The Development and Persistence of Ethnic Voting*
Published online by Cambridge University Press: 02 September 2013
Mass immigration ended fifty years ago, but national origins continue to be a salient dimension in many people's perceptions of themselves and of others. Where this salience is widespread, ethnicity plays a major role in politics. Ethnicity is often an important independent variable in voting behavior. “Ethnic voting,” as I shall call it, has two manifestations. (1) Members of an ethnic group show an affinity for one party or the other which cannot be explained solely as a result of other demographic characteristics. Voters of Irish descent, to take a familiar example, are more likely than other voters of similar economic status to be Democrats. (2) Members of an ethnic group will cross party lines to vote for—or against—a candidate belonging to a particular ethnic group.
- Research Article
- Copyright © American Political Science Association 1965
1 For a recent statement of this theme see Glazer, Nathan and Moynihan, Daniel Patrick, Beyond the Melting Pot (Cambridge, MIT Press and Harvard University Press, 1963)Google Scholar. “Ethnic consciousness” or “ethnic salience” exists when: (1) many people think of themselves, and are regarded by others, as members of a particular nationality group; and (2) such classification is salient. The two aspects of ethnic consciousness reinforce each other.
2 Conflict among ethnic groups is a central topic in descriptions of politics in the Northeast; see, for example, Lockard, Duane, New England State Politics (Princeton, Princeton University Press, 1959)CrossRefGoogle Scholar. For treatments of ethnicity in personnel appointments see Lowi, Theodore J., At the Pleasure of the Mayor (New York, 1964)Google Scholar; and Moynihan, Daniel Patrick and Wilson, James Q., “Patronage in New York State, 1955–1959,” this Review, Vol. 58 (06, 1964), pp. 296–301Google Scholar. For a discussion of the social and political consequences of ethnic politics see Wolfinger, Raymond E., “Some Consequences of Ethnic Politics,” in Zeigler, Harmon and Jennings, Kent, eds., The Electoral Process (Englewood Cliffs, N. J., 1966)Google Scholar.
3 A good deal of data to support these propositions will be found in this article. For additional evidence see Campbell, Angus, Gurin, Gerald and Miller, Warren E., The Voter Decides (Evanston, 1954), pp. 77–79Google Scholar; Banfield, Edward C. and Wilson, James Q., City Politics (Cambridge, Harvard University Press, 1963), pp. 230–231Google Scholar; and Davidowicz, Lucy S. and Goldstein, Leon J., Politics in a Pluralist Democracy (New York, Institute of Human Relations Press, 1963)Google Scholar. Ethnicity is only one variable in voting behavior. This article concerns secular trends in its importance. Many short-term influences on voting decisions that also affect its importance are not discussed here; this omission should not be interpreted as an implicit assertion that these short-range factors are not relevant.
4 I will use the terms “ethnic group” and “nationality group” interchangeably to refer to individuals whose national origins set them apart from the predominantly Protestant old American society.
5 The assimilation theory is most clearly and explicitly stated in Dahl, Robert A., Who Governs? (New Haven, Yale University Press, 1961), pp. 34–36Google Scholar.
6 Data on New Haven are from an intensive study of that city's politics conducted primarily by Dahl, William H. Flanigan, Nelson W. Polsby, and Raymond E. Wolfinger. It is reported ibid.; and in Polsby, , Community Power and Political Theory (New Haven, Yale University Press, 1963)Google Scholar, ch. 4; and Wolfinger, , The Politics of Progress (New Haven, Yale University Press, forthcoming)Google Scholar.
7 The sample was randomly chosen from voting lists. Sampling procedures are described in greater detail in Dahl, pp. 338–339. The survey was directed by William H. Flanigan.
8 This figure undoubtedly underrepresents the number of Irish in New Haven, since 83 % of all Irish immigrants came to the United States before 1900; see U. S. Bureau of the Census, Statistical Abstract of the United States: 1956 (Washington, 1955), p. 95Google Scholar.
10 This table, taken from Banfield and Wilson, op. cit., p. 39, shows the proportion of first- and second-generation Americans in cities with more than 500,000 population in 1960:
There are some regional differences in the national origins of these ethnic populations; the Western cities tend to have more Scandinavians, for example. On the other hand, San Francisco, for one, has sizable Irish and Italian groups.
11 For similar observations about regional differences see Glazer and Moynihan, op. cit., pp. 10, 250. They mention an alternate explanation: residents of western cities moved there after living in the East, hence they are less conscious of their European origins.
