Published online by Cambridge University Press: 10 August 2010
This article offers a first-ever comprehensive empirical assessment of a key Progressive reform, the direct primary, and its impact on competition in American elections. We begin with a review of the problems Progressives diagnosed in the American electoral system and reasons to expect the direct primary to be a pro-competitive, democratizing reform. We then consider prior research into the direct primary and electoral contestation and describe the database of primary and general election outcomes that we have constructed to trace competition in primaries for federal and statewide offices. Finally, we examine the historical trajectory of competition in primary elections, starting with the first decades after the introduction of the reform and then the succeeding decades.
Consistent with the hopes of reformers, we find primary elections indeed provided a forum for contestation for federal and statewide elections. Although primaries were never broadly competitive, even at the outset, they accounted for about a third of the serious electoral tests faced by statewide officeholders and about a fifth faced by U.S. representatives. The role of primaries as a venue for robust contestation, however, was short-lived, as the competitiveness of federal and statewide primaries decreased sharply starting in the 1940s. The last section of this article explores whether two recent developments in American elections—the extension of two-party competition and the rise in the value of incumbency—conspired to temper the contribution of direct primaries to electoral competition.
2. And beyond: a Democrat did not win the governorship in Vermont until 1962. Vermont's gubernatorial elections were annual until 1970.
3. Unlike in the South, where Republicans were merely a nuisance after about 1880—and almost invisible after about 1900—Democrats in the North were rarely shut out even in the most Republican states. Republicans won fifteen of the seventeen gubernatorial races in Michigan between 1878 and 1910 but never once achieved 70 percent of the vote. Democrats in South Carolina not only won every race for governor in the same period, but their nominees failed only once to top 70 percent, in 1894, when the nominee drew just 69.6 percent.
4. On the history of efforts to reform U.S. political parties, see Ranney, Austin, Curing the Mischiefs of Faction: Party Reform in America (Berkeley: University of California Press, 1975)Google Scholar. Merriam, Charles E. and Overacker, Louise, Primary Elections (Chicago: University of Chicago Press, 1928), 267–70;Google ScholarKey, V. O. Jr., American State Politics: An Introduction (New York: Alfred A. Knopf, 1956), 85–97;Google ScholarKey, V. O. Jr., Politics, Parties, and Pressure Groups, 4th ed. (New York: Thomas Y. Crowell Co., 1958), 411–13Google Scholar; and Sorauf, Frank J. and Beck, Paul Allen, Party Politics in America, 6th ed. (Glenview, IL: Scott, Foresman and Co., 1988); 243–45Google Scholar, among others, cite limited two-party contestation as a central consideration in the adoption of the direct primary. For a critique and an alternative view, see Ware, Alan, The American Direct Primary: Party Institutionalization and Transformation in the North (New York: Cambridge University Press, 2002)Google Scholar.
5. The authorship of the remark is uncertain. It is most often attributed to H. L. Mencken, although it seems out of character for him to associate with the sentiments of “goo-goo” reformers. It is also often credited to Al Smith, on whom it fits more comfortably. It has been ascribed to numerous others as well, including Alexis de Tocqueville, Edward Abbey, Jane Addams, Winston Churchill (of course), and Sidney Hook.
6. Merriam and Overacker, Primary Elections, Chapter 5.
7. Lovejoy, Allen Fraser, La Follette and the Establishment of the Direct Primary in Wisconsin, 1890–1904 (New Haven, CT: Yale University Press, 1941): 36Google Scholar. See also Ranney, Curing the Mischiefs of Faction, 115–26; and Hofstadter, Richard, The Age of Reform (New York: Random House, 1955), 257–71Google Scholar. Merriam was both a social scientist and a Progressive reformer, but his evaluation of the direct primary reform, two decades into its operation, is also longer on exposition than empiricism. See Merriam and Overacker, Primary Elections.
8. From The Old Order Changeth, as quoted in Hofstadter, The Age of Reform, 258.
9. Anthony Downs, An Economic Theory of Democracy (Harper & Row, 1957). The opposition to the direct primary, in fact, made this very argument, that nominees chosen by caucus were more likely to be “electable” than candidates chosen by primary. The critics warned that the primary would preclude ticket balancing and raise the chances that a weak candidate with limited support would win the nomination in a multicandidate race. See Merriam and Overacker, Primary Elections, 211–12; Ranney, Curing the Mischiefs of Faction, 125–26.
