2.1 Senate Voting on the League of Nations, 1919–1920

The “League fight” was arguably the most extensive public debate about the country's role in the world that the United States has ever had.Footnote ³¹ As John Milton Cooper put it, “Democrats and Republicans alike believed they were contending for the soul of American foreign policy.”Footnote ³² During the course of the long debate over the treaty, there were more than 160 roll call votes and seventy-two of the ninety-six senators gave at least one speech on the issue. The salience of this debate made opposition to multilateralism into an article of faith among many Republicans for the remainder of the interwar period. Every Republican Party platform from 1920 through 1936 included language explicitly objecting to U.S. membership in the League of Nations.Footnote ³³

Pre-1914 foreign policy commitments fueled Republican opposition to the League of Nations. Senator Henry Cabot Lodge (R-MA), the Republican Majority Leader as well as the chair of the Foreign Relations Committee, had been a central figure in developing Republican foreign policy during the McKinley and Roosevelt administrations. Lodge's objections to the League centered on its potential to restrict the freedom of action and claims of privilege on which American foreign policy had rested under previous Republican administrations. The Foreign Relations Committee distilled these concerns into a set of formal reservations about the resolution of ratification. All of these “Lodge reservations” sought to rescind or limit multilateral commitments entailed in joining the League.Footnote ³⁴ For instance, one reservation forbade the League from objecting to American actions taken under the aegis of the Monroe Doctrine or from questioning U.S. interpretation of the doctrine. Another prohibited the League from taking action on issues such as trade and immigration that would have domestic repercussions in the United States. Reservations like these prefigured later conservative Republican objections to multilateral commitments. The Senate considered each of them in November 1919, prior to the final rejection of the treaty in March 1920, taking roll call votes on all but one.

According to Lodge's memoir, Senator William Borah, one of the leaders of the “irreconcilable” faction committed to blocking the treaty regardless of the reservations attached to it, informed him that he intended to vote for all of the reservations, then against the final treaty. Borah reasoned that the reservations would improve the treaty if it passed.Footnote ³⁵ We expect that other senators who opposed the treaty would follow Borah's example, so we take support for each reservation as evidence of skepticism about the multilateral commitments embodied in the League Covenant.

Figure 1 shows the partisan and ideological character of the supporters and opponents of the Lodge reservations. It reports predicted probabilities from a logit model that includes party identification, ideology, and a dummy variable for each roll call taken on these reservations during the November 1919 debate.Footnote ³⁶ As in most studies of roll call voting, we use the first dimension of the DW-NOMINATE score to indicate liberal-conservative ideology.Footnote ³⁷ It ranges from −1 (most liberal) to 1 (most conservative). The mean Republican and Democratic senators in Figure 1 reflect the central tendency of the party. The DW-NOMINATE scores for the conservative Republican and liberal Democrat in Figure 1 are set to 0.5 and −0.5, respectively, and will be used as points of comparison with debates in other Congresses. As Figure 1 indicates, there was a stark party division on the Lodge reservations, with little intraparty ideological difference. Republican support for the reservations exceeded 90 percent on twelve of the fourteen votes and never dropped below 50 percent. Democratic support never rose above 22 percent and was below 15 percent on twelve of the fourteen votes.

Fig. 1. Partisan-Ideological Differences in Support for the Lodge Reservations, 1919.

Notes. Predicted probabilities were computed from a logit model including party affiliation, the first dimension of the DW-NOMINATE score, and a dummy variable for each roll call. The predicted probabilities were computed for the closest roll call vote taken during the debate. The whiskers indicate 95 percent confidence intervals adjusted for clustering on the individual member. The Appendix describes each roll call and provides full model results.

Voting on the Lodge reservations offers another reason for questioning ideational accounts of the constraining effect of ideology. The political lineup before the New Deal is inconsistent with the claim that what would later be called “conservative” positions on core economic issues implied opposition to multilateral rules. On the eve of World War I, Republicans were divided between party regulars and insurgent Progressives, who had supported Theodore Roosevelt's third-party campaign for president in 1912. The Progressives took a range of leftist positions on domestic economic policy issues, such as labor rights, antitrust regulation, and consumer protection, that have led some historians to identify them as the antecedents of modern liberals.Footnote ³⁸ David Brady and David Epstein identify twelve progressive Republicans in the Senate during the 66th Congress, when the League fight took place.Footnote ³⁹ These senators had a mean DW-NOMINATE score of 0.22, making them substantially more liberal than party regulars, whose mean was 0.49. Both groups were more conservative than any Democrat, whose DW-NOMINATE scores ranged from −0.03 to −0.48, with a mean of −0.28. Further underscoring the anachronism of the term during this period, the fifteen Democrats with the most liberal DW-NOMINATE scores were Southerners whose views on many issues, particularly race, were far from liberal.

