflictual (see Appendix). Events records compiled by the GEDS group provided the same information, but the number of fields and the descriptions were expanded. The text describing events averages 500 characters, with some summaries five times that length.


Examples Of Events Data

Some examples give a better feel for the nature of this methodology and how the information has changed over time. "Japan agrees with U.S. to lower import controls on some goods,"8 a COPDAB economic event from October 7, 1969, is assigned a cooperative value of 4. A conflictual economic event with a value of 10 occurred March 10, 1970: "U.S. blames Japanese competition for layoffs." Another 10-rated conflictual economic event with Japan as the actor and the United States as the target was dated July 5, 1969: "Japan prohibits U.S. purchase of Sony shares by foreign investor." A political event from June 4, 1972 scored a 5 on the cooperation scale: "U.S. visit to Japan by Kissinger; discuss U.S. relations, USSR, China."

GEDS event descriptions include much more information than was possible in the abbreviated COPDAB format, although the subject matter often has a haunting familiarity. A mildly cooperative event with a value of 7 occurred July 7, 1993:

U.S. President Bill Clinton called for a tearing down of trade barriers with Japan on July 7. Clinton issued a plea on Wednesday to the Japanese people to join in a "common cause" to tear down trade barriers that not only hurt American workers but drive up prices in Japan. "The persistent trade imbalance … has hurt the Japanese people, deprived you of the full benefits of a strong economy," he told students at Waseda University. "I would send this message to all of you and to the people beyond the walls here in this hall," Clinton said. "You have a common cause with the people of America — a common cause against outdated practices that undermine our relationship and diminish the quality of your lives." "You are entitled to no less, and it will be a part of your role as a great nation for the foreseeable future to have that sort of open relationship," he said. "We look to Japan to address its own economic agenda with equal vigor, including help with promoting global economic growth and removing both formal and informal barriers to the flow of goods, services and investment," he said.

A more critical economic event with a value of 10 took place later that year:

U.S. Trade Representative Mickey Kantor said he was ready to act if no progress was made in trade negotiations with Japan on October 22. The row comes over what the U.S. says are unfair practices in Japan's construction bidding system. If nothing is done, the U.S. says it would close part of its construction bidding market to Japan. Earlier in Tokyo, a construction ministry official said Japan was considering an open bidding system for public works projects in a last-minute bid to head off the U.S. sanctions. "We have not made significant progress," Kantor said in a telephone interview. "This has been a seven-year struggle and the November 1 deadline is real." "I would like to compile within this month basic plans to introduce an open, competitive bidding system on projects above a certain value, and to improve the bidding and contracting system of public works based on the principles of transparency and competitiveness," Japanese Prime Minister Hosokawa said.


Lingering Analytical Issues

Several issues concerning analytical methods arise in using the wealth of information contained in the 4,300-plus events records. Especially debatable are the issues of aggregation and scaling of the categories of cooperation and conflict. Some scholars insist that events records should be used "as is" — that is, without any aggregation — in order to understand the unfolding nature of a negotiation or a crisis.

That objective, however, is not the only one that occupies the minds of researchers. A study of U.S. and North Atlantic Treaty Organization relations with Serbia during the Bosnian conflict, for example, used a weekly window to examine the shifting balance of diplomatic behavior and its impact on results.9 Other studies use monthly or annual events totals. This report employs yearly aggregates, which have the advantage of identifying broader tendencies and influences on bilateral ties. Monthly data better illuminate cross-national moves as one country reacts directly and immediately to the behavior of the other.

Assessing levels of cooperation and conflict is a little more complicated. Mr. Azar's original 15-point scale was intended to provide an ordinal ranking, with a higher number representing more conflict but not necessarily indicating the intensity of an event. Some studies simply use the ratio of cooperative to total events, giving no further consideration to intensity. Others use the 15 COPDAB categories to calculate a weighted sum of events per time period.

Mr. Azar himself first pointed out that the ordinal rankings provided little sense of the importance of a category. Is the difference between a 13 and a 14, which involve acts of war, the same as that between a 9 and a 10, which span verbal actions or other nonviolent acts? He asked a panel of international relations experts and diplomats to assign a relative intensity value to each of his 15 categories. These values are shown in the Appendix description of the categories and are used in the statistical analysis below.


