1. See Santis, Vincent De, Republicans Face the Southern Question: The New Departure Years, 1877–1897 (Baltimore, MD: Johns Hopkins University Press, 1959)Google Scholar; Hirshson, Stanley, Farewell to the Bloody Shirt: Northern Republicans and the Southern Negro, 1877–1893 (Bloomington: Indiana University Press, 1962)Google Scholar; Wang, Xi, The Trial of Democracy: Black Suffrage and Northern Republicans, 1860–1910 (Athens: University of Georgia Press, 1997)Google Scholar; Calhoun, Charles W., Conceiving a New Republic: The Republican Party and the Southern Question, 1869–1900 (Lawrence: University Press of Kansas, 2006)Google Scholar.

2. See, for example, Black, Earl and Black, Merle, The Rise of Southern Republicans (Cambridge, MA: Harvard University Press, 2002)Google Scholar; Lublin, David, The Republican South: Democratization and Partisan Change (Princeton, NJ: Princeton University Press, 2004)Google Scholar; Mickey, Robert, Paths Out of Dixie: The Democratization of Authoritarian Enclaves in America's Deep South, 1944–1972 (Princeton, NJ: Princeton University Press, 2015)Google Scholar.

3. The sole exceptions to this were Warren G. Harding in 1920 (winning Tennessee) and Herbert Hoover in 1928 (winning Florida, North Carolina, Tennessee, Texas, and Virginia).

4. Heersink, Boris and Jenkins, Jeffery A., “Southern Delegates and Republican National Convention Politics, 1880–1928,” Studies in American Political Development 29 (2015): 68–88CrossRefGoogle Scholar.

5. Ibid.

6. For background on the composition and arguments of the Black-and-Tans and Lily-Whites, see Walton, Hanes Jr., Black Republicans: The Politics of the Black and Tans (Metuchen, NJ: The Scarecrow Press, 1975)Google Scholar.

7. See Walton, Black Republicans; Lisio, Donald J., Hoover, Blacks, & Lily Whites: A Study of Southern Strategies (Chapel Hill: The University of North Carolina Press, 1985)Google Scholar.

8. Case in point was the passage of various “Black Codes” by Southern state legislatures, which would have established a second-class citizenship status for blacks through the adoption and implementation of draconian vagrancy laws. Such laws would have forced blacks into low-wage work contracts on plantations, in order to pay fines accrued because they were considered unemployed “vagrants.” Simply put, the Black Codes were meant to mimic as closely as possible the political-economic aspects of the slave economy, given the post–Civil War realities (the abolition of slavery via the 13th Amendment)—and, in doing so, maintain the pre–Civil War system of white supremacy. See Wilson, Theodore Brantner, The Black Codes of the South (Tuscaloosa: University of Alabama Press, 1965)Google Scholar; Cohen, William, At Freedom's Edge: Black Mobility and the Southern White Quest for Racial Control, 1861–1915 (Baton Rouge: Louisiana State University Press, 1991)Google Scholar.

9. Abbott, Richard H., The Republican Party and the South, 1855–1877 (Chapel Hill: University of North Carolina Press, 1986)Google Scholar; Foner, Eric, Reconstruction: America's Unfinished Revolution, 1863–1877 (New York: Harper & Row, 1988)Google Scholar.

10. Foner, Reconstruction, 354–55.

11. Woodward, C. Vann, Reunion and Reaction: The Compromise of 1877 and the End of Reconstruction (New York: Oxford University Press, 1966)Google Scholar; Polakoff, Keith Ian, The Politics of Inertia: The Election of 1876 and the End of Reconstruction (Baton Rouge: Louisiana State University Press, 1973)Google Scholar.

12. De Santis, Republicans Face the Southern Question; Hirshson, Farewell to the Bloody Shirt.

13. As Charles Calhoun notes: “Key's appointment to head the department with the largest patronage signaled Hayes's intention to use the bestowal of office to woo conservative southern Democrats, especially those of Whig antecedents, to the Republican Party.” Calhoun, Conceiving a New Republic, 138.

