The new politics of voting alignments in the United Nations General Assembly.

The end of the cold war saw the dismantling of the bipolar system that had characterized international politics for more than forty years. The ensuing changes reflect a new era in international politics and raise questions about which issues are emerging with new importance and about how the preferences of countries are aligned on those issues. The United Nations (UN) General Assembly, as a forum in which international politics is played out, can provide insight into the content and texture of these changes. The General Assembly offers a unique context in which to study post-cold war international politics, providing a great deal of information about the issues most salient to its member states and about their preferences. It also provides information and insights that aid an understanding of the political implications of change in the UN Security Council.

This article uses voting alignments on issues that have figured in General Assembly debates to identify and analyze preferences among member states. Our primary objective is descriptive inference: that is, to illustrate the unobserved underlying issue-dimensions of General Assembly roll-call votes and the alignments inherent in these votes by analyzing General Assembly voting records.(1) The analysis covers resolutions put to a roll-call vote during the forty-sixth through the forty-eighth sessions (1991 through 1993). We delineate the issue-dimensions that drive the activities of the General Assembly and analyze influences on the voting patterns of member states. These patterns have changed greatly since the end of the cold war. The East-West division no longer prevails in General Assembly deliberations; a North-South cleavage has superseded cold war alignments, giving rise to state preferences defined along developmental lines.

Earlier patterns

A study by Hayward Alker marks an early stage in a body of research that has sought to identify issue-dimensions and voting alignments in the UN General Assembly since its establishment in 1945.(2) The studies as a group span the entire history of the UN and, through applications of different statistical techniques, largely confirm one another's findings, especially in the delineation of voting blocs. The literature is thus consistent in its substantive results across different methods, inspiring confidence in studies that illustrate the evolution of state behavior in the UN.

Alker found two major behavioral voting alignments, East-West and North-South, in the Sixteenth General Assembly (1961). In an alternative interpretation, which lays the basis for our presentation here, he identified four main substantive issue-dimensions underlying roll-call votes, which he labeled self-determination, cold war and related membership questions, UN supranationalism, and Muslim questions.(3) Together these four issues accounted for 85 percent of the variance in General Assembly roll-call votes.

An expanded study of four different General Assembly sessions - 1947, 1952, 1957, and 1961 - by Alker and Bruce Russett found three of these substantive issue-dimensions - cold war, colonialism, and supranationalism - to occur repeatedly.(4) The cold war dimension separated the voting groups into Communists versus the West. The colonial dimension divided the colonial powers from the former colonies. Finally, supranationalism distinguished those that favored a stronger UN from those that did not. These three issue-dimensions, or "superissues," accounted for more than half of the variance in roll-call voting in each of the four General Assembly sessions. Two other superissues, albeit with less explanatory power, included intervention in southern Africa and matters concerning Palestine. These five factors together regularly accounted for 59 to 70 percent of all the variance in roll-call voting in each session. A study of the eighteenth session by Russett identified the same five factors, which accounted for two-thirds of the total variance and in which the cold war issue was most prominent. He also identified voting groups. For the eighteenth session, he found five major voting blocs, in addition to two small groups and several marginal countries, and suggested that a simplified East-West-neutral categorization could be misleading. In another study, Russett found four major voting blocs in 1952 and four major and four smaller groups in 1957.(5) Later studies of General Assembly voting behavior generally confirmed these findings.(6)

More recently, Steven Holloway analyzed five General Assembly sessions - 1946, 1955, 1965, 1975, and 1985.(7) He found that by 1985 three large voting blocs had formed. The most cohesive was the Warsaw Pact group, which also included a small group of radical Third World states including Cuba, Afghanistan, Vietnam, and Syria. The second large voting group was the Nonaligned Movement (NAM) bloc, which appeared to have retained its cohesion through 1985, ostensibly through the organizational efforts of the NAM and the Group of 77.(8) The third bloc, consisting of Western or Organization for Economic Cooperation and Development (OECD) countries, was located far from the first two groups and spread over a large space, reflecting the wide spectrum of left-to-right differences among its members. Holloway found these groups to be stable over time, especially between 1975 and 1985. In 1985 Poland appeared as a point separate from the former Soviet Union for the first time, reflecting the preglasnost configuration presaging the loosening of the Soviet bloc.

Previous studies of General Assembly voting records have thus focused on voting alignments, which delineate the major voting groups to be found among member states, and issue-dimensions, which characterize the clusters of resolutions put to a roll-call vote that in turn give rise to distinct voting alignments. The most useful studies identify voting alignments on particular issues or issue-dimensions. Since voting alignments vary greatly over different issue-dimensions, studies that identify voting alignments regardless of the issues that give rise to them often obscure important information. Thus, we shall follow a procedure that preserves the two kinds of information, first identifying the major issue-dimensions and then showing the voting alignments on each dimension.

Data, measurement, and methods

Our primary data are derived from voting records on those resolutions put to a roll-call vote in the forty-sixth through the forty-eighth sessions of the UN General Assembly, spanning the sessions that began in each of the years 1991-93. The analysis examines a total of 131 resolutions for 150 countries. Votes were ranked so as to reflect as closely as possible the distribution of voting preferences for any particular resolution.(9) They were then standardized so as to assign a zero - the truly middle position - to those countries that were absent or not participating at the time of these roll-call votes. Since it would be difficult to justify the assignment of a neutral position to countries that are absent from a large fraction of the roll-call votes, we exclude countries that were absent or not participating for more than 30 percent of the total number of votes.

Factor analysis reduces a set of variables, based on their correlations, to a smaller number of variables, known as "factors," which are the theoretical constructs that are latent to the variables and of which the variables are indicators. This technique is thus appropriate for determining which resolutions produce similar voting patterns and how many groupings of such similarities can account for countries' patterns of roll-call votes. With the variables reduced to their underlying dimensions, it then becomes possible to offer an interpretation of what these issue-dimensions are and to identify the major voting groups with respect to them. Furthermore, since the voting …