Trends In U.S.-Japan Relations

Transpacific events have grown in number and changed in character over the decades. Military activities were the most numerous type of event in the 1950s, a reflection of the establishment of a formal military alliance between the United States and Japan (see Figure 1).10 Political actions came to the fore in the 1960s, as Cold War diplomacy and activities in Vietnam and the People's Republic of China assumed greater importance. With the growth of U.S.-Japan trade and other economic interactions, economic events edged out political activities in the 1970s. This trend accelerated in the next two decades. Economics came to dominate and define affairs between Tokyo and Washington.

The annual volume of both economic and political events started to rise in the second half of the 1980s (see Figure 2). Furthermore, they moved up and down in tandem. In fact, the apparent relationship suggested in the Figure is not an apparition: the simple correlation between the annual numbers of economic and political events is a highly significant 0.76.

Of course, correlation does not indicate causality. However, it is suggestive that in the 1980s, the U.S. trade deficit with Japan rose to historic highs and imports from Japan had a harsh impact on several important American industries — important in both economic and domestic political terms. A search for explanations for the surge in government-to-government economic activities probably should begin with trade.

A review of U.S.-Japan relations that relies on press reports would seem to indicate that the United States is the initiator of most events, with Japan the target. Such a conclusion is accurate in terms of economic affairs, but Japan more frequently is the actor in the political and military fields. Nevertheless, since economics are so important (see Figure 3), the common perception is not entirely incorrect. For events where the identity of the actor is unambiguous, the United States figured in 2,200 cases compared with 1,925 for Japan.

Most bilateral events are routine, dictated by the sheer number of activities necessary to maintain relations between the world's two largest economies. Although the incidence of conflictual developments has been increasing, the frequency of cooperative events has grown even faster. Despite the overwhelming number of recurring transactions between the United States and Japan and the general tendency toward cooperation, not surprisingly, it is the occasional conflict that catches the eye of the news media. The Table shows a breakdown of events by decade, divided into cooperative, routine and conflictual. Neutral events are defined as those falling in cooperation-conflict categories seven to nine. Cooperation includes categories four to six and high cooperation one to three. Conflictual events are those in categories 10 to 12.11

U.S.-Japan Cooperation and Conflict, 1950-97


High Cooperation






























Untangling Relationships

As was suggested above, the broad course of bilateral relations may be affected by trade, international tensions or other major occurrences on the world scene. One question that has caught the attention of both scholars and diplomats is whether a nation's relations in one area — for example, trade — are independent of developments in other areas, possibly because specialized bureaucracies are single-minded in the pursuit of their objectives or because the attention of leaders is focused on one issue at a time. Alternatively, some experts speculate that a nation's leaders coordinate policies across subjects, making trade-offs when an advance in a preferred direction requires a setback elsewhere.

In one view, the United States has traded economic interests for political, military and diplomatic gains in its relationship with Japan. From this perspective, American interests are best served by stimulating exports while discouraging imports. Although such mercantilist ideas were discredited more than 200 years ago by Adam Smith, they still hold a certain popular appeal.

In any event, the question raised by the assertions is interesting in its own right — and it is answerable. Do domestic political interests and narrow economic considerations affect the conduct of economic affairs? Or, conversely, as many diplomats complain, does the rough-and-tumble of trade relations adversely influence the course of diplomacy?

What Influences Volume of Activity? - To get at the question of causality behind U.S.-Japan relations as revealed by the 50 years of events data, the information was aggregated annually. Statistical regression equations12 were formulated using variables thought to be likely determinants of the direction of bilateral affairs. These included:

  • Imports, exports, the trade balance and total trade between the United States and Japan.
  • American and Japanese military spending as an indicator of the saliency of cold wars and full-scale military conflicts.
  • The relative sizes of the real American and Japanese economies as a measure of the economic and political importance of Japan on the world scene.
  • U.S. unemployment as a possible index of domestic political sensitivity to trade.
  • The political party of the president.

The starting point of the analysis was a consideration of economic events initiated by the United States. Confining the study to those events in which Washington was the actor allowed a clearer delineation of the domestic factors that may have motivated relations. The first conclusion is that trade, indeed, consistently was a statistically significant influence on the number of events. A rise in total trade increased the number of economic events, higher exports decreased it and more imports raised the volume.13 The size of the trade deficit, however, was not as important as the separate trade flows.