If the issue positions we now consider liberal and conservative logically implied positions on multilateralism, then Progressives should have been less likely to support the Lodge reservations than other Republicans were. In fact, the two factions were indistinguishable on these roll call votes. Models like those used to produce Figure 1 find no statistically significant differences between them.Footnote ⁴⁰ Indeed, some leading Progressives like William Borah (R-ID) and Hiram Johnson (R-CA) were “irreconcilables,” who would have opposed the League even if all the reservations had been adopted.

2.2 Senate Voting on Adherence to International Courts, 1923–2002

Senate debates on the jurisdiction of international courts offer an opportunity to observe partisan and ideological positions on multilateral rules over a longer period of time. Beginning in the 1920s, the Senate considered whether the country should accept the jurisdiction of, first, the Permanent Court of International Justice, established in the Versailles Treaty; second, the International Court of Justice, established in the UN Charter; and finally, the International Criminal Court, set up by the Rome Statute of 1998. Although there were substantial differences among these courts and in the circumstances surrounding the debates, acceptance or rejection of their jurisdiction always bore on the broader question of whether the United States should accept multilateral rules. All of the debates considered here focused on this issue.Footnote ⁴¹

Figure 2 depicts the partisan and ideological character of the courts’ supporters and opponents in each debate. It is based on a series of logit models identical to those we used to produce Figure 1. In spite of the enormous domestic and international political changes that occurred between 1923 and 2002, as well as the important differences in the courts under consideration, Republicans were always more likely to oppose international courts’ jurisdiction than Democrats were. In 1935, 1946, and 1985, there were also statistically significant intraparty ideological differences, with conservatives being more likely to oppose the courts’ jurisdiction. These intraparty differences are not apparent in 1994 and 2002 because the Republican Party had become almost monolithically conservative by this time. Indeed, the mean DW-NOMINATE score among Republicans in 2002 was actually 0.56, slightly to the right of the hypothetical conservative we used as a point of comparison in the other debates (DW-NOMINATE = 0.5).

Fig. 2. Partisan-Ideological Differences in Opposition to International Courts, 1923–2002.

Notes. Predicted probabilities were computed from logit models of each debate that include party affiliation, the first dimension of DW-NOMINATE, and a dummy variable for each roll call. The predicted probabilities were computed for the closest roll call vote taken in the debate. The whiskers indicate 95 percent confidence intervals adjusted for clustering on the individual member. The Appendix describes each roll call and provides full model results.

2.3 Senate Voting on the Bricker Amendment, 1954

While Senate debates over international courts show conservative Republican opposition to multilateralism at several points over a long period of time, the fact that there was only one vote on the issue during the entire Cold War era poses a problem. Conservative Republicans were arguably most likely to have abandoned their skepticism of multilateralism during this period. Multilateral institutions played a critical role in cementing the alliance system that the United States used to contain the Soviet Union. The absence of debate on this topic during the Cold War might indicate that Republicans abandoned their opposition to multilateralism after World War II and that its recent resurgence reflects current conditions rather than a continuity from the first half of the twentieth century.

Our third source of data helps address this possibility. It arises from the 1954 debate over a proposed constitutional amendment to limit the legal force of international agreements and the president's power to negotiate them. The amendment, sponsored by Senator John Bricker (R-OH), focused on many of the same issues that concerned opponents of the League of Nations and the international courts, particularly the prospect that these organizations could have jurisdiction over domestic legal matters within the United States. The amendment affected more than just multilateralism, but the issue played a major role in the debate. The nature and timing of the amendment indicate the robustness of conservative Republican opposition to multilateralism. Among other things, the amendment would have required enabling legislation before the provisions of any international agreement could have the force of law. Perhaps most radically, it would have prohibited the president from pursuing executive agreements without Senate ratification. Floor votes on the Bricker amendment took place at the height of the Cold War under a Republican president who strongly opposed the measure. Indeed, at the time Bricker introduced his amendment, the United States was still fighting in Korea under UN auspices.Footnote ⁴² These conditions should arguably have minimized conservative Republican support for the measure. Nevertheless, the proposed amendment enjoyed the backing of sixty-three cosponsors, and the final version missed the necessary two-thirds threshold by just one vote.

Senate consideration of the amendment involved seven roll call votes. Figure 3 summarizes the results of a logit model of these votes identical to those used for the votes on the Lodge reservations and international courts. It treats agreement with the position of Senator Bricker on these votes as a function of party and ideology. The pattern here resembles what we found for the court votes between 1935 and 1985, with conservatives and Republicans tending to support the measure. Conservative Republicans almost unanimously supported the Bricker amendment, in spite of the Cold War and President Eisenhower's objections. Although some moderate Republicans defected, most also backed the measure. Democrats were far less supportive. The measure's narrow defeat came about only through their opposition.

