1. See, for example, Schattschneider, E. E., Politics, Pressures and the Tariff (New York: Prentice-Hall, 1935)Google Scholar; Bauer, Raymond A., Pool, Ithiel de Sola, and Dexter, Lewis Anthony, American Business and Public Policy (New York: Atherton Press, 1963)Google Scholar; and Lowi, Theodore J., “American Business, Public Policy, Case Studies, and Political Theory,” World Politics 16 (07 1964), pp. 677–715CrossRefGoogle Scholar.

2. See, for example, Kindleberger, Charles P., The World in Depression, 1929–1939 (Berkeley: University of California Press, 1973)Google Scholar; Krasner, Stephen D., “State Power and the Structure of International Trade,” World Politics 28 (04 1976), pp. 317–47CrossRefGoogle Scholar; Katzenstein, Peter J., Between Power and Plenty (Madison: University of Wisconsin Press, 1978)Google Scholar; Keohane, Robert O., Beyond Hegemony (Princeton, N. J.: Princeton University Press, 1984)Google Scholar; and Lake, David A., Power, Protection, and Free Trade (Ithaca, N. Y.: Cornell University Press, 1988)Google Scholar.

3. See Schattschneider, Politics, Pressures and the Tariff; Bauer, Pool, and Dexter, American Business and Public Policy; Lowi, “American Business, Public Policy, Case Studies, and Political Theory”; McKeown, Timothy J., “Firms and Tariff Regime Change,” World Politics 36 (01 1984), pp. 215–33CrossRefGoogle Scholar; Cassing, James, McKeown, Timothy J., and Ochs, Jack, “The Political Economy of the Tariff Cycle,” American Political Science Review 80 (09 1986), pp. 843–62CrossRefGoogle Scholar; Wallerstein, Michael, “Unemployment, Collective Bargaining, and the Demand for Protection,” American Journal of Political Science 31 (11 1987), pp. 729–52CrossRefGoogle Scholar; and the studies reviewed by Nelson, Douglas in “Endogenous Tariff Theory: A Critical Survey,” American Journal of Political Science 32 (08 1988), pp. 796–837CrossRefGoogle Scholar.

4. See Kindleberger, The World in Depression; Krasner, “State Power and the Structure of International Trade”; Keohane. Beyond Hegemony; and Lake, Power, Protection, and Free Trade.

5. See Kindleberger, The World in Depression; Krasner, “State Power and the Structure of International Trade”; Lake, Power, Protection, and Free Trade; and Frieden, Jeff, “Sectoral Conflict and Foreign Economic Policy, 1914–1940,” International Organization 42 (Winter 1988), pp. 59–90CrossRefGoogle Scholar.

6. See Pastor, Robert, Congress and the Politics of U. S. Foreign Economic Policy (Berkeley: University of California Press, 1980)Google Scholar; and Haggard, Stephan, “The Institutional Foundations of Hegemony,” International Organization 42 (Winter 1988), pp. 91–120CrossRefGoogle Scholar.

7. Schattschneider, Politics, Pressures and the Tariff.

8. Lowi, “American Business, Public Policy, Case Studies, and Political Theory.”

9. U. S. Congress, Joint Committee on Printing, Tariff Acts Passed by the Congress of the United States from 1789 to 1909, Document no. 671. 61st Congress, 2d sess., 1909, p. 13Google Scholar.

10. See Downs, Anthony, An Economic Theory of Democracy (New York: Harper & Row, 1957)Google Scholar; and Kramer, Gerald H., “Short-Term Fluctuations in U. S. Voting Behavior, 1896–1974,” American Political Science Review 65 (03 1971), pp. 131–43CrossRefGoogle Scholar. Parties and their leaders are the focus here for two reasons. First, tax policy is of consistent and vital interest to political parties, leading them to expend more resources in its development than in the development of other policies. Second, while nineteenth-century members of Congress seem not to have been very interested in reelection, the political parties that sponsored them demonstrably were. See Manley, John F., The Politics of Finance (Boston: Little, Brown, 1970)Google Scholar; and Polsby, Nelson W., “The Institutionalization of the U. S. House of Representatives,” American Political Science Review 62 (03 1968), pp. 144–68CrossRefGoogle Scholar.

