Abstract
The failure of the United States to join the League of Nations is often considered to be an outcome of isolationist influence. The supermajority requirement of treaty ratification in the US Senate also is blamed for allowing a minority of isolationists to block the will of the majority that supported the treaty. To determine the cause of the failure, I analyze the Senate debate over the treaty using the concepts of the supermajority core and supermajority winset. Using all 157 votes on the treaty, I estimate senatorial positions and the locations of both the status quo and the treaty on the same metric space. From this analysis, I find that isolationists were not influential enough to block the ratification. Instead, President Wilson’s unwillingness to compromise is found to have played a critical role in the treaty’s defeat. The treaty’s defeat thus was not an indication of the power of isolationism. This study contributes to the growing body of literature that debunks claims about the dominance of isolationism in the interwar period. At the same time, the paper demonstrates how the core and winset concepts can be useful in answering substantive collective choice questions.
Similar content being viewed by others
Notes
Strictly speaking, the Senate cannot amend the text of a treaty. However, the Senate can attach reservations or understandings that can affect the interpretation or implementation of the treaty. Therefore, we can treat “reservations” as modifying the treaty, similar to amendments to regular legislation.
The classification rates of the two-dimensional models are slightly better than those of one-dimensional models (92 vs. 91% for 1919 and 85 vs. 82% for 1920). However, issue space is determined by the substance of the bills and amendments being debated. As discussed previously, the treaty debate concerned two major issues that were substantively independent. That is, no reason exists to assume that support for multilateralism is necessarily related to support for imperialism or opposition to Irish independence. See the online appendix for additional justifications for the use of a two-dimensional model.
See Miller (2015) for an introduction to the spatial voting model.
The supermajority core and supermajority winsets are computed using the simulation method discussed in the next section.
To test whether a simple majority or more of the Senate supported the status quo (which would reflect, of course, the dominance of isolationists), we need to use a different concept that captures the will of the majority in multidimensional spaces. The ‘uncovered set’ provides an ideal basis for this. For more information, see Miller (1980), McKelvey (1986), Bianco and Sened (2005), and Jeong et al. (2014). However, in the case of the Treaty of Versailles, we do not need to examine whether the status quo was supported by a majority or by as few as one-third of the Senate because both hypotheses are rejected altogether in the following section.
This amendment proposed releasing all members of the League of Nations from any oblication to the United States under Article X, which called for assistance to be given to a member nation that was a victim of external aggression.
This exclusion did not affect the estimation results, as the amendment was defeated (4–68) and thus did not change the agenda structure.
The salience of both dimensions is assumed to be equal for model identification.
When the Monte Carlo Markov Chain (MCMC) method is used to estimate a model, we draw samples from the target density. Those samples (or chains of samples) are used to draw inferences (see Hastings 1970).
The movement from TO to TR(R) might seem impossible when TO is inside the supermajority core. However, the vote to adopt the reservations was made on the basis of simple majority rule. Only the ratification of the treaty required a two-thirds majority.
See Poole and Rosenthal (1997).
The estimated linear regression model is: Position on the Irish Dimension = 0.73 (0.67) + 0.28 (1.28) Democrats + 1.31 (0.65) Ideology + 0.52 (0.36) Wilson’s Vote Share + 0.13 (0.32) Irish Population. The numbers in the parentheses are standard errors. Note that the positive coefficient on Ideology indicates that conservative senators were more likely to oppose Irish independence. Also, Ideology (i.e., the DW-NOMINATE dimension 1 score) is highly correlated with the position on the first dimension (the correlation coefficient is −0.73; p < 0.01), indicating that conservative senators were more likely to support nationalistic positions on the treaty.
See Duff (1968) for Irish-Americans’ reactions to the treaty.
References
Adler, S. (1966). The isolationist impulse, its twentieth century reaction. New York: Free Press.
Ambrosius, L. (1972). Wilson, the Republicans, and French security after World War I. Journal of American History, 59, 341–352.