Fig. 3. Partisan-Ideological Differences in Support for the Bricker Amendment, 1954.

Notes. Predicted probabilities were computed from a logit model including party affiliation, the first dimension of DW-NOMINATE, and a dummy variable for each roll call. The predicted probabilities were computed for the closest roll call vote taken during the debate. The whiskers indicate 95 percent confidence intervals adjusted for clustering on the member. The Appendix describes each roll call and provides full model results.

Support for the Bricker amendment is sometimes attributed to segregationist opposition to multilateral human rights agreements then under consideration.Footnote ⁴³ This claim is certainly correct for the Southern Democrats who supported the measure, but conservative Republicans who had other motives devised the amendment and provided the bulk of legislative support for it. Only half of the senators from the eleven former Confederate states chose to cosponsor the amendment when it was introduced in January 1953. By contrast, forty-three of forty-eight Republicans did so. Many of the amendment's Republican supporters, including Bricker himself, would later vote in favor of the Civil Rights Act of 1957. Their opposition to multilateralism long predated the civil rights struggles of the 1950s and 1960s.

In spite of the amendment's near success and the overwhelming Republican support for it, subsequent observers have usually treated it as an aberration. The conventional view is that it was “the last hurrah of conservative isolationism.”Footnote ⁴⁴ In fact, it set the stage for the persistent American refusal to ratify multilateral human rights agreements.Footnote ⁴⁵ Conservative Republicans may have failed to seize control of American foreign policy, but they nevertheless got some of what they wanted. Moreover, they did not change their position. The Bricker amendment never again reached the floor of either the Senate or the House, but support for it persisted in conservative political circles. Various versions of the Bricker amendment were reintroduced sixty-five times between its February 1954 defeat and the end of the 115th Congress in 2018. These quixotic efforts have become less frequent over time but still recur. Most recently, Rep. John Culberson (R-TX) introduced such a measure at the beginning of 111th through 114th Congresses (2011–15).Footnote ⁴⁶

2.4 Sponsorship of Anti-UN Measures in the House and Senate, 1973–2018

Floor debates are informative but episodic. They do not reveal whether conservative Republican opposition to multilateralism was continuous over time or arose only in a few isolated instances. The sponsorship data help address this problem. Sponsorship and cosponsorship decisions tap opposition to multilateralism that failed to get past agenda setting by the congressional leadership. For much of the postwar era, efforts to undermine the United Nations were the province of a persistent minority in Congress. Few of these proposals ever reached the floor. Nevertheless, members were free to introduce them, either alone or with the support of cosponsors. That fact that many of them did so on a regular basis gives us a window onto conservative opposition to multilateral rules that would otherwise be difficult to observe. Data on bill sponsorship are available from the 93rd Congress (1973–74) through the present—a long period that overlaps the end of the Cold War.

Like opponents of the League of Nations, critics of the United Nations have expressed concern that it might limit American freedom of action, empower critics and enemies of the United States, and perhaps even infringe on the country's domestic sovereignty. Suspicion of the UN has been a staple of right-wing rhetoric since the organization's founding. D. J. Mulloy notes that Alger Hiss's role in setting up the UN helped fuel right-wing opposition, “but at a deeper level it was really about the fear that by joining such ‘international monstrosities’ as the UN, the World Health Organization, UNESCO, and NATO, the United States was willingly circumscribing its ability to project its enormous power onto the world stage for its own motives and in furtherance of its own interests—that it was yet another step away from the deserved spoils of victory.”Footnote ⁴⁷

Using the congress.gov database provided by the Library of Congress, we gathered data from the 1973–2018 period on bills that would have (1) withdrawn the United States from the UN, (2) reduced or eliminated American financial support for the UN, or (3) limited UN authority in other ways. We excluded omnibus measures in which actions against the United Nations were only one of many provisions, as well as measures that criticized UN actions without proposing to undermine the organization itself. We then identified the sponsors and cosponsors of each of the 292 measures we identified using the data gathered by James Fowler and his colleagues.Footnote ⁴⁸ Most of the 3,402 cosponsorships in the House and 578 cosponsorships in the Senate proposed cutting funds to the UN, but 131 of the House sponsorships were to measures that would have entirely withdrawn the United States from the organization. The Appendix provides a list of these bills and more information about how we identified them.

Figure 4 summarizes the results of a count model of the number of anti-UN bills each member sponsored or cosponsored in each Congress. It compares four hypothetical members identical to those we used to examine the roll call votes, showing the predicted probability that each would sponsor at least one anti-UN bill in a given Congress. The same pattern once again emerges. Conservative Republicans were substantially more likely to sponsor these measures than other Republicans, and many times more likely to do so than almost any Democrat. The patterns are quite similar in the House and Senate. It is worth noting that sponsoring bills to undermine the United Nations was not a rare event among conservative Republicans over the last five decades.