11. Downs, , An Economic Theory of Democracy, pp. 69–71Google Scholar.

12. Another option, of course, is to finance the deficit by borrowing, but that possibility will not be explicitly considered here. Except in times of war or other national emergencies, governments have avoided chronic recourse to debt finance, at least until recently. The reasons for their restraint are puzzling (see Buchanan, James M. and Wagner, Richard E., Democracy in Deficit [New York: Academic Press, 1977])Google Scholar, but they fall outside the scope of this article, except for two comments. First, as Musgrave and others point out, debt finance is no different analytically from other methods of public finance. The burdens of government simply fall on different voters and, in this case, most directly on other borrowers (see Musgrave, Richard A., The Theory of Public Finance [New York: Mc Graw-Hill, 1959], pp. 208 and 612–14)Google Scholar. Thus, the political appeal of deficit finance was limited, perhaps, by an electorate comprising more borrowers than lenders. Second, the argument here asserts only that parties move policy toward budget balance, not that they achieve it.

13. See Riker, William H., The Theory of Political Coalitions (New Haven, Conn.: Yale University Press, 1962)Google Scholar; and Lowi, “American Business, Public Policy, Case Studies, and Political Theory.”

14. See Lowi, “American Business, Public Policy, Case Studies, and Political Theory”; and Downs, , An Economic Theory of Democracy, pp. 247–58Google Scholar. Exclusive and inclusive coalitions bear a family resemblance to minimum-winning and universalistic (or logrolling) coalitions, respectively (see Riker, The Theory of Political Coalitions; and Shepsle, Kenneth A. and Weingast, Barry R., “Political Preferences for the Pork Barrel,” American Journal of Political Science 25 [02 1981], pp. 96–111)CrossRefGoogle Scholar. The uncommon language, borrowed from Olson, Mancur (The Logic of Collective Action [Cambridge, Mass.: Harvard University Press, 1971, pp. 36–43)Google Scholar, avoids the unfortunate connotations of bipartisanship that universalism has acquired. As implied below, inclusive coalitions might be partisan or bipartisan, depending upon other characteristics of the tax.

15. The exact object of the inclusive coalitions depends also on the concentration and dispersion of costs and rents because these in turn affect party ideologies. Even if the distribution, though, is most favorable to logrolling (concentration opposed by dispersion), heavy dependence necessarily limits the practice, while minimal dependence allows it.

16. Two control variables related to this discussion also appear in the empirical analysis: the percentage of revenues raised from sources other than the tariff and the percentage of imports admitted without duty. As the government substitutes other tax revenues for tariff revenues, obviously, tariff rates should fall (see Baack, Bennett D. and Ray, Edward John, “The Political Economy of the Origin and Development of the Federal Income Tax,” in Higgs, Robert, ed., Emergence of the Modern Political Economy [Greenwich, Conn.: JAI Press, 1985], especially pp. 132–33)Google Scholar. Likewise, as the government exempts more goods from duties, tariff rates on the remainder should rise (see Taussig, F. W., The Tariff History of the United States, 8th ed. [New York: Putnam, 1931], especially pp. 178–89)Google Scholar.

17. Musgrave, , The Theory of Public Finance, p. 207Google Scholar.

18. See Downs, , An Economic Theory of Democracy, pp. 82–95Google Scholar; and Olson, , The Logic of Collective Action, pp. 51–52Google Scholar.

19. Downs, , An Economic Theory of Democracy, pp. 97–103Google Scholar.

20. Ibid., pp. 114–41.

21. Musgrave, , The Theory of Public Finance, pp. 355–64Google Scholar.

22. Because the tariff is a discriminatory assessment, scholars have traditionally analyzed tariffs as rents rather than taxes. There are several good arguments, however, for considering tariffs together with other taxes. First, a common theoretical structure for all kinds of revenue policies is analytically advantageous. Whether the government uses tax policy primarily for extraction or primarily for incentives is, as argued here, an interesting theoretical question that is obscured by considering tax-created rents wholly as government largess. Second, negative incidence arises, in truth, when the government declines to extract funds that it is legally and technically able to extract. The zero incidence of the income tax for many corporations resulted from the government's decision not to collect income taxes from them because they invested in durable capital goods. Similarly, the negative incidence of the tariff for many corporations resulted from the government's decision to tax imports competitive with domestic producers rather than goods produced only overseas (for example, tea and coffee). In both cases, the government could legally have extracted the same revenue by doing things differently, and it chose not to. Finally, tax-created rents differ in kind from the benefits of other government interventions, such as franchises, quotas, and subsidies. These other interventions have as their sole purpose the creation of private rents; taxes do not.

