¹ See its final report: Commission on Party Structure and Delegate Selection, Mandate for Reform (Washington, D.C.: Democratic National Committee, 1970)Google Scholar.

² For the chronology, see White, Theodore H., The Making of the President, 1972 (New York: Atheneum, 1973)Google Scholar.

³ Among the 22 states with presidential primaries, there were, in fact, 18 recognizably different primary plans. A heavy majority of states did use Districted regulations, but the variety within that designation was such that no state needed to apologize for a lack of institutional creativity. There were Districted states which put their delegates in Congressional Districts (Wisconsin), State Senatorial Districts (Pennsylvania), and Counties (New Jersey). There were Districted states which apportioned these delegates by prior Democratic vote (Ohio), by population (Maryland), or by a combination of the two (New Hampshire). There were Districted states which allowed persons elected to other offices (e.g., the Democratic State Committee) to pick a small share of the delegates on a Winner-Take-All basis (Florida) or on a Proportional basis (New York), Districted states which allowed delegates who were elected under Winner-Take-All (Maryland) or Proportional (Pennsylvania) rules, and Districted states which reserved seats for persons not connected with that election at all (the Democratic National Committeemembers from Nebraska). There were Districted states in which voters could pick individual delegates from different slates pledged to different presidential candidates (West Virginia), Districted states in which voters could pick the delegates but could not tell whom they were pledged to (Indiana), and Districted states in which voters could decide to whom the delegates should be pledged but could not pick the delegates (Tennessee).

Three more states adopted Proportional plans, incorporating a percentage cut-off (5 per cent in Michigan), a candidate cut-off (the top two finishers in New Mexico), or both (a 15 per cent cut-off and a four-man cut-off in North Carolina). Finally, four states used ‘simple’ Winner-Take-All rules, but even these varied according to whether the candidates had greater (Rhode Island) or lesser (California) control over who could stand as a delegate for them. Nine other states reserved some (usually small) share of their delegations for the statewide front-runner on a Winner-Take-All basis, a share that ran from 2 delegates (Oregon) to 38 (Ohio).

A good example of the intricacies which could arise from an imaginative mix of the three basic plans was Pennsylvania, where: (a) 137 delegates were apportioned among State Senatorial Districts according to a formula based 50 per cent on Democratic vote and 50 per cent on population: (b) 27 more delegates were chosen, at-large, by the first 137 delegates (who were elected on a Districted basis), by standards that were roughly (but only roughly) Proportional to the presidential preferences of this first 137; and (c) 18 additional delegates were chosen by the (independently elected and Party-Districted) State Democratic Committee, and were then pledged to presidential candidates in (very) rough approximation to those candidates' support among the first 137 (Senate-Districted) delegates. To get this process moving, the Pennsylvania citizen had to mark his ballot at least three times—once at the top of the ballot to indicate his presidential preference, and then two to four times at the bottom to elect delegates who might (but need not) be committed to that presidential preference.

⁴ For states which asked their citizens to mark the ballot twice—once at the top for a presidential preference, and once at the bottom for district delegates—we have used the top (presidential preference) line as the “recorded vote.” This insures comparability across the three basic sets of primary rules, since both Winner-Take-All and Proportionality demand a statewide vote. It has the additional research advantage of underestimating our findings. If we were, instead, to construct an artificial “statewide vote” by averaging the votes for district delegates whenever possible, this second figure would magnify our estimates of the impact of Proportionality and, especially, of Districting. Thus the top ballot line, which is more useful for comparison, is also more likely to provide a tough test of our argument.