Fig. 4. Partisan-Ideological Orientation and the Cosponsorship of Anti-UN Bills, 1973–2018.

Notes. Predicted probabilities were computed from a negative binomial model that included party and ideology. The whiskers indicate 95 percent confidence intervals adjusted for clustering on the individual member. Full results are reported in the Appendix.

2.5 House Votes on the WTO and NAFTA, 1993–1994

One possible objection to the evidence presented thus far is that Republican opposition to international courts and the United Nations might be anomalous. Conservatives could have objected to these institutions for reasons other than opposition to multilateral rules in general. If this is the case, then conservatives should not object when multilateral rules advance a goal they support.

The votes on NAFTA and the Uruguay Round of the GATT offer a way to test this possibility. The House approved both agreements during the 103rd Congress, on November 17, 1993, and November, 29, 1994, respectively. Most of the same members thus voted on both measures. At the time, conservatives generally favored trade liberalization, while liberals were more likely to oppose it. Both agreements lowered trade barriers, but the Uruguay Round agreement also established the WTO, a broad multilateral organization with a quasi-judicial dispute resolution mechanism. The NAFTA agreement also contained a dispute resolution mechanism, but it was not broadly multilateral. Indeed, it was essentially two bilateral trade agreements. As such, NAFTA was less open to the objection that it could erode American sovereignty or constrain American freedom of action. Thus, despite a generally similar set of legislators, we should expect to see differences in voting behavior due to the scope of the multilateral rules in the two agreements.

Figure 5 depicts opposition to NAFTA and the WTO by party and ideology using the House votes on final passage of these two agreements. The models used to produce the chart are similar to those we estimated on other roll call votes, including only party and ideology. In this case, however, we also included a squared term for ideology, allowing it to have a nonlinear relationship to the way House members voted on these agreements.Footnote ⁴⁹ Ideological voting patterns differed on the two agreements in ways consistent with conservative concerns about the multilateral rules embodied in the WTO. Conservative Republicans were far more supportive of the NAFTA agreement than Democrats were. By contrast, these same conservatives were nearly as likely to object to the GATT agreement setting up the WTO as were liberal Democrats. Support for the WTO was found mainly among relatively moderate members of both parties.

Fig. 5. Opposition to the Creation of the WTO and NAFTA, 1993–1994.

Notes. Predicted probabilities were computed from a logit model including party affiliation, the first dimension of DW-NOMINATE, and the square of the DW-NOMINATE score. The whiskers indicate 95 percent confidence intervals. Both roll calls took place during the 103rd Congress. Numerical results are reported in the Appendix.

The floor debate confirms the reason for conservative Republican reluctance to support the WTO. For example, Rep. Howard Coble (R-NC) noted that while he had voted in favor of NAFTA, he “had not yet attained a similar comfort zone regarding the passage of GATT” in part because he worried it might erode American sovereignty. Rep. Duncan Hunter (R-CA) objected to the loss of “bilateral leverage” under the WTO. “Well, we are going to lose all of that in this World Trade Organization because now we are going to give this power away to a committee.” Comparing the WTO to the UN, he raised the specter of the United States being outvoted in the WTO by small countries that could be susceptible to bribery by the Japanese or other American trade competitors.Footnote ⁵⁰ Coble and Hunter were not alone among conservatives in raising these objections. In spite of their general support for trade liberalization evident on the NAFTA vote, conservative Republicans remained suspicious of multilateral rules and organizations.

In sum, there is overwhelming evidence that conservative Republicans have been the most consistent critics of multilateralism since the early twentieth century. This position followed naturally from the policy positions the party staked out when it controlled the White House between 1897 and 1913. It hardened into party orthodoxy during the debate over the League of Nations in 1919–20. Debates over adherence to the jurisdiction of international courts show the persistence of this pattern over time, though with some intraparty division during the mid-twentieth century. Voting on the Bricker amendment shows that it held even under a Republican president at the height of the Cold War. Sponsorship of anti-UN bills since 1973 shows that the episodic character of the roll call votes was not responsible for this pattern but instead that it held up consistently over time. Finally, the votes on the NAFTA and GATT agreements in 1993 and 1994 show that conservative objections were to the principle of multilateral rules rather than to the substantive issues at stake. Conservative Republicans objected to multilateralism even when it advanced policy goals they supported.

3.1 The Cold War

Before World War I, and perhaps even during the interwar period, Republican unilateralism made sense in ways it did not after World War II. It dovetailed with other aspects of American foreign policy. American protectionism and the country's efforts to build and enforce a privileged position in the Western Hemisphere made the acceptance of multilateral rules problematic. This project brought the United States into competition with other major powers, making cooperation with them more difficult. By contrast, American security during and after World War II depended on multilateral cooperation with other developed states. These circumstances should have diminished conservative Republican opposition to multilateral rules.