10. Key, American State Politics, 94–97; Lovejoy, La Follette and the Establishment of the Direct Primary in Wisconsin, Chapter 3; Merriam and Overacker, Primary Elections, Chapter 9; Ranney, Curing the Mischiefs of Faction, 119–20, 124–25; Sorauf and Beck, Party Politics in America, 244.
11. Merriam and Overacker, Primary Elections, 267–68; Key, American State Politics, 85–92; Key, Politics, Parties, and Pressure Groups, 411–13; Sorauf and Beck, Party Politics in America, 243–45.
13. Massachusetts and Ohio voted Republican in every presidential election between 1864 and 1908. In the same period, Illinois and Nebraska each defected to the Democratic nominee just once, Illinois to Grover Cleveland in 1892 and Nebraska to its favorite son William Jennings Bryan in 1896.
14. Federalism also reinforced single-party monopolies at the local level. The ability to control state and local offices allowed party leaders to extract resources without competing for national political offices. For a theoretical treatment of this issue, see Ansolabehere, Stephen and Snyder, James M. Jr., “Valence Politics and Equilibrium in Spatial Models,” Public Choice 103 (1999): 327–36Google Scholar. For an example of how this logic played out historically, see Burner, David, The Politics of Provincialism: The Democratic Party in Transition, 1918–1932 (Cambridge, MA: Harvard University Press, 1986)Google Scholar.
15. See, e.g., Key, American State Politics, 104–16; Key, V. O. Jr., Southern Politics in State and Nation (New York: Alfred A. Knopf, 1949Google Scholar), Part 1; Morlan, Robert L., Political Prairie Fire: The Nonpartisan League (Minneapolis: University of Minnesota Press, 1955)Google Scholar; Jensen, Richard J., The Winning of the Midwest: Social and Political Conflict, 1888–1896 (Chicago: University of Chicago Press, 1971)Google Scholar.
16. See Merriam and Overacker, Primary Elections, Chapter 10; Key, Politics, Parties, and Pressure Groups, 419–21.
17. As Key put it, “the ties of party—given the recency of the Civil War—made it simpler to advance popular government by introducing the direct primary than to proceed by a realignment of the parties” (Key, Politics, Parties, and Pressure Groups, 412).
18. Key, Southern Politics, Part 1.
19. Key, American State Politics, Chapter 4.
20. See Ewing, Cortez A. M., Primary Elections in the South: A Study in Uniparty Politics (Norman: University of Oklahoma Press, 1953)Google Scholar; Grau, Craig H., “Competition in State Legislative Primaries,” Legislative Studies Quarterly 6 (1981): 35–54Google Scholar; Jewell, Malcolm E., Legislative Representation in the Contemporary South (Durham, NC: Duke University Press, 1967)Google Scholar; and Jewell, Malcolm E. and Olson, David M., American State Political Parties and Elections (Homewood, IL: Dorsey, 1978)Google Scholar.
21. Grau, “Competition in State Legislative Primaries,” 35–54; Hogan, Robert E., “Sources of Competition in State Legislative Primary Elections,” Legislative Studies Quarterly 28 (2003): 103–26Google Scholar; Schantz, Harvey L., “Contested and Uncontested Primaries for the U.S. House,” Legislative Studies Quarterly 5 (1980): 542–62Google Scholar.
22. Turner, Julius, “Primary Elections as the Alternative to Party Competition in ‘Safe’ Districts,” Journal of Politics 15 (1953): 210Google Scholar; see also Jewell, Malcolm E. and Breaux, David, “Southern Primary and Electoral Competition and Incumbent Success,” Legislative Studies Quarterly 14 (1991): 129–43.Google Scholar
23. John R. Alford and Kevin T. Arceneaux, “Isolating the Origins of the Incumbency Advantage: An Analysis of House Primaries, 1956–1990,” paper presented at the annual meeting of the Southern Political Science Association, 2000; and Gerber, Elizabeth R. and Morton, Rebecca B., “Primary Election Systems and Representation,” Journal of Law, Economics, and Organization 14 (1998): 304–24Google Scholar; Westlye, Mark C., Senate Elections and Campaign Intensity (Baltimore, MD: Johns Hopkins University Press, 1991)Google Scholar; Berry, William D. and Canon, Bradley C., “Explaining the Competitiveness of Gubernatorial Primaries,” Journal of Politics 55 (1993): 454–71Google Scholar; Grau, “Competition in State Legislative Primaries,” 35–54.