This point about regional differences applies only to concern about national origins among whites and does not deal with racial prejudice.
13 This shared experience is thought to be the reason for the inclusion of Jews in San Francisco high society: “the early Jews in the West could boast that they were pioneers among pioneers” (ibid., p. 65).
14 For a description of these social patterns see Handlin, Oscar, The Uprooted (Boston, 1952)Google Scholar, chs. 7, 8.
15 An excellent description of ethnic politics that expresses the style and flavor of these negotiations may be found in Whyte, William F., Street Corner Society, enlarged edition (Chicago, University of Chicago Press, 1955)Google Scholar.
16 Lowi suggests that, unlike success in other fields, political eminence has a strong impact on ethnic perspectives: “Success in economic fields ia highly individualized; … there is relatively little group symbolization of success. In contrast, political success, particularly in the big cities, is symbolized very highly in group terms” (op. cit., p. 46).
17 The present Mayor of New Haven was one of the first politicians to make use of professional sample surveys of the electorate. Data from the surveys he commissions are analyzed by ethnic group, rather than income, occupation, or education.
18 McConnell, John W., The Evolution of Social Class (Washington, American Council on Public Affairs, 1942), p. 214Google Scholar. The data are from the Sample Family Survey conducted by the Yale Institute of Human Relations in 1931–33.
19 Myers, Jerome K., “Assimilation in the Political Community,” Sociology and Social Research, Vol. 35 (1951), pp. 175–182Google Scholar. Myers estimated the number of Italians in various categories of municipal jobs from the names in city directories and manuals. The Republicans lost control of City Hall in 1932 to John W. Murphy, a Democrat of the old school. Murphy, who stayed in office until 1946, usually carried the Italian wards, but by somewhat smaller margins than he received in other working-class neighborhoods.
As these findings suggest, there are impediments to a rational strategy of ethnic politics. (1) Party leaders may refrain from cultivating ethnic groups out of prejudice. In much of the East Coast the exclusiveness of Yankee Republican politicians aided the Democrats' proselytizing of immigrants. (2) Party leaders may be reluctant to share political spoils with “outsiders.” This seems to have been true of many Irish Democrats. Until a very few years ago, one Democratic ward organization in a New Haven Irish neighborhood would not let Italians participate in any form of campaign activity. (3) There may be principled objections to making appointments on the basis of ethnicity rather than other forms of merit. For examples of this attitude see Wilson, James Q., The Amateur Democrat (Chicago, University of Chicago Press, 1962), pp. 283–288Google Scholar. It is quite possible, however, that such scruples are forgotten when the men who hold them actually attain power. Moynihan and Wilson describe the appointment policy followed by the young, liberal, intellectual staff of the newly elected New York Governor Averell Harriman: “… great efforts were extended to ‘recognize’ certain groups and careful records were kept of the racial and religious identity of the appointees” (op. cit., p. 296).
20 See Myers; and Dahl, op. cit., pp. 43–44. In the 1950s, during a period of Democratic success, Italians held slightly more than their share of municipal elective offices.
21 For other data casting doubt on the notion of Italian Republicanism see Lockard, op. cit., pp. 210, 305–319; Lubell, Samuel, The Future of American Politics (Garden City, N. Y., 1956), pp. 225–226Google Scholar; Davidowicz and Goldstein, pp. 11–12, 30–32; Berelson, Bernard R. et al. , Voting (Chicago, University of Chicago Press, 1954), p. 62Google Scholar; and Huthmacher, J. Joseph, Massachusetts People and Politics 1919–1932 (Cambridge, Harvard University Press, 1959), pp. 173, 179–184, 252, 260–261CrossRefGoogle Scholar.
22 Jews in New Haven, like those elsewhere, tended to vote Republican until the New Deal; see, e.g., Glazer and Moynihan, op. cit., pp. 168–169.
23 Except for the 1928 election, when the 10th was 18% ahead of the city in its support of Al Smith, and the 1932 and 1936 elections, when Roosevelt ran about 10% ahead of the city-wide vote there. In other mayoralty and presidential elections prior to 1939, differences were generally minor; see Dahl, op. cit., pp. 48–50.
24 President Roosevelt's 1940 “stab-in-the back” speech and World War II are supposed to have cost him some Italian votes in 1940 and 1944. Whatever the extent of this loss, it seems to have been recouped in the 1948 election; see Lubell, op. cit., 225–226. For differing assessments of the impact of Roosevelt's speech, see Whyte, op. cit., pp. 230–231; and Key, V. O., Public Opinion and American Democracy (New York, 1961), pp. 271–272Google Scholar.