To get an idea of the impact of these trade variables on the number of economic events, a 1 percentage point increase in imports from Japan as a share of the U.S. total (say, from the 1997 level of 12 percent to 13 percent) would have added only one event to the total of 47 U.S.-actor economic events in 1997. A 1-point rise in the Japan-bound share of U.S. exports would have reduced the number of events by a more substantial six. A gain in two-way trade with Japan of $10 billion was estimated to generate about three events. (The value of 1997 two-way trade with Japan was about $210 billion in 1990 dollars.) It is this last variable — the value of real total trade — that mainly was responsible for the jump in economic events over the 1980s.

One other variable was found to be especially key to estimating the number of events initiated by the United States: the total for the preceding year. Roughly 40 percent of a prior year's volume will continue into the next one regardless of other variables. Thus, a change, whatever its cause, has a tendency to continue. That finding really is not surprising since trade relations are ongoing affairs that do not end with the turning of a calendar page. If, for example, alleged dumping of steel by Japan is being discussed, the talks will extend over a period of months. One lesson from this equation is that the effects of random events on the pace of bilateral economic discussions will linger until time and the intervention of new influences diminish their impact.

To consider the question of whether economics influence political and military relations or vice versa, the problems created by simultaneity and estimation biases must be addressed. Consider, for example, a situation in which the number of economic events affects the number of political events, and politics, in turn, impact economics. A naive use of statistical regression techniques would produce nonsense. The intermingled effects cannot be separated unless something more about the relationships is known. Specifically, factors that work on one of the dependent variables but not the other must be found.

As noted, the number of economic events is swayed by trade, and political events likely are affected by strategic affairs and politics. It may be possible to determine the effect of broad changes in international relations on political interactions by looking at such indicators as military spending. Then, using the statistical methodology known as two-stage least squares with instrumental variables, the predicted value of political events could be used in the equation for economic events, thereby disentangling the simultaneity in the economic relationship. Independent variables like military spending (assumed to be unrelated to economic events) are called instruments or instrumental variables. A similar process could be used for political matters.

This approach can be applied to events data to show the relationships between and among economic, political and military occurrences (see Figure 4). Political activity had a slight and always statistically insignificant effect on the number of economic events, while economic actions had a small positive impact on political events. Interestingly, both economics and politics had strong positive impacts on military happenings.

As will be discussed below, a conflictual economic relationship induces more cooperation in the military sphere; put somewhat cynically, the Pentagon cleans up the mess left by the Office of the U.S. Trade Representative. Political activity has a positive effect on the number of military events, with behavior in these two domains tending to move together. Other variables influencing the number of military events include the level of American defense spending and the political party of the president. A Democrat in the White House, for example, tends to reduce the events total by about five annually.

Unraveling Cooperation and Conflict - Differing opinions about appropriate methodologies are apparent when the subject turns from the number of events to the level of cooperation and conflict in the relationship. One approach is to examine the average value of the original COPDAB 15-point scale across the spectrum of events.14 Another weights each event by the COPDAB scale and averages the weighted events over the relevant time period. A third method weights events by the intensity scale developed by Mr. Azar and calculates the net sum of cooperative and conflictual events. Finally, some scholars maintain that cooperation and conflict are two different kinds of activities and should not be combined in a single net measure. All methods were tried. The last one seemed to provide the most informative results.

Cooperative and conflictual events were aggregated separately, and weighted according to Mr. Azar's intensity scale. However, almost all U.S.-Japan events fell into a middle range of plus 27 to minus 29. Unlike the analysis of the number of events, which included only those where Washington was the actor, all events were part of this weighted cooperation-conflict analysis.

The annual weighted total of cooperative and conflictual economic events is shown in Figure 5. Several points jump out. The tally of weighted negative events stayed in the range of minus 200 to zero until the mid-1980s. In 1985, it increased sharply; however, cooperative events jumped at about the same time.

The other feature of Figure 5 is that the two curves appear to be mirror images of each other. When negative events peak, positive ones follow with a lag of about a year. In fact, the simple correlation between the two variables is minus 0.85. A straightforward explanation is that talks on a difficult trade issue usually begin on a negative note and stay that way until resolution — a positive outcome. Such a mirror-image relationship between cooperative and conflictual events does not appear in the political sphere.