14. De Santis, Republicans Face the Southern Question; Hirshson, Farewell to the Bloody Shirt.

15. See, e.g., Donald, David Herbert, Baker, Jean Harvey, and Holt, Michael F., The Civil War and Reconstruction (New York: Norton, 2001), 584, 598–601Google Scholar.

16. De Santis, Vincent P., “President Arthur and the Independent Movements in the South in 1882,” The Journal of Southern History 19 (1953): 346–63CrossRefGoogle Scholar; Doenecke, Justin D., The Presidencies of James A. Garfield and Chester A. Arthur (Lawrence: University Press of Kansas, 1981), 114–23Google Scholar

17. On Mahone and the Readjuster movement, see Moger, Allen W., Virginia: Bourbonism to Byrd, 1870–1925 (Charlottesville: University of Virginia Press, 1968)Google Scholar; Dailey, Jane, Before Jim Crow: The Politics of Race in Postemancipation Virginia (Chapel Hill: University of North Carolina Press, 2000)CrossRefGoogle Scholar; Valelly, Richard M., The Two Reconstructions: The Struggle for Black Enfranchisement (Chicago: University of Chicago Press, 2004), 57–59CrossRefGoogle Scholar; Tartar, Brent, A Saga of the New South: Race, Law, and Public Debt in Virginia (Charlottesville: University of Virginia Press, 2016)Google Scholar.

18. Mahone had optimal leverage to secure such patronage, as he was the pivotal voter in giving the Republicans organizational control of the Senate in the 47th Congress (1881–83). When the Congress assembled, there were 37 Republicans, 37 Democrats, one Independent (David Davis of Illinois), and one Readjuster (Mahone). Davis agreed to caucus with the Democrats, which meant the GOP needed Mahone (and Vice President Arthur as the eventual tiebreaker) to achieve majority control. In exchange for his support, Mahone received four key committee assignments and a share of executive patronage for his party from the GOP. See Dailey, Before Jim Crow, 53–54.

19. De Santis, Republicans Face the Southern Question, 142–53.

20. On the conflict between the Black-and-Tans and Lily-Whites, see Walton, Black Republicans. And while some Lily-Whites wanted to throw all blacks out of the Republican Party, most understood the importance of keeping blacks within the organization (but out of leadership roles). As Ralph Bunche explained some years later: “The lily-white Republican organizations do not generally exclude Negroes entirely. There is no such thing as a pure white Republican primary in the South. In some states … the Negro Republican registrants are needed in order to give the party sufficient representation in the state to continue the party's legal recognition and keep its place on the ballot.” See Bunche, Ralph, The Political Status of the Negro in the Age of FDR (Chicago: University of Chicago Press, 1973), 82CrossRefGoogle Scholar.

21. Barr, Alwyn, Reconstruction to Reform: Texas Politics, 1876–1906 (Austin: University of Texas Press, 1971)Google Scholar; Casdorph, Paul D., A History of the Republican Party in Texas, 1865–1965 (Austin, TX: Pemberton Press, 1965)Google Scholar.

22. Hirshson, Farewell to the Bloody Shirt, 175–76, 179–81.

23. See Kousser, J. Morgan, The Shaping of Southern Politics: Suffrage Restriction and the Establishment of the One-Party South (New Haven, CT: Yale University Press, 1974)Google Scholar; Perman, Michael, Struggle for Mastery: Disfranchisement in the South, 1888–1908 (Chapel Hill: University of North Carolina Press, 2001)Google Scholar.

24. Many states also adopted “grandfather clauses,” which allowed poor whites to vote (despite their illiteracy and/or poverty), while still excluding blacks.