Ambrosius, L. (1987). Woodrow Wilson and the American diplomatic tradition. Cambridge: Cambridge University Press.
Bailey, T. (1945). Woodrow Wilson and the great betrayal. New York: The Macmillan Co.
Bianco, W., Jeliazkov, I., & Sened, I. (2004). The uncovered set and the limits of legislative action. Political Analysis, 12(3), 256–276.
Bianco, W., & Sened, I. (2005). Uncovering evidence of conditional party government: Reassessing majority party influence in Congress and state legislatures. American Political Science Review, 99(3), 361–371.
Braumoeller, B. (2010). The myth of American isolationism. Foreign Policy Analysis, 6, 349–371.
Clinton, J., & Meirowitz, A. (2004). Testing explanations of strategic voting in legislatures: A reexamination of the Compromise of 1790. American Journal of Political Science, 48(4), 675–689.
Cooper, J., Jr. (2001). Breaking the heart of the world: Woodrow Wilson and the fight for League of Nations. Cambridge: Cambridge University Press.
Darling, H. (1929). Who kept the United States out of the League of Nations. Canadian Historical Review, 10, 196–211.
Duff, J. (1968). The Versailles Treaty and the Irish-Americans. The Journal of American History, 55, 582–598.
Fensterwald, B., Jr. (1958). The anatomy of American ‘isolationism’ and expansionism. Part I. Journal of Conflict Resolution, 2, 111–139.
Greenberg, J. (1979). Consistent majority rule over compact sets of alternatives. Econometrica, 47, 627–636.
Guinsburg, T. (1982). The pursuit of isolationism in the United States Senate from Versailles to Pearl Harbor. New York: Garland.
Guinsburg, T. (1994). The triumph of isolationism. In Gordon Martel (Ed.), American foreign relations reconsidered, 1890–1993. London: Routledge.
Hammond, T., & Miller, G. (1987). The core of the constitution. American Political Science Review, 81, 1155–1174.
Hastings, W. (1970). Monte Carlo sampling methods using Markov chains and their applications. Biometrika, 57, 97–109.
Holt, W. (1933). Treaties defeated in the Senate. Baltimore: Johns Hopkins Press.
Jeong, G. (2008). Testing the predictions of the multidimensional spatial voting model with roll call data. Political Analysis, 16, 179–196.
Jeong, G., Lowry, W., Miller, G., & Sened, I. (2014). How preferences change institutions: The 1978 Energy Act. Journal of Politics, 72(2), 430–445.
Johnson, R. (1995). The peace progressives and American foreign relations. Cambridge, MA: Harvard University Press.
Krehbiel, K. (1998). Pivotal politics: A theory of U.S. lawmaking. Chicago: The University of Chicago Press.
Margulies, H. (1989). The mild reservationists and the League of Nations controversy in the Senate. Columbia, MO: University of Missouri Press.
McKelvey, R. (1976). Intransitivities in multidimensional voting models and some implications for agenda control. Journal of Economic Theory, 12(3), 472–482.
McKelvey, R. (1986). Covering dominance and institution free properties of social choice. American Journal of Political Science, 30(2), 283–314.
Mervin, D. (1971). Henry Cabot Lodge and the League of Nations. Journal of American Studies, 4, 201–214.
Miller, N. (1980). A new solution set for tournament and majority voting. American Journal of Political Science, 24(1), 68–96.
Miller, N. (2015). The spatial model of social choice and voting. In Jac C. Heckelman & Nicholas R. Miller (Eds.), Handbook of social choice and voting. Cheltenham: Edward Elgar.
Poole, K., & Rosenthal, H. (1997). Congress: A political economic history of roll call voting. Oxford: Oxford University Press.
Rivers, D. (2003). Identification of multidimensional spatial voting models. Typescript. Stanford University.
Schofield, N. (1978). Instability of simple dynamic games. Review of Economic Studies, 45(3), 575–594.