Analyzing the cross-relationships among cooperative and conflictual economic, political and military events requires the same statistical techniques applied earlier. Economic conflict was strongly influenced by trends in exports and imports, with U.S. imports from Japan increasing conflict and exports to that country sharply diminishing it (see Figure 6). Imports reduce political cooperation. A Democratic president, on average, is associated with greater economic cooperation. The mirroring effect between economic cooperation and conflict is indicated by the minus signs on the arrows connecting the two. Holding other factors constant, a change in weighted cooperation induces about a 60 percent movement in weighted conflict and vice versa.

In unreported results from an examination of the average level of the cooperation-conflict variable, the cross-product of unemployment and imports from Japan strongly affected economic cooperation. That is, neither unemployment nor imports alone had much kick. However, imports tend to be strongest when the U.S. economy is surging and unemployment is low. But if rising imports accompany a weak economy, political calls for stronger action on the trade front can be expected to shift into high gear.

The statistical explanation of the interactions among the areas of bilateral relations suggests that political conflict is largely random.15 A look at conflictual political events lends some support to this interpretation. Events in political conflict categories 10 and 11 on the COPDAB scale included problems associated with the U.S.-Japan security treaty in the 1960s, Japanese unhappiness over delays in the return of Okinawa, American dissatisfaction with Japan's implementation of certain environmental treaties and U.S. government reactions to perceptions of closed Japanese markets. Regardless of their origins, these conflicts spilled over into all other areas.

Higher levels of political conflict were associated with increased economic conflict. An intriguing result is that political conflict appears to boost cooperation in the economic and military arenas (note the minus signs in Figure 6). This outcome can be rationalized by suggesting that negative occurrences in one field provoke positive spillover elsewhere, perhaps as a way of repairing deteriorating relations triggered by domestic and/or international events.

How well do the equations estimate weighted cooperation and conflict events? As noted, negative political events were largely random. However, events in the other spheres of U.S.-Japan relations were predicted quite well, with some 70 percent to 80 percent of the variance "explained" by the equations. Figure 7 plots the actual level of weighted economic conflict and the projected value based on the equation presented schematically in Figure 6. The equation tracks the main trends quite closely, except for conflicts in 1987 and 1995, which were somewhat underpredicted.



U.S.-Japan economic relations are driven largely by trade. These effects are to be expected since the sheer volume of activity will uncover many issues whose resolution requires government intervention. However, the link between trade and negative economic relations almost certainly is political in origin rather than the result of routine bureaucratic problem-solving. The sharp increase in economic events in the early years of the Clinton administration, for example, was part of the domestic political strategy deployed by U.S. Trade Representative Mickey Kantor to reelect the president.16

According to the average annual aggregates of the events data, economic relations between the United States and Japan had at most a weak influence on political and military ties. The spillover into other areas of national interest was minimal. The other general result was the impact, both positive and negative, of conflictual political relations on the other aspects of the bilateral relationship.

Politics swayed economic and military ties but were not affected by them.

These conclusions are fairly robust econometrically. One interpretation is that governments attempt to manage domestic and international events that, left unchecked, might dictate the nature of fundamental relations. Nevertheless, economic conflict appears to feed off political friction.

A recent study of trade protection came to conclusions roughly consistent with the results reported here. It found that for politically unorganized industries, protection tended to increase with import penetration. Both total employment in an industry and the unemployment rate boosted the likelihood of protection.17 These results support the notion that trade policy is driven to a large extent by domestic politics.

Another study used statistical methods to explain the actions of the Office of the U.S. Trade Representative in an attempt to determine the sources of American trade policy.18 It found that for the years from 1984 through 1993, a country's trade balance with the United States was the single most important variable related to the attention paid by USTR to that trading partner. Attention was measured by the number of pages devoted to the country in USTR's annual National Trade Estimate Report. However, that total had no relationship to measurable implicit or explicit trade barriers. A variable designating Japan was positive and significant, indicating that more moves were made against it than Japan's economic characteristics seemed to warrant. Additional analysis was undertaken of all formal, unilateral trade measures initiated by the U.S. government to achieve enhanced market access between 1984 and 1993. Again, a variable for Japan was strongly related to official U.S. actions beyond the prediction of the other variables.