25. Note that the technical definition of the term “rotten borough”—which was introduced to refer to parliamentary boroughs in the United Kingdom in the Eighteenth Century that had such a small electorate that voters could not vote freely since they depended upon the “owner” of the borough for employment—does not entirely match the reality of the South's role in the Republican Party. However, the term was used at the time, and since then, to describe the South's role in the GOP. Most notoriously, Theodore Roosevelt used it to describe the role Southern delegates played in the 1912 Republican National Convention: “In the Convention at Chicago last June the breakup of the Republican Party was forced by those rotten-borough delegates from the South … representing nothing but their own greed for money or office” and who (in Roosevelt's perspective) “betrayed the will of the mass of the plain people of the party.” Cited in Cowan, Geoffrey, Let the People Rule: Theodore Roosevelt and the Birth of the Presidential Primary (New York: W.W. Norton, 2016), 265Google Scholar. Similarly, historian George B. Tindall argued that “William McKinley's campaign for president illustrated perfectly the standard uses of what had become the rotten boroughs of Republicanism in the South.” Tindall, George B., “Southern Strategy: A Historical Perspective,” The North Carolina Historical Review 48 (April, 1971): 137Google Scholar.

26. See Heersink and Jenkins, “Southern Delegates and Republican National Convention Politics.”

27. In addition to McKinley, Theodore Roosevelt, William Howard Taft, Calvin Coolidge, and Herbert Hoover all relied on this version of a Southern strategy to help ensure their (re)nomination at often unpredictable national conventions. In response, their political opponents within the party began to openly call for a reduction of the Southern delegation size. Thus, the South became an often controversial, but always relevant, element in national Republican politics. See Heersink and Jenkins, “Southern Delegates and Republican National Convention Politics.” At the same time, national Republican leaders never entirely abandoned hope for a Southern resurgence—though their (frequently-halfhearted) attempts at reinvigorating the party in the former Confederacy were rarely successful. In particular, Presidents Harding, Hoover, and, to a lesser extent, Roosevelt attempted to replace some of the corrupt party organizations with new ones built around handpicked local leaders who were deemed “high quality”—and nearly always white. While those attempts largely failed, they did place these presidents at the center of major clashes between Black-and-Tan and Lily-White factions over control of state party organizations.

28. The sole exception being a set of Southern state delegations (Florida, Mississippi, South Carolina, and Texas) printed in the proceedings of the 1896 convention: for these states, black delegates are identified as “colored” while white delegates receive no racial identification. See Republican National Convention, St. Louis, June 16th to 18th, 1896 (St. Louis: I. Haas Publishing and Engraving Company, 1896), 175–210.

29. Specifically, Article I, Paragraph 3, Section 2 of the Constitution of the United States says: “The actual Enumeration shall be made within three Years after the first Meeting of the Congress of the United States, and within every subsequent Term of ten Years, in such Manner as they shall by Law direct.”

30. Anderson, Margo J., The American Census: A Social History (New Haven, CT: Yale University Press, 1988), 9–12Google Scholar.

31. Race—together with age and sex—has been one of the few items consistently asked in every census. However, the option available to respondents has changed over time: in censuses collected between 1790 and 1840, information was collected by household and not individual—and a distinction was made between free white males, free white females, all other free persons, and slaves. After 1850, the Census Bureau began relying on a form that identified each individual person in a household, whereby each free individual was identified as being white, black, or mulatto. After the Civil War, the distinction between free and slave was dropped, but the three-fold definition of race remained in use. For the 1890 census, workers were given instructions as to how to further characterize black Americans (noting a distinction between black, mulatto, quadroon, and octoroon, depending on the extent to which an individual was deemed to have “black blood”). The term “Negro” was introduced in the 1900 census. The term “mulatto” was not included in the 1900 census, but reappeared in 1910 and 1920. For the purposes of this study, we identify any delegates whose census lists their race as any of the terms listed above as black. Anderson, Margo J., ed., Encyclopedia of the U.S. Census (Washington, DC: CQ Press, 2000), 19.Google Scholar

32. The end of slavery and, thereby, the three-fifths compromise provides no reason to believe local Democratic Southern leaders would subsequently have an incentive to frustrate attempts by census workers to incorporate black Southerners; while black Southerners were banned from voting, post–Civil War they did count as full citizens, increasing the population count for the South and the number of House seats provided to Southern (solidly Democratic) states. See Anderson, The American Census, 72.