23. Stolper, Wolfgang F. and Samuelson, Paul A., “Protection and Real Wages,” Review of Economic Studies 9 (11 1941), pp. 58–73CrossRefGoogle Scholar.

24. See Wallerstein, , “Unemployment, Collective Bargaining, and the Demand for Protection,” pp. 732–34Google Scholar; and Magee, Stephen P., “Three Simple Tests of the Stolper-Samuelson Theorem,” in Oppenheimer, Peter, ed., Issues in International Economics (Stocksfield, U. K.: Oriel Press, 1978), pp. 138–53Google Scholar.

25. See Rogowski, Ronald, “Political Cleavages and Changing Exposure to Trade,” American Political Science Review 81 (12 1987), pp. 1121–38CrossRefGoogle Scholar. See also Brock, William A. and Magee, Stephen P., “The Economics of Special Interest Politics,” American Economic Review Papers and Proceedings 68 (05 1978), pp. 246–50Google Scholar; and Brock, William A. and Magee, Stephen P., “Tariff Formation in a Democracy,” in Black, John and Hindley, Brian, eds.. Current Issues in Commercial Policy and Diplomacy (New York: St. Martin's Press, 1980), pp. 1–9Google Scholar.

26. This commonsense notion is also consistent with the pressure group accounts of endogenous tariff theory in economics and the systemic explanations of hegemonic stability theory in political science, although the interpretation given it here is probably more friendly to the former.

27. Tufte, Edward R., Political Control of the Economy (Princeton, N. J.: Princeton University Press, 1978)Google Scholar.

28. See, for example, Schattschneider, , Politics, Pressures and the Tariff, pp. 103–213Google Scholar.

29. See Mc Keown, “Firms and Tariff Regime Change”; Gallarotti, Giulio M., “Toward a Business-Cycle Model of Tariffs,” International Organization 39 (Winter 1985), pp. 155–87CrossRefGoogle Scholar; Cassing, Mc Keown, and Ochs, “The Political Economy of the Tariff Cycle”; Wallerstein, “Unemployment, Collective Bargaining, and the Demand for Protection”; and Magee, Stephen P. and Young, Leslie, “Endogenous Protection in the United States, 1900–1984,” in Stern, Robert M., ed., U. S. Trade Policies in a Changing World Economy (Cambridge, Mass.: MIT Press, 1987), pp. 145–95Google Scholar.

30. Thus, the analysis presented here differs from most studies of tariff policy, and not only because it is quantitative. Among quantitative studies, most model the rates of duty in the cross-section, across industries, usually in recent periods (see the studies reviewed by Nelson in “Endogenous Tariff Theory”). Only a handful of quantitative studies examine policy variations over time. Mc Keown and Gallarotti each present a bivariate analysis of longitudinal data. Baack and Ray conduct their analysis across industries at three historical junctures, while Conybeare conducts his across nations at two junctures. Closest in empirical spirit to the study presented here is the analysis of Magee and Young, which aggregates by presidential administration for the years 1900 to 1984. In contrast to these earlier works, the empirical analysis offered here takes legislative acts as its cases, examines a longer period of time, and tests different hypotheses. See McKeown, “Firms and Tariff Regime Change”; Gallarotti, “Toward a Business-Cycle Model of Tariffs”; Baack, Bennett D. and Ray, Edward John, “The Political Economy of Tariff Policy,” Explorations in Economic History 20 (01 1983), pp. 73–93CrossRefGoogle Scholar; Conybeare, John A. C., “Tariff Protection in Developed and Developing Countries,” International Organization 37 (Summer 1983), pp. 441–67CrossRefGoogle Scholar; and Magee and Young, “Endogenous Protection in the United States.”