There are two mechanisms through which the increased importance of multilateralism during the Cold War might have influenced conservative Republicans. First and most obviously, the reliance of American Cold War strategy on multilateral institutions might have directly led them to reconsider their position. After all, these institutions were a means to fight international communism, a goal that they strongly supported. Second, Republican presidential leadership might have reduced conservative opposition to multilateralism. Regardless of their party affiliation, presidents were directly responsible for managing American foreign policy, a task for which multilateral institutions were useful. Republican presidents were arguably in a position to persuade members of their own party to temper or abandon their hostility to multilateralism.

Because the data on sponsorship of anti-UN bills provide continuous coverage through several Republican and Democratic presidents and overlap the end of the Cold War, they allow us to test these two mechanisms. We focus on the House of Representatives here because the larger number of House members provides more explanatory leverage. Figure 6 shows the number of representatives who sponsored or cosponsored at least one anti-UN bill, as well as the total number of sponsorships and cosponsorships, in each Congress from the 93rd (1973–74) through the 115th (2017–18). It suggests that members of Congress were indeed less likely to sponsor anti-UN bills before the end of the Cold War. The Berlin Wall came down during the 101st Congress, and the Soviet Union dissolved during the 102nd.

Fig. 6. Cosponsorship of Anti-UN Bills in House of Representatives, 1973–2018.

Figure 7 graphically displays the results of count models using interaction terms to test both mechanisms. It depicts the probability that a very conservative Republican would sponsor one or more anti-UN bills in a given Congress under different conditions. As expected, the Cold War had a substantial impact. Very conservative Republicans were nearly twice as likely to sponsor at least one anti-UN bill per Congress after it ended. Although there is no way to be certain that the end of the Cold War, rather than other historical changes happening around the same time, is responsible for this effect, the evidence is consistent with that claim. Republican presidents also made a difference, especially when they were relatively sympathetic to multilateralism. The most multilateralist Republicans in our sample—Richard Nixon, Gerald Ford, and George H. W. Bush—had substantially larger effects on Republicans in Congress, reducing the probability of sponsoring at least one anti-UN bill to 0.41, compared to 0.59 under Reagan, Trump, and the younger Bush. Under Democratic presidents, this probability rose to 0.71. International conditions thus made a difference, but they did not entirely erode Republican skepticism of multilateral rules.

Fig. 7. Political Circumstances and Republican Cosponsorship of Anti-UN Bills, 1973–2018.

Notes. Predicted probabilities were computed for a very conservative Republican (DW-NOMINATE = 0.75). The whiskers indicate 95 percent confidence intervals. Numerical model results are reported in the Appendix.

Our finding that the end of the Cold War was associated with an upsurge in Republican opposition to multilateralism is not new. Previous research on bipartisanship in foreign policy has advanced much the same argument.Footnote ⁵¹ However, these writers tend to overstate the extent of the consensus in support of multilateralism and other elements of the mainstream foreign policy consensus while the Cold War was going on. Opposition to multilateralism remained strong within the conservative faction of the Republican Party even during the Cold War. This pattern matters because it determined the direction of the Republican Party on the issue once the Cold War ended and conservatives became the dominant faction within it.

3.2 The Changing Interests of American Manufacturing

Another potentially important source of pressure for change in conservative opposition to multilateralism is the changing competitiveness of the American manufacturing sector. The foreign policy of the Republican Party during the 1890–1914 period was rooted in this sector's demands for trade protection. The unilateralist policy that prevailed before World War I sought to limit the economic impact of competition with other developed states by excluding those states’ manufactured products from the American domestic market. The policy also aimed at carving out an economic sphere of interest in Latin America, and to a lesser extent in East Asia, where American exporters and investors would have privileged access. By contrast, a multilateral order like the one the United States pursued after World War II promised greater access to developed country markets and sites for investment but would have required greater American economic openness than Republicans could countenance. For this reason, Henry Cabot Lodge specifically excluded American tariffs from the jurisdiction of the League of Nations in his reservations to the Versailles Treaty.

The changing international position of the American economy during the last century raises two issues that are important for our analysis. The first concerns in the increasing competitiveness of American manufacturing during the interwar period and especially in the immediate aftermath of World War II. Given the importance of the manufacturing sector to the Republican Party in the early twentieth century, this development should have influenced their position on multilateralism in much the way the Cold War did, making participation in a multilateral order more attractive.

The second issue concerns the distributive impact of globalization in recent decades. Participation in the global economy created winners and losers in American society, especially as it deepened. This development suggests an alternative explanation for the post–Cold War resurgence in Republican skepticism toward multilateralism rooted in current conditions rather than ideological continuity with the past. Evidence that this skepticism was more widespread among Republican members of Congress whose constituents tended to lose from globalization would support this alternative explanation. We will examine several measures of constituent interests to test this possibility.