Schofield, N., Grofman, B., & Feld, S. (1988). The core and the stability of group choice in spatial voting games. American Political Science Review, 82, 195–211.
Stone, R. (1970). The irreconcilables: The fight against the League of Nations. Lexington: The University Press of Kentucky.
Toth, C. (1961). Isolationism and the emergence of Borah: An appeal to American tradition. The Western Political Quarterly, 14, 555–568.
Acknowledgements
I thank Arjun Chowdhury, Dominik Stecula, Editors, three anonymous reviewers, and participants at the Canadian-Comparative Workshop at the University of British Columbia for their helpful comments.
Author information
Authors and Affiliations
Corresponding author
Electronic supplementary material
Below is the link to the electronic supplementary material.
Appendix
Appendix
In this appendix, I explain how the model is identified and provide a sensitivity analysis. When using an ideal point estimation technique, some anchors are necessary in order to identify the model. In my analysis, two points are fixed a priori: the status quo—i.e., SQ = (−2, −2)—and the original treaty submitted by President Wilson—i.e., TO = (2, 2). No ideal points are fixed a priori. This set of anchors is used to set the direction of the space such that a vote for the original treaty is considered as indicative of support for both multilateralism and imperialism as well as opposition to Irish independence. Both senatorial ideal points and the modified versions of the treaty are estimated relative to those anchors. Instead of fixing a set of bill locations, one can alternatively fix a set of senatorial ideal points to identify the model. However, my experience suggests that the model is identified more effectively by fixing the locations of the two bills than by fixing a set of ideal points. For instance, when I fixed William Borah—a well-known isolationist—at (−2, −2) and Gilbert Hitchcockat (2, 2), other isolationists, such as Robert LaFollette, were estimated to be around (−5, −5), despite the fact that Borah and LaFollette had almost identical voting records during the debate. This indicates that fixing ideal points is not effective in identifying the model.
In addition to fixing the SQ and TO, two more anchors are needed to fully identify the model (see Rivers 2003). However, instead of constraining two additional parameters a priori, I let the multiple Monte Carlo Markov Chain (MCMC) chains run and selected the chains that estimate the positions of the irreconcilables as occupying the southwest corner of the figures, which is consistent with the choice of the anchors discussed above. The advantage of this approach is that it minimizes the number of parameters to be fixed a priori, while still allowing the model to be identified.
An additional reason for fixing the location of the status quo is to avoid having to take into account two layers of uncertainty in computing the supermajority winset of the status quo. If the location of the status quo is estimated, the computation of the supermajority winset of the status quo has to take into account the uncertainty in estimating senatorial positions as well as the uncertainty in estimating the status quo. For this reason, fixing the status quo is computationally convenient while at the same time minimizing the number of ideal points or bill locations that need to be fixed a priori for model identification.
Nevertheless, to check the sensitivity of the estimation to the choice of anchors, I re-estimated the model using a different set of anchors. In this case, the only point I fix is the status quo, at (−2, − 2). All other proposal locations and ideal points are estimated relative to that point. The results are similar to the results that we obtain by fixing the status quo and the original treaty. The results are reported in the online appendix.
The model is estimated with a MCMC method. In order to minimize the role of the priors, I assigned diffuse priors. A uniform distribution with a range of −5 and 5 were assigned as prior distributions for location parameters. WinBUGS (Spiegelhalter et al. 1999) is used to fit the model (see the online appendix for the code). The total number of iterations was 55,000. The first 5000 draws were discarded to remove the effects of initial values. The chains were thinned every 50th draw to reduce autocorrelation. The convergence of the MCMC chains is checked using density plots and traceplots (see the online appendix).
Rights and permissions
About this article
Cite this article
Jeong, GH. The supermajority core of the US Senate and the failure to join the League of Nations. Public Choice 173, 325–343 (2017). https://doi.org/10.1007/s11127-017-0481-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11127-017-0481-5