31. In fact, the federal government enacted twenty-eight general tariff bills in that time, including three in 1861 and two in 1875. Because the data on which the analysis is based are yearly, it is impossible to distinguish acts that occur in the same year.

32. See Bauer, Pool, and Dexter, American Business and Public Policy; Lowi, “American Business, Public Policy, Case Studies, and Political Theory”; and Pastor, Congress and the Politics of U. S. Foreign Economic Policy.

33. In conducting the analysis, I was attentive to the possibility that years in which tariff legislation passed might be systematically different from years in which it did not. Nonrandom selection into the sample is a problem if the error term of the selection equation is correlated with the error term of the outcome equation (in this case, if the unincluded factors that lead to passage of a general tariff bill are associated with the unincluded factors that influence the magnitude of the tariff revision). In the presence of nonrandom selection with correlated errors, the direct application of ordinary least squares to the outcome equation yields biased and inconsistent estimates (see Achen, Christopher H., The Statistical Analysis of Quasi-Experiments [Berkeley: University of California Press, 1986], pp. 1–16Google Scholar). Consequently, I initially estimated a two-stage selection model for censored samples. The first-stage equation, estimated on the entire period (n = 112 years), predicted the probability that a tariff bill would be enacted given the values of independent variables implied by the theory. The second-stage equation, estimated on the uncensored (that is, non-zero) part of the sample (n = 25 years), predicted the magnitude of changes in tariff rates given both the values of the independent variables specified above and (roughly) the probability that an observation was selected into the uncensored part of the sample (for the procedure and its rationale, see Heckman, James J., “The Common Structure of Statistical Models of Truncation, Sample Selection, and Limited Dependent Variables and a Simple Estimator for Such Models,” Annals of Economic and Social Measurement 5 [Fall 1976], pp. 475–92Google Scholar; Heckman, James J., “Sample Selection Bias as a Specification Error,” Econometrica 47 [01 1979], pp. 153–61CrossRefGoogle Scholar; and Achen, , The Statistical Analysis of Quasi-Experiments, pp. 97–137Google Scholar). These analyses found no evidence of selection bias; in the outcome equation, the coefficient for the sample selection term was dwarfed by its standard error. Rather than introduce needless complexity, therefore, I omit the selection equation and report only the outcome equation.

34. The validity of average ad valorem rates as measures of protection has sometimes been challenged. The problem is that they indicate nominal rather than effective rates of duty. Two points need to be made in their defense. First, even if they are not perfect measures of changes in protection, they are still good ones. They have the singular virtue of weighting each duty by its importance in consumption, allowing the analyst to judge the overall character of the legislation even if some duties go up and others go down. As measures of policy change, they are at least as valid as, say, Taussig's judgments. Second, and more important, even if nominal rates measure effective rates with error, their use does not necessarily jeopardize the conclusions of the analysis. Errors in measurement of a dependent variable undermine the efficiency of ordinary least squares estimates, but they do not bias the coefficients.

35. The second equation, in fact, is a bit superfluous. The two dependent variables are related to one another by the inverse of the change in the proportion of imports subject to duty. Thus, the coefficients estimated from the two equations are related by the same quantity if all the independent variables are truly exogenous. In this case, however, some of the independent variables are probably not wholly exogenous; some of the independent variables likely cause other independent variables. (Unfortunately, I do not have the data necessary to unpack all those relationships.) Thus, the specifications of the two equations (and the coefficients) are slightly different, although the variations make theoretical sense and conform to expectations.

36. This finding is subject to several indistinguishable interpretations, based as it is on aggregate data. Several analysts have offered accounts of stepped-up political pressure by importcompeting sectors during recessions (see Mc Keown, “Firms and Tariff Regime Change”; Gallarotti, “Toward a Business-Cycle Model of Tariffs”; Cassing, Mc Keown, and Ochs, “The Political Economy of the Tariff Cycle”; and Magee and Young, “Endogenous Protection in the United States”). Wallerstein adds intensified pressure by labor (see “Unemployment, Collective Bargaining, and the Demand for Protection”). The interpretation given here is best understood as the complement to these interpretations. It emphasizes the government's eagerness to grant relief to voters for hard times, while the previous accounts emphasize voters’ demands for compensation.