24. These sources are listed in Table A1 in the Appendix.
25. For more information on the general election data, see Ansolabehere, Stephen and Snyder, James M., “The Incumbency Advantage in U.S. Elections,” Election Law Journal 1 (2002): 313–38Google Scholar.
26. The Democratic share of the two-party vote itself consists of the underlying partisan division of the state, or “normal vote”; the “personal vote” attributable to the candidates' own characteristics, including incumbency advantages and challenger quality; national party tides, which take the form of year effects; and idiosyncratic variation. For a decomposition of the vote into these components, see Ansolabehere and Snyder, “The Incumbency Advantage in U.S. Elections.”
27. The inclusion of “empty” primaries that draw no candidates for nomination has no substantive effect on our analysis. The same conclusions hold for primaries with single unopposed candidates.
28. Key, Southern Politics, Chapter 14; and American State Politics, 104–18.
29. We do not have estimates for competition in Democratic or Republican primaries in Republican states in the 1960s and 1970s. No states met the normal vote threshold for Republican domination in the 1960s, and only Alaska and Wyoming met it in the 1970s. For more on the estimation of the “normal vote,” see Ansolabehere and Snyder, “The Incumbency Advantage in U.S. Elections.”
30. We follow the literature in examining the electoral tests that officeholders face. By omitting the degree of competition in the primary won by the losing nominee, we do not introduce the bias that would result from introducing a large number of cases in which nominations were not worth enough to contest. In the first half of the twentieth century, for instance, the Republican Party did not regularly even hold primaries in most of the states in the South. Later in this article we compare competitiveness in dominant- and subordinate-party primaries in one-party states.
31. As one would expect, the statewide offices at the top of the ticket, governor and U.S. senator, were more competitive in the primaries than the down-ballot offices like attorney general and secretary of state. From 1910 to 1938, about half of gubernatorial and senatorial primaries were competitive, versus about two-fifths of the primaries for the lesser statewide offices.
32. Schattschneider, The Semi-Sovereign People, Chapter 5.
34. Gelman, Andrew and King, Gary, “Estimating Incumbency Advantage without Bias,” American Journal of Political Science 34 (1990): 1142–64Google Scholar; Levitt, Steven D. and Wolfram, Catherine D., “Decomposing the Sources of Incumbency Advantage in the U.S. House” Legislative Studies Quarterly 22 (1997): 45–60Google Scholar; Ansolabehere and Snyder, “The Incumbency Advantage in U.S. Elections.”
35. See also Turner, “Primary Elections as the Alternative to Party Competition in ‘Safe’ Districts.”
36. As we have already shown, statewide elections are more competitive than U.S. House elections, both overall and at the primary stage. Incumbency accounts for part of the difference. Because of term limits on many statewide executive offices, more than 30 percent of statewide elections compared to fewer than 10 percent of House elections fill open seats. See Ansolabehere and Snyder, “The Incumbency Advantage in U.S. Elections.”
37. In a separate analysis, we find that the value of incumbency rose sharply in primary elections a decade or so earlier than it did in general elections. We treat the topic of the incumbency advantage in primary elections more fully in a forthcoming article.
38. Ranney, Curing the Mischiefs of Faction, 127–29; Key, American State Politics, Chapter 5; and Politics, Parties, and Pressure Groups, 415–16.
39. See, e.g., Key, American State Politics, Chapters 4–6; and Polsby, Nelson W., Consequences of Party Reform (New York: Oxford University Press, 1983)Google Scholar. Progressives differed among themselves on the wisdom of the reform; see Hofstadter, The Age of Reform, 264–65.