33. These sources include any references to specific delegates in historical accounts of Republican party politics—during the Reconstruction era and beyond—that explicitly identify the race of individual delegates. Two books that were especially helpful in this regard were Foner, Eric, Freedom's Lawmakers: A Directory of Black Officeholders During Reconstruction (Baton Rouge: Louisiana State University Press, 1996)Google Scholar and Walton, Black Republicans. We also used lists of black delegates featured in newspaper articles or Monroe Work. ed., The Negro Year Book (Tuskegee Institute, Alabama), a reference volume that appeared in eleven editions between 1912 and 1952. The Negro Year Book provided lists of black delegates by state from the 1912 through 1936 conventions in the following editions: 1912 and 1916 convention delegates (1918–19 edition, 208–10), 1920 (1921–22 edition, 183), 1924 (1925–26 edition, 245), 1928 (1931–32 edition, 92), 1932, and 1936 (1937–38 edition, 100–01).

34. Finally, for delegates to the 1948 and 1952 conventions, there is a limitation in that the U.S. Census follows the “72 years” rule: individual census forms are not released until 72 years after the census was taken. As a result, the 1940 census is currently the last census that has been fully released. For the 1948 and 1952 conventions, delegate data is based on repeat delegates (that is, delegates who were also present at previous conventions) or on census data that was eight or twelve years old. As a result, coverage drops for these last conventions—from 8.9 percent and 11.2 percent of delegates for which a match could not be made in 1940 and 1944 to 12.1 percent and 21.6 percent in 1948 and 1952, respectively.

35. Dorman, Robert L., “The Creation and Destruction of the 1890 Federal Census,” The American Archivist 71 (Fall-Winter 2008): 350–83CrossRefGoogle Scholar.

36. See Lisio, Hoover, Blacks, & Lily-Whites. Black voters in the North also began shifting into the Democratic Party around this time. Most historians believe the 1936 election was the point at which a majority of blacks voted Democratic in presidential elections. See, for example, Weiss, Nancy J., Farewell to the Party of Lincoln: Black Politics in the Age of FDR (Princeton, NJ: Princeton University Press, 1983)Google Scholar.

37. The Black-and-Tans in Mississippi would finally be replaced in 1960.

38. Woodward, C. Vann, Origins of the New South, 1877–1913 (Baton Rouge: Louisiana State University Press, 1951), 324Google Scholar.

39. This notion of respectability was widespread at the time. In 1928, for example, Horace A. Mann, a Tennessee politician, worked to recruit Democrats for Republican presidential nominee Herbert Hoover in the South—the so-called Hoovercrats. Mann stated the following in discussion with a Hoovercrat leader in Florida: “We are going to have a respectable party in every southern state and that means the elimination of the negro in so far as political activities and office holding is concerned.” Quoted in Lisio, Hoover, Blacks, & Lily-Whites, 116–17.

40. This section is based on Ginzl, David J., “Lily-Whites versus Black-and-Tans: Mississippi Republicans during the Hoover Administration,” Journal of Mississippi History 42 (1980): 194–211Google Scholar; McMillen, Neil R., “Perry W. Howard, Boss of Black-and-Tan Republicanism in Mississippi, 1924–1960,” Journal of Southern History 48 (1982): 205–24CrossRefGoogle Scholar.

41. Frederick Sullens, “Editorial,” Jackson Daily News, March 27, 1929.

42. McMillen, “Perry W. Howard,” 218.

43. Frederick Sullens, “Editorial,” Jackson Daily News, December 15, 1928.

44. Indeed, Howard would be acquitted again in a second trial in the spring of 1929. Again, twelve white jurors found in his favor. See Ginzl, “Lily-Whites versus Black-and-Tans,” 201; McMillen, “Perry W. Howard,” 220.