37. In addition, the greatly worsened performance of U. S. trade during the Depression is a major reason why Roosevelt's revision of the tariff was more timid than Wilson's, despite Roosevelt's unilateral authority under the Reciprocal Trade Agreements Act.

38. As before, the finding is subject to several indistinguishable interpretations. One line of argument highlights the changes in pressure group demands that result from changes in economic interest. Another holds that a unitary state bureaucracy reassesses trade policy in light of the nation's repositioning in the world economy (see, for example, Krasner, “State Power and the Structure of International Trade”). Each arrives at the same conclusion as the median voter argument offered here.

39. For the most part. Democratic presidents signify Democratic congresses as well. Passage of tariff legislation when control of the government was divided was quite uncommon. Of the twenty-eight general tariff bills approved from 1829 to 1940, only two won passage when control was split. In lame-duck sessions in 1857 and 1861, Democrats under Pierce and Buchanan held the White House and the Senate, but Republicans held the House of Representatives.

40. Taussig, , The Tariff History of the United States, pp. 178–89Google Scholar.

41. See Witte, John F., The Politics and Development of the Federal Income Tax (Madison: University of Wisconsin Press, 1985), pp. 67–87Google Scholar; and Waltman, Jerold L., Political Origins of the U. S. Income Tax (Jackson: University Press of Mississippi, 1985), pp. 3–31Google Scholar.

42. In “The Political Economy of the Origin and Development of the Federal Income Tax,” pp. 132–33, Baack and Ray uncover the equivalent effect of revenue substitution on income taxes.

43. The Smoot-Hawley bill is so often presented as an exemplar of universalist politics that it is well to recall the sharp partisanship involved in its passage. In the House, 95 percent of the Republicans voted or paired for the bill, while 86 percent of the Democrats voted or paired against it. In the Senate, 89 percent of the Republicans indicated support, while 77 percent of the Democrats indicated opposition. The votes on the Reciprocal Trade Agreements Act in 1934 were just as partisan.

44. Schattschneider, , Politics, Pressures and the Tariff, pp. 135ffGoogle Scholar.

45. Of course, these were major products of the Democratic South.

46. In “The Political Economy of the Origins and Development of the Federal Income Tax,” pp. 132–33, Baack and Ray found a similar effect on the income tax.

47. In this interaction term, the percentage of funds derived from the tariff is a five-year moving average beginning at time t. This allows policymakers to anticipate the consequences of enacting new taxes, especially at major turning points such as 1861 and 1913.

48. As it happens, in all instances before the Civil War, Democrats presided over surpluses and Whigs over deficits. In all instances after World War I, on the other hand, Democrats presided over deficits and Republicans over surpluses. Thus, the argument faces a more exacting test later than it does earlier. In between, however, the argument works perfectly. Every time Republicans ran a surplus between 1863 and 1918, they lowered tariff rates, with one exception. That was in 1909, a deficit year, when Republican progressives, allied with Democrats, forced Republican stalwarts to lower tariffs, in defiance of both fiscal and partisan logic.

49. Schattschneider, , Politics, Pressures and the Tariff, p. 135Google Scholar.

50. What might have happened had the United States enjoyed hegemony before it cut itself loose from tariff revenues is an interesting but moot question. The more serious issue is the relationship between trade interests and the adoption of the income tax. The tariff was central, of course, to the debates over the income tax, but concerns about its impact on trade were entirely eclipsed by concerns about its incidence (see Witte, , The Politics and Development of the Federal Income Tax, pp. 67–109Google Scholar; and Waltman, , Political Origins of the U. S. Income Tax, pp. 1–31Google Scholar). Baack and Ray, however, offer an account of the triumph of the income tax that introduces trade considerations much more obliquely. They attribute the success of the 16th Amendment to the support of maritime interests, which sought the revenues necessary to maintain a global navy. See Baack, Bennett D. and Ray, Edward John, “Special Interests and the Adoption of the Income Tax in the United States,” Journal of Economic History 45 (09 1985), pp. 607–25CrossRefGoogle Scholar.