The three major primary plans to which this vote was applied were defined as follows: (a) WINNER-TAKE-ALL: All of the state's delegates were awarded to the candidate who got the largest popular vote (on the top ballot line); (b) PROPORTIONAL: All of the state's delegates were divided among the candidates who got at least 15 per cent of the popular vote, in proportion to their share of that vote. The 15 per cent figure will be required in all Proportional primaries in 1976, and seems likely to be continued in the future. (But see footnote 9 for an estimate of the effect of substituting a 5 per cent cut-off.); (c) DISTRICTED: All of the state's delegates were placed in Congressional Districts and apportioned according to a formula based 50 per cent on prior Democratic vote and 50 per cent on population; the candidate with the largest vote in each district then received its delegates. This was the most common version of Districting in 1972 (all but four Districted states adopted it), and most of the other variants in use that year have been outlawed for 1976.

In view of the nearly infinite minor variations that are possible within these three arrangements (see footnote 3), it might seem that the general impact of Winner-Take-All, Proportionality, or Districting could be substantially altered by manipulating subrules within any one plan, rather then by having to scrap the plan itself. In fact, this does not appear to be true. Some of these variations actually have no effect at all, and most have measurable but trivial effects when compared with the differences between the three basic primary laws. An example of the former is the provision which permits some states' elected delegates to choose a small additional share of their delegations. As long as they choose by Proportional rules (as most do), this has no impact on the strengths of presidential aspirants or on the power of states within the party (although it might still influence the behavior of these delegates at the national convention). An example of the latter (clear but minimal impact) is the provision which permits some states' voters to select individual delegates from different candidates' slates. In 1972, exactly 11.5 out of 3,016 convention votes were changed by this subrule, and no candidate faced a loss/gain of more than 4 delegates through it. In the same way, the permission for states which used Districted rules to apportion their delegates either entirely by past Democratic vote or by half-Democratic-vote/half-population would have shifted no more than 2 delegates in any one state, and all the shifts (though trivial) would have reinforced our argument in the third section.

⁵ Two primaries which occurred during this period, Illinois (March 21; 170 delegates) and the District of Columbia (April 18; 15 delegates), are ignored because they could not meet both of these criteria.

⁶ A detailed example of the way this was done might be useful:

⁷ In states with a different kind of social base, the rules had a much more muted impact on delegate disbursements. Tennessee is such a slate (footnote 6), and analogous rules changes would have had little effect on the loyalties of its delegates. This interaction between primary rules and the various types of states is analyzed in the final two sections of this paper.

⁸ The actual results of these fifteen contests—Humphrey 284, Wallace 291, McGovern 401.5, Muskie 56.5, Others 59—do not, of course, match the outcomes which a consistent application of any of these three plans would have created, mainly because different states used different basic plans. Thus McGovern's ability to surpass his apparent “optimum,” and the inability of Wallace and Humphrey to attain their apparent “minimums,” are partially due to the way their varying levels of electoral support within states meshed with the type of primary rules in effect in those same states.

Humphrey and Wallace each suffered from one other rule-related problem, stemming from the way the social status of their respective supporters interacted with the rules. Because lower-income voters were most likely to fail to mark their ballots completely, and because Humphrey's votèrs were more likely to be from the lower-class parts of the party. Districted rules (with their usual two-step voting requirement) deprived him of delegates in areas where he was the favorite; conversely, McGovern, with higher-income support, picked up delegates in areas where he could do no better than third in the preference poll. In Pennsylvania alone, Humphrey lost 20 delegates (more delegates than there are in the New Hampshire primary) through this “voter fatigue;” McGovern picked up 15 of them.

Wallace's supporters were a “problem” at a much earlier stage. Unlike the more middle-class followers of George McGovern or the party/union activists of Hubert Humphrey, George Wallace's early enthusiasts were often not sufficiently organized (or perhaps, organizable) to mount delegate petition drives. Consequently, Wallace occasionally placed his name on the presidential preference line without providing delegates who would stand for him on the lines below. In Pennsylvania, he finished comfortably ahead in districts which should have delivered 16 delegates, but since they contained only 2 pledged Wallaceites, he lost 14 of these 16 to the other candidates (Humphrey, Muskie, and especially, McGovern).