45. To be sure, the Whiteness Index remains an indirect measure of the racial division within the broader state party at any given moment in time. However, changes in the index do correlate with specific historical events. For example, the Whiteness Index for North Carolina shows considerable black representation at the national convention, up through 1896. Subsequently, black representation drops dramatically until it reaches zero in 1908. This decline is in line with the history of the Republican Party in North Carolina, which represented a mixed-race coalition until the introduction of Jim Crow disfranchisement laws in 1900. In 1902, Senator Jeter R. Pritchard (R-NC) began a campaign to exclude blacks entirely from the party, resulting in all white delegations from 1908 onwards. For Arkansas, the Whiteness Index shows a peculiar development in which black representation declines to zero in 1920, but subsequently recovers to around 20 percent for most of the rest of the period. This temporary drop coincides with the short-term takeover of the Arkansas GOP by Lily-White leader Gus Remmel in 1914. At the state convention in 1920, Remmel successfully blocked blacks from being nominated to the national convention, resulting in an all-white national delegation. However, Remmel died shortly after the 1920 elections, and Black-and-Tans leaders succeeded in regaining control of the state party and began sending black delegates to the national committee again from 1924 onwards. Finally, in Florida, a similar image of the decline and subsequent reappearance of black representation at the national convention emerges: through 1928, blacks made up at least some part of Florida's national convention delegation. In 1932 and 1936 black representation was zero, yet by 1940 black representation reappeared. The cause of this reemergence of black delegates was the repeal of the poll tax in 1937. In subsequent years, black activists in the state began to mobilize black citizens to register to vote again, and sent a rival delegation to the 1940 national convention. As part of a compromise, a small number of black delegates were seated at this and subsequent conventions.

46. We note that our use of “whiteness” as an analytic concept to explain political behavior is not new. In contemporary American politics, as Ashley Jardina notes, “a large portion of whites actively identify with their racial group and support politics and candidates that they view as protecting whites’ power and status.” Jardina, Ashley, White Identity Politics (Cambridge, UK: Cambridge University Press, 2019)CrossRefGoogle Scholar, i. The 2016 presidential election has been viewed as one in which whiteness played a significant role, as Donald Trump successfully appealed to white consciousness—and threats posed to whites from nonwhites—for political gain and, arguably, to achieve election. Trump, however, did not invent the use of whiteness in presidential politics. George Wallace (1968) and Pat Buchanan (1992), for example, both used whiteness to enhance their presidential bids and score some electoral successes—they just did not ride whiteness politics all the way to the White House, as Trump did. See John Sides, Michael Tesler, and Lynn Vavreck, “Donald Trump and the Rise of White Identity Politics.” Paper presented at the “2016 U.S. Presidential Election: Tumult at Home, Retreat Abroad?” conference, Mershon Center, Ohio State University, November 2017.

47. These various party-strength measures are based on David, Paul T., Party Strength in the United States, 1872–1970 (Charlottesville: University of Virginia Press, 1972)Google Scholar; Paul T. David and William Claggett, Party Strength in the United States: 1872–1996 (Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2008-09-10), https://doi.org/10.3886/ICPSR06895.v1

48. Note that our results hold if we lag Whiteness Index two or three time periods, or use the average of the last three time periods.

49. See Kousser, The Shaping of Southern Politics, 238–40. Note that we choose to begin our count of disfranchisement provisions with the first “high tide” (per Kousser) in the disfranchisement movement, which extended from 1888 to 1893. This ignores a small number of provisions enacted in Georgia, South Carolina, and Virginia before then (between 1871 and 1882).

50. The U.S. Supreme Court found the grandfather clause to be unconstitutional in Guinn v. United States, 238 U.S. 347 (1915).

51. The mean of the WI (Lagged) for the entire period (N = 242) is 71.51, and the standard deviation is 23.83. The mean of the WI (Lagged) under disfranchisement (N = 171) is 75.18, and the standard deviation is 24.91. The mean of the WI (Lagged) prior to disfranchisement (N = 71) is 62.67 and the standard deviation is 18.34.

52. The mean of GOP presidential vote for the entire period (N = 242) is 29.07 and the standard deviation is 16.01. The mean of GOP presidential vote under disfranchisement (N  = 171) is 25.25, and the standard deviation is 15.86. The mean of GOP presidential vote prior to disfranchisement (N = 71) is 38.27, and the standard deviation is 12.26.

53. See, for example, Lublin, The Republican South, 33–65.

54. See Mickey, Paths Out of Dixie, 25. The Outer South is sometimes called the Peripheral South. See Black and Black, The Rise of Southern Republicans, 17.