Indicators of constituent interests such as the size of the manufacturing sector could influence the foreign policy positions of their members of Congress in at least two ways. First, constituent interests could directly shape the views of representatives. Members might consider these interests either because of lobbying or simply because they understood their importance in the economy of their region. This is the direct effect we wish to estimate. Second, constituent interests could indirectly shape members’ positions by affecting the party and ideology of those elected to Congress. For instance, during the early twentieth century, Republicans tended to win elections in areas with large manufacturing sectors, while Democrats had more success in agricultural areas. This indirect effect of constituent interests is less interesting here than it would be in other settings. Multilateralism was rarely salient enough to shape election outcomes, so it makes sense to treat party and ideology as if they were exogenous to constituent economic interests here. With this in mind, we will control for the effects of party and ideology when estimating the impact of constituent interests in this analysis and focus on their direct effect on members’ positions.

Another model specification issue concerns the likelihood that the manufacturing sector had different effects on Republicans and Democrats. This is a common pattern in the analysis of congressional voting because some interests are more important in one party than in the other.Footnote ⁵² In our previous research, we found that the trade interests of the manufacturing sector were strongly associated with Republican foreign policy positions during the early twentieth century.Footnote ⁵³ We therefore expected its changing interests over time to have a greater impact on Republicans than on Democrats, an expectation the evidence supports.Footnote ⁵⁴

Figure 8 shows the impact of the manufacturing sector on Republicans in each debate, holding ideology at the party mean. The extent of employment in this sector had little impact on the first two debates. This began to change as American manufacturing became more internationally competitive during the interwar period and later. At this point, Republicans from manufacturing states—mainly located in the Northeast—became more supportive of multilateralism. These changing interests contributed to a substantial rift within the party that persisted through the early Cold War era. The best-remembered internationalist Republicans of the mid-twentieth century, such as Senator Arthur Vandenberg (R-MI), reflected these changing interests. These internationalist Republicans were more active participants in the making of American foreign policy than their more conservative copartisans and thus play a more prominent role in most historical narratives of the early Cold War era. Their prevalence within the party should not be overstated, though. Internationalists were a minority among congressional Republicans.Footnote ⁵⁵ The intraparty division associated with the manufacturing sector appears to have disappeared by the end of the Cold War. While the size of the manufacturing sector remained statistically significant in the 1994 and 2001–02 debates, its substantive effect was extremely small.

Fig. 8. The Manufacturing Sector and Republican Opposition to Multilateralism.

Notes. Predicted probabilities are for a Republican with the party mean ideology score for that Congress. The whiskers indicate 95 percent confidence interval. Low and high manufacturing employment are 1 standard deviation below and above the mean, respectively. Numerical model results are reported in the Appendix.

The data on sponsorship of anti-UN bills offer a better test of the effect of constituent interests for the last five decades. In addition to providing more complete and continuous coverage of legislative opposition to multilateralism during these years, data on House districts provide a more fine-grained picture of constituent interests than do the state-level data we used in our analysis of Senate voting.

Using these data, we evaluate not only the size of the manufacturing sector but also two other indicators of constituents’ economic stakes in a relatively open world economy, a central goal of postwar multilateralism. While some manufacturing industries have remained highly competitive as the exposure to the world economy grew during the last fifty years, the effect of globalization on manufacturing employment has been a major political concern throughout this period.Footnote ⁵⁶ We expect manufacturing employment to be positively associated with opposition to multilateralism, especially among Republicans, during the 1973–2018 period. The other two constituent interest variables are factor-based indicators of the distributive impact of globalization. Because high-skill workers in a capital-abundant country like the United States should see their incomes rise with greater participation in the international trading system, we expect representatives from districts with relatively skilled populations to be less skeptical of multilateralism. To capture this effect, we will examine the percentage of college-educated persons and the percentage of persons in white-collar occupations.Footnote ⁵⁷

Figure 9 shows predicted probabilities of sponsoring at least one anti-UN bill for members of both parties with different constituencies. The size of the manufacturing sector had the expected effect on Democrats, though it was not substantively large. As expected during a time when American manufacturing faced growing international competition, Democrats representing districts with larger manufacturing sectors were somewhat more likely to oppose multilateralism. However, the effects of these constituent interests on Republicans were both large and the opposite of what we hypothesized. Those from districts with small manufacturing sectors were substantially more likely to sponsor at least one anti-UN measure than were Republicans from districts with large manufacturing sectors. While the source of this surprising pattern is unclear, it rules out the possibility that the negative impact of globalization on manufacturing employment explains Republican skepticism of multilateralism in recent decades.