25 The success of the Socialist party in nearby Bridgeport was due to different causes. Capitalizing on public revulsion at corrupt “double machine” collusion between the two major parties, the Socialist leader, Jasper McLevy, won the mayoralty there in 1933 and stayed in office for 24 years.
27 Except for the 1941 and 1943 elections, when Celentano was not a candidate, and 1947, when both candidates were Italians.
28 The 1st Ward was excluded because most of its residents are Yale students.
29 Dahl has an interesting table comparing the Negro 19th Ward to the Italian 11th. The two wards have similarly low occupational, income and educational levels, yet the 19th is overwhelmingly Democratic and the 11th is very Republican. This is the reverse of their partisan affinities in 1930 (Dahl, op. cit., p. 57).
30 Cf. Berelson et al., op. cit., pp. 67–72; and Campbell, Angus et al. , The American Voter (New York, 1960)Google Scholar, ch. 12.
31 This may be a dubious assumption, although it is not crucial to my argument. There are indications that the previous identification of many immigrants was not with the old nation, but the old village or old province. In this view, it was not until the immigrants saw that Americans classified them by nationality that they themselves developed some sense of belonging to a nationality group.
One could also argue that, whatever the locale of his previous identity, the immigrant's first impulse was to forget this old identity and become an American, but that he was forced back into ethnic consciousness by old-settler prejudice. Recurring nativist phenomena like Know-Nothingism, the Ku Klux Klan, Prohibition, and the end of mass immigration probably increased many ethnics' self-consciousness. See, e.g., Hofstadter, Richard, The Age of Reform (New York, 1955), p. 297Google Scholar.
32 Cf. the discussion of “political proximity” in Campbell et al.: “Groups as perceived objects may be located according to their proximity to the world of politics … at the individual level: as perception of proximity between the group and the world of politics becomes clearer, the susceptibility of the individual member to group influence in political affairs increases” (op. cit., p. 311; emphasis in original).
33 Cf. Handlin, “The immigrant might sometimes read an article on such a matter in his newspaper but was less likely to be persuaded by any intrinsic ideas on the subject than by the character of the persuader” (op. cit., p. 211). Such persuaders included, in addition to overt party workers, priests and the types of immigrant leaders discussed earlier.
35 Source: data from the 1956 SRC study reported in Greenstein, Fred I., “The Changing Pattern of Urban Party Politics,” The Annals, Vol. 353 (05, 1964), pp. 8–9Google Scholar.
37 Whyte's discussion of bribery indicates that cash can have a powerful, if temporary, distracting effect on ethnic loyalties.
38 Cf. Lowi, “The representation of a new minority in places of power occurs long after it has reached considerable size in the population and electorate” (op. cit., p. 39). Elmer E. Cornwell, Jr. reports that in Providence, Rhode Island, “Members of a new group are not likely to appear as ward committeemen at all until some three decades after their first arrival in substantial numbers”; see his “Party Absorption of Ethnic Groups: The Case of Providence, Rhode Island,” Social Forces, Vol. 38 (03, 1960), p. 208Google Scholar.
See also Lubell: “The key to the political progress of any minority element in this country would seem to lie in just this success in developing its own middle class” (op. cit., p. 79). Lubell does not discuss specifically the importance of candidacy for major office, which is the key point in any group's mobilization.
39 Huthmacher has an interesting description of this competitive bidding process in Massachusetts (op. cit., pp. 119–126). His discussion makes clear the dangers of such strategies because of the jealousies aroused when newer groups are recognized.
40 Prior to Celentano's nomination the Italian community in New Haven was reported to be fragmented; see McConnell, op. cit., pp. 159–160.
41 For present purposes “major office” may be loosely but serviceably defined as any public elective office which is the central prize in a political system: mayor, governor, perhaps U. S. senator, and, of course, the presidency. Candidacy for minor office does not seem to produce so much ethnic impact, at least where candidates for such positions appear on the ballot below more important ones. This is particularly true in states like Connecticut where one can vote for an entire party slate with one choice. Such arrangements discourage split-ticket voting; see Campbell et al., op. cit., pp. 275–276.