55. Mickey, Paths Out of Dixie, 25. V. O. Key makes a similar argument about the Deep South states, via a discussion of the “black belt,” those counties in which blacks made up a substantial proportion of the population. Key states: “It is the whites of the black belts who have the deepest and most immediate concern about the maintenance of white supremacy.” V. O. Key, Jr., Southern Politics in State and Nation (New York: Knopf, 1949), 5.

56. Key, Southern Politics, 9.

57. Key, Southern Politics, 669.

58. We find similar results when looking at the marginal effects and plotting the predicted GOP support for congressional vote and governor vote. See Table A1 for the OLS regression results. Overall, the model fits are better for the federal elections (president and Congress) than the state elections (governor) in both the Outer South and the Deep South.

59. The mean of the WI (Lagged) for the Outer South under disfranchisement (N = 94) is 87.78, and the standard deviation is 13.78. Thus, we actually measure the high end up through 100 (whereas a 1 standard deviation move above the mean would extend beyond 100). The mean of the WI (Lagged) for the Deep South under disfranchisement (N = 77) is 59.80, and the standard deviation is 26.82.

60. We also find a similar relationship when looking beyond GOP vote totals and focusing on a particular set of GOP electoral “wins.” Note, though, that the Republican Party only began winning consistently in the South after our period of analysis here. But there were enough GOP wins in presidential elections through 1956, by state, to perform a systematic analysis. We thus replicated our presidential-election model, but instead of GOP vote percentage as the dependent variable, we specified a binary variable for whether the Republican Party candidate won the election in the state (1) or not (0). A linear probability model reveals the same relationship between the WI (Lagged), conditioned by disfranchisement, and the likelihood of the GOP winning: negative before disfranchisement and positive after. The average marginal effects, however, do not meet standard levels of statistical significance (p < .173 for the pre-disfranchisement era and p < .175 for the post-disfranchisement era). See Figure A7 for a visual illustration. Digging deeper, we do find, though, that the predictive margins reach statistical significance (p < .05) for WI (Lagged) values of 60 percent and greater (see Table A2). In effect, for every 5 percentage point increase in the WI (Lagged), the probability of a Republican winning a Southern state in the period after the introduction of disfranchisement laws increases by 0.79 percentage point. At a WI (Lagged) of 100 percent in this period, a Republican presidential candidate has a 15.13 percent chance of winning a Southern state. While not a huge likelihood, the ability of the GOP to win in the South during this era—when the party was long seen by generations of Southerners as the “black party” and the party that initiated the “War of Northern Aggression”—is meaningful. And it was achieved by the Republican Party going Lily-White.

61. Election forecasts based on statistical models gave Trump little chance to win in his matchup with Democratic nominee Hillary Clinton—anywhere from 29 percent (FiveThirtyEight) to 15 percent (New York Times) to less than 1 percent (Princeton Election Consortium). Election forecasts based on betting markets yielded similar, low odds—11 percent (PredictWise). For these various forecasts and predictions, see Josh Katz, “Who Will Be President?” New York Times, November 8, 2018, https://www.nytimes.com/interactive/2016/upshot/presidential-polls-forecast.html.

62. See, for example, Cavari, Amnon, Powell, Richard J., and Mayer, Kenneth R., eds., The 2016 Presidential Election: The Causes and Consequences of a Political Earthquake (Lanham, MD: Lexington Books, 2017)Google Scholar; Ceaser, James W., Busch, Andrew E., and Pitney, John J. Jr., Defying the Odds: The 2016 Elections and American Politics (Lanham, MD: Rowman and Littlefield, 2017)Google Scholar; Sabato, Larry J., Kondik, Kyle and Skelley, Geoffrey, eds., Trumped: The 2016 Election that Broke All the Rules (Lanham, MD: Rowman and Littlefield, 2017)Google Scholar; Sides, John, Tesler, Michael, and Vavreck, Lynn, Identity Crisis: The 2016 Presidential Campaign and the Battle for the Meaning of America (Princeton, NJ: Princeton University Press, 2018)Google Scholar.

63. Trump won 304 electoral votes overall. While he won Texas, Trump received only 36 of the state's 38 electoral votes, as two members of the Electoral College who were pledged to vote for him did not do so. These “faithless electors” cast their votes for John Kasich and Ron Paul instead.