Fig. 9. Constituent Interests and Cosponsorship of Anti-UN Bills in the House.

Notes. Predicted probabilities assume the party mean ideology score. The whiskers indicate 95 percent confidence intervals. Numerical model results are reported in the Appendix.

The results concerning college education and white-collar employment in the district population present the same puzzle. Among Democrats, the proportion of persons with a college degree was negatively associated with sponsorship of anti-UN bills, as we expected. This effect is small but meaningful. Among Republicans, though, the effect was the opposite of what we expected. Because college education and white-collar employment capture the same theoretical relationship and are highly correlated, it is not surprising that they produce nearly identical substantive results. In this case, too, the expected relationship occurs among Democrats but not among Republicans. For Republican representatives, white-collar employment in their district is associated with more anti-UN sponsorship activity, not less. As with the results concerning manufacturing employment, this pattern is strikingly inconsistent with the claim that the negative effects of globalization are responsible for continuing Republican opposition to multilateralism in American foreign policy.

Taken together, the evidence concerning constituent interests suggests that they played a role in moving some Republicans away from their traditional skepticism of multilateralism during the middle of the twentieth century. It helped produce the split between nationalist and internationalist Republicans during the early Cold War era. This split had largely disappeared by the end of the Cold War. Results for more recent decades are puzzling but entirely inconsistent with the argument that recent Republican opposition to multilateralism reflects the interests of constituents who lose from globalization. If anything, Republicans from districts that lost from globalization were actually less skeptical of multilateralism than were Republicans from districts that tended to benefit from it.

4.1 The Costs of Changing an Ideological Position

Even though the internal logic of an ideology is not tight enough to force its adherents to accept a particular position, abandoning an established position is costly for political and social reasons. Once elites publicly articulate a position, partisan and ideological loyalists in the general public will tend to adopt it as well. This pattern of opinion leadership is well established in previous research.Footnote ⁶¹ Changing the position risks alienating supporters who are invested in the old one. It also makes the leader articulating the new rationale appear inconsistent and perhaps insincere. The longer the old position has been held, and the more salient it has been, the more risky changing it is likely to be. Opposition to multilateral rules might appear esoteric, but it has been a staple of conservative rhetoric for a long time, often formulated in terms of national sovereignty or opposition to “globalism.”Footnote ⁶²

The frequent reintroduction of the Bricker amendment illustrates the social processes that reinforce the connection between opposition to multilateralism and membership in the conservative Republican political faction in Congress. By the 1970s, it was surely clear to those who introduced these bills that they had little chance of serious consideration. They nevertheless acted to demonstrate their loyalty to the conservative social circles from which they drew political support and inspiration. These demonstrations of loyalty, in turn, reinforced the faction's commitment to the policy. The remarks of Rep. John Ashbrook (R-OH) on reintroducing the Bricker amendment in January 1978 illustrate the social significance of his action:

Ashbrook was not alone in using the Bricker amendment to demonstrate his commitment to conservative positions to activists in his home state long after its 1954 defeat. Rep. Helen Chenoweth (R-ID) introduced the measure in 1997 and 1999, receiving praise in the right-wing press for doing so.Footnote ⁶⁴ Earlier in her political career, Chenoweth had served as the chief of staff for Rep. Steve Symms (R-ID), who had introduced the measure four times during his tenure in the House.Footnote ⁶⁵ The Bricker amendment was not central to the careers of either Symms or Chenoweth—neither chose to discuss the measure on the House floor when they introduced it—but their actions suggest how a policy position can be transmitted socially through time. Abandoning an established position, even one as extreme as the Bricker amendment, is costly. It risks alienating some group members, impugning past and present leaders who espoused the old position, and calling into question the solidity of the group's other ideological commitments. For party leaders, reiterating the old ideological line makes sense unless there is a compelling reason to change it.

4.2 Agenda Setting and Gatekeeping

By itself, the costliness of changing an ideological position is not enough to explain continuing Republican opposition to multilateral rules. Changing the party's orthodox position was not impossible, and adherence to the old position also had costs. By the mid-twentieth century, multilateral rules had become important for managing American relationships with developed allies, as well as offering economic advantages for the American manufacturing sector with its longstanding ties to the Republican Party. The evidence reviewed in the last section suggests that these considerations led some Republicans to modify their position on the issue. Why wasn't this change broader, as it was on other foreign policy issues?

The answer to this question concerns the efforts of party leaders. They engaged in gatekeeping to keep conservative opponents of multilateralism away from leadership positions on foreign policy. They also used their agenda-setting powers to exclude from active consideration measures would have seriously undermined the role of multilateralism in American foreign policy, particularly treaty debates where a minority could prevail. These efforts shielded multilateralism from its domestic political opponents, but also prevented conservative Republicans from confronting the costs of actually repudiating American commitment to multilateral rules. It meant they had little reason to challenge party orthodoxy.