42 Lowi, who has analyzed top-level mayoral appointments in New York City from 1898 to 1958, reports similar findings: “It has been the role of the minority party in New York to provide a channel of mobility for new ethnic groups …. The dominant Democratic organizations of the twentieth century have made efforts to attract the immigrants, but the minority Republicans made greater use of top patronage for these purposes” (op. cit., pp. 37–39).
43 Eldersveld, Samuel J., “The Influence of Metropolitan Party Pluralities in Presidential Elections Since 1920: A Study of Twelve Cities,” this Review, Vol. 43 (12, 1949), p. 1196Google Scholar.
44 For a statement of this point of view see Dahl, op. tcit., pp. 34–36, 59–62.
45 Berelson et al., op. cit., pp. 61–71; Campbell, Gurin and Miller, op. cit., p. 71; Campbell et al., op. cit., ch. 12; and Greer, Scott, “Catholic Voters and the Democratic Party,” Public Opinion Quarterly, Vol. 25 (Winter, 1961), pp. 611–625CrossRefGoogle Scholar. Berelson's study found that Catholicism was a stronger independent variable than socio-economic status in voting behavior. Like ethnicity, its importance is subject to much short-term variation. For example, Catholicism was much more important in the 1960 presidential election than in 1956; see Converse, Philip E. et al. , “Stability and Change in 1960; a Reinstating Election,” this Review, Vol. 55 (06, 1961), pp. 269–280Google Scholar.
46 On the “disappearance” of the Germans see Glazer and Moynihan, op. cit., p. 311; and Moynihan and Wilson, op. cit., pp. 299–300.
47 Greer, op. cit., p. 621; and Campbell, Gurin and Miller, op. cit., p. 79. The latter study found that the partisan difference between Catholics and Protestants was as great in the fourth generation as in the first.
48 All ethnic groups are not, of course, equally represented in various occupations; see, e. g., Glazer and Moynihan, op. cit., pp. 317–324.
49 This assumes that the direction of ethnic influences will favor the Democratic party. When ethnic pressures are pro-Republican, as in the case of New Haven Italians, the problem of predicting the political consequences of social mobility becomes more complicated.
50 Berelson, et al., op. cit., p. 65.
51 Barber, James A. Jr., “Social Mobility and Political Behavior,” unpublished dissertation, Stanford University, 1965Google Scholar. See also Berelson et al., op. cit., p. 91.
53 Berelson et al., op. cit., p. 90. For discussion of the varying strength and characteristics of the relationships between social class and voting behavior, see Campbell et al., op. cit., ch. 13; and Eulau, Heinz, Class and Party in the Eisenhower Years (New York, 1962)Google Scholar. As these books make clear, associations between class and party are mediated by a number of other personal and historical variables. One such is the difference between social class as measured by objective indicators like income, and subjective class, i.e., what the individual considers his class position to be. When middle-class people identify with the working class their political attitudes and behavior tend to resemble those of members of the working class. Possibly middle-class ethnics are more likely to consider themselves working class than are middle-class Yankees. This suggests one mechanism that would modify the political impact of social mobility.
54 This proposition is stated in Berelson et al., op. cit., p. 74; and is supported by data in their ch. 6.
Ethnic groups may differ in their willingness to move from old urban habitats. Glazer and Moynihan report that Italians in New York, unlike some other groups, seem to remain, generation after generation, in the same areas where they first settled. The areas of Italian concentration in 1920 and 1960 are substantially the same except where land clearance has displaced people (op. cit., pp. 186–187).
55 Immigration continues to provide a diminished but by no means negligible fresh supply of ethnics. Most of the 2,500,000 people who entered the United States as immigrants from 1950 to 1959 probably settled in neighborhoods inhabited by earlier arrivals from their respective countries.
57 Greer, op. cit. p. 621. Even in the suburbs, however, Catholicism is a potent independent variable in voting behavior.
59 See Dahl, op. cit., pp. 49–51.
60 Key, V. O. Jr., and Munger, Frank, “Social Determinism and Electoral Decision: the Case of Indiana,” in Burdick, Eugene and Brodbeck, Arthur J., eds., American Voting Behavior (Glencoe, Ill., 1959), pp. 281–299, at p. 457Google Scholar n. Similar findings for Ohio are reported in Key, V. O. Jr., “Partisanship and County Office: The Case of Ohio,” this Review, Vol. 47 (June, 1953), pp. 529–531Google Scholar; and Flinn, Thomas A., “The Outline of Ohio Politics,” Western Political Quarterly, vol. 13 (September, 1960), pp. 702–721CrossRefGoogle Scholar.