COPDAB events data have been used to examine the volume of trade as a function of diplomatic relations, which reverses the direction of causality identified in this report.19 The finding that countries that have relatively cooperative interactions trade more with each other fits the U.S.-Japan pattern. In comparative terms, relations between Washington and Tokyo are quite close, and each country is among the other's largest trading partners.

Finally, research on the influence of international relations on trade and the reverse came to the conclusion that close political ties increased trade but that even military disputes failed to reduce it.20 Moreover, trade did not lower the probability of such disagreements.

The results from this review of six countries from 1907 to 1990 are consistent with the evidence from U.S.-Japan relations. Economic issues seem to be fairly well isolated from activities on the political, diplomatic and military fronts. Trade has its own rationale and domestic politics, which can intrude onto the world scene and make life more or less comfortable for governments. Nonetheless, trade has relatively little lasting effect on other facets of bilateral relationships.

The views expressed in this report are those of the author
and do not necessarily represent those of the Japan Economic Institute

Top aaaa Issue Index aaaa Subscriber Area aaaa Publications aaaa Home


1aa This description and history are taken from Richard Merritt, Robert Muncaster and Dina Zinnes (eds.), International Event-Data Development: Data Development for International Research, Phase II (Ann Arbor, Michigan: University of Michigan Press, 1993), p. 2. Return to Text

2aa Ibid., p. 16. Return to Text

3aa Jack Vincent, "WEIS vs. COPDAB: Correspondence Problems," International Studies Quarterly, XXVII, No. 2, 1983, p. 161. Return to Text

4aa Charles McClelland, "Let the User Beware," International Studies Quarterly," XXVII, No. 2, 1983, p. 170. Return to Text

5aa Ibid. Return to Text

6aa Merritt, Muncaster and Zinnes, op. cit., p. 1. Return to Text

7aa Reuters data for the first two months of 1979 were unavailable. An earlier report analyzed COPDAB-GEDS data through the first half of 1996. See Arthur Alexander, "U.S.-Japan Relations Since 1948: A Preliminary Analysis Of Events Data," JEI Report No. 32A, August 21, 1998. Return to Text

8aa This event has been lightly edited to make its telegraphic style somewhat more readable. Return to Text

9aa Joshua Epstein and Jon Pevehouse, "Reciprocity, Bullying and International Cooperation: Time-Series Analysis of the Bosnia Conflict," American Political Science Review, XCI, No. 3, September 1997. Return to Text

10aa To simplify the analysis, the eight issue types in the data were combined into four. Symbolic-rhetorical political events were combined with political actions to form a unified "political" category. Four issue types were combined into a single "other" category: cultural, scientific and education issues; physical environment, ecology and natural resource issues; minorities, human rights and health issues; and other issues. Most of the events in this combined category involve scientific and cultural cooperation. Return to Text

11aa The 18 events in the 1948-49 period are added to the 1950s and a single conflictual event with a value of 13 was added to the conflict column. Return to Text

12aa The estimated equations are available from the author. Return to Text

13aa The variables were defined as imports from Japan as a percentage of total U.S. imports, exports to Japan as a share of total exports and the sum of U.S.-Japan exports and imports in real (1990) dollars. Return to Text

14aa This method was used in Alexander, op. cit. Return to Text

15aa An indicator is the small value of the squared correlation coefficient of the best equation: only 0.095. Return to Text

16aa This conclusion is supported in greater detail in Alexander, op. cit. Return to Text

17aa Pinelopi Koujianou Goldberg and Giovanni Maggi, "Protection For Sale: An Empirical Investigation," American Economic Review, XCIII, No. 5, December 1999, p. 1136. Return to Text

18aa Marcus Noland, "Chasing Phantoms: The Political Economy of USTR," International Organization, LI, No. 3, Summer 1997, p. 365. Return to Text

19aa Brian Pollins, "Does Trade Still Follow the Flag?" American Political Science Review, LXXXIII, No. 2, June 1989, pp. 475-476. Return to Text

20aa James Morrow, Randolph Siverson and Tressa Tabares, "The Political Determinants of International Trade: The Major Powers, 1907-90," American Political Science Review, XCII, No. 3, September 1998, p. 649. Return to Text

The views expressed in this report are those of the author
and do not necessarily represent those of the Japan Economic Institute

Top aaaa Issue Index aaaa Subscriber Area aaaa Publications aaaa Home