During the Cold War, presidents and congressional leaders generally supported multilateralism in foreign policy more than conservative Republicans did. This was obviously the case for Democrats, who controlled both the House and the Senate for nearly the entire period, but it remained true even when Republicans held congressional majorities. The Senate was especially important because of its treaty-ratification power. When Republicans took charge, members of its internationalist wing occupied key foreign policy positions. For instance, all of the Republicans who chaired the Senate Foreign Relations Committee during the Cold War had DW-NOMINATE scores to the left of the party mean. Indeed, this was true of every Republican Foreign Relations Committee chair between the death of Henry Cabot Lodge in 1924 and the accession of Jesse Helms in 1995. The famously internationalist Arthur Vandenberg, who sponsored the resolution ratifying the UN Charter in 1945, was actually the closest to the Republican mean during this long period. Republican presidential nominees were also more internationalist than the rest of the party during the mid-twentieth century, when the United States made its most important multilateral commitments. Things might have turned out differently if Republicans had chosen an isolationist in 1940, rather than the internationalist Wendell Willkie, or Robert Taft instead of Dwight Eisenhower in 1952.

The nature of the issue made it possible for presidents and congressional leaders to keep the question of multilateral rules off the legislative agenda. Once the United States made its major initial multilateral commitments in the immediate postwar era, multilateralism required less legislative action than military spending, where a vote on the annual budget was unavoidable, or trade policy, where occasional votes were also difficult to avoid. Presidents could use executive agreements to avoid the necessity of Senate treaty ratification. As one account of Senate treaty powers noted, “since the initial post-World War II security treaties … security commitments have been made almost entirely by means other than treaties.”Footnote ⁶⁶ Indeed, ending this practice was one of the main conservative motives for the Bricker amendment. The central role of the United States in world politics also made it possible for American policymakers to exercise influence over the shape of international institutions even when the United States has not ratified—or even signed—the multilateral agreements that established them.

Avoiding floor votes about multilateral rules no doubt appeared prudent to congressional leaders. The near-passage of the Bricker amendment in 1954 vividly illustrated what might happen if there were a serious debate on the matter. The measures that died in the Foreign Relations Committee included all of the reintroduced versions of the Bricker amendment and nearly all of the anti-UN bills in our sample. Of the 292 anti-UN bills we identified between 1973 and 2018, just 10 (3.4 percent) received floor consideration and only one (0.3 percent) became law. By comparison, of the 15,406 bills related to international affairs introduced during this same period, 21.3 percent received floor consideration, and 3.0 percent became law.Footnote ⁶⁷ Agenda setting affected not only efforts to turn back multilateralism but also measures that would have advanced American adherence to it. Many multilateral agreements that the Senate might have embarrassingly rejected were instead simply never debated. For example, in a direct response to the Bricker amendment, the Eisenhower administration shelved the Genocide Convention, then before the Foreign Relations Committee.Footnote ⁶⁸ The same practice extended to other multilateral human rights agreements. As of 2017, the United States had ratified only seventeen of the 46 UN-sponsored human rights treaties it had signed.Footnote ⁶⁹

While this strategy allowed the executive branch to operate within existing multilateral institutions largely unimpeded during the Cold War, it had the perverse effect of giving conservative Republicans no reason to reconsider their position on the issue. If they had been forced to confront the costly consequences of rejecting multilateral rules, they might have chosen to take a new position and explain it to the party faithful in terms that made it seem consonant with their other partisan and ideological positions. Vandenberg, long considered a leading conservative, did precisely this when supporting the establishment of the United Nations in 1945.Footnote ⁷⁰ Even though most Republicans ended up supporting some version of the Bricker amendment in 1954, the measure discomfited some who had initially been inclined to support it, especially given the strong opposition of the Eisenhower administration. Some of the sixty-three senators who had sponsored the original measure ended up voting to weaken it, and twelve even voted against its final passage.Footnote ⁷¹ If such consequential votes had happened more often, the process might have led to a lasting change in the position associated with the conservative Republican faction.

Congressional gatekeeping and agenda setting to protect American multilateral commitments ended in the 1990s. The Republican Party became increasingly conservative, and, as we have seen, the connection between conservatism and opposition to multilateralism became stronger. The party no longer consistently chose congressional leaders or presidential candidates who supported multilateralism. When Republicans gained control of the Senate following the 1994 elections, Jesse Helms (R-NC) became chair of the Senate Foreign Relations Committee. He sponsored and secured floor passage of a series of measures intended to undermine the functioning of the International Criminal Court, the UN, and other multilateral organizations. These actions are often seen as the end of the bipartisan consensus on American foreign policy. It would be more accurate to understand them as evidence that the bipartisan consensus was never as far-reaching as it sometimes seemed.