⁹ If these Proportional rules had used a 5 per cent cut-off, rather than the tougher 15 per cent, Muskie's total would have continued to grow—to 113. These additional delegates would have arrived after he had withdrawn from active candidacy, courtesy of states where his name was still on the ballot. The biggest gainer from the reduced vote threshold, however, would have been the category “Others,” where a collection of long-shots (Chisholm, Lindsay, Sanford, Jackson, and Yorty) stood to amass 74 delegates. Humphrey, Wallace, and McGovern would have surrendered delegates to these others in about the same proportions, coming to rest at totals of 292, 324, and 289, respectively.

¹⁰ See its report to the Democratic National Committee: Commission on Delegate Selection and Party Structure, Democrats All (Washington, D.C.: Democratic National Committee, 1973)Google Scholar. Guideline 8(“Unit Rule”) abolishes the Winner-Take-All primary. Guideline 9.A covers Districted primaries, and Guideline 11 contains the Commission's recommendations on Proportionality. The Democratic National Committee raised the cut-off figure in the Proportionality section (from 10 per cent to 15 per cent), but otherwise accepted the Commission's guidelines as written.

¹¹ If the argument of the last two sections is correct, however, this claim (that Districting is a “mean” between Winner-Take-All and Proportionality) is true only in formal-structural terms. In terms of the plan's effect on state power within the national Democratic party, it is, instead, the most extreme case (see below).

¹² For example, Tennessee gives its first-place finisher a margin of 31 delegates under Proportional rules, compared to 49 under Winner-Take-All; thus it retains 63 per cent of its original political value—the best record among these fifteen states. Comparable figures for the remaining states are: Massachusetts 43 per cent (44/102); Oregon 41 per cent (14/34); Florida 38 per cent (31/81); West Virginia 31 per cent (11/35); Michigan 26 per cent (34/132); Rhode Island 18 per cent (4/22); North Carolina 16 per cent (10/64); Pennsylvania 14 per cent (26/182); Maryland 13 per cent (7/53); New Hampshire 11 per cent (2/18); Wisconsin 9 per cent (6/67); Nebraska 8 per cent (2/24); Indiana 5 per cent (4/76); Ohio 2 per cent (3/153).

¹³ It might be argued that our definition of a state's “political value” (the delegate margin which it can provide to its favorite) is not the ideal measure of the impact of primary rules on power within the Democratic party. One could argue that, for purposes of understanding presidential politics in the long run, the crucial consideration is not what a state can give to any particular candidate, but how much of a margin it can generate for its favorite wing of the party, i.e., for each group of candidates (the left-Democrats, the liberal-Democrats, or the conservative-Democrats) as a group.

In 1972, the question of which measure to use—the “candidate margin” or an artificial “ideological margin”—was, of course, irrelevant, because two of the first four primaries, and all of those after the fourth (Pennsylvania), featured a maximum of three serious contenders (Mc-Govern, Humphrey, or Wallace), one for each wing of the party. Collectively, they would have amassed well over 90 per cent of the delegates from the first 15 contests under any set of rules (Table 4).

Future nomination campaigns are very likely to follow this same pattern—a rapid thinning of the field until there is a champion for each major faction—because the central problem in seeking a major party nomination is the scarcity of political resources, especially activists, finances, and popularity. An aspirant must succeed in monopolizing these within his wing of the party, and must then defeat the standard-bearers of the other wings. If he fails to drive other contenders who resemble him out of the race early, he will surely find that the resource situation allows one of them to drive him out instead—witness Lindsay (eliminated early by McGovern), Muskie (isolated from his potential resource base by Humphrey), or Mills (smothered by Wallace) in 1972.

It is probably extremely fortunate that investigation of the political role of primary rules does not require the analyst to lump one candidate's voters or delegates with another's, and then to assert that they form a bloc, since this would introduce severe distortions into the results. An observer might conclude, for example, that Mc-Govern, Chisholm, and Lindsay make up one bloc, that Humphrey. Muskie. Sanford, and Jackson are in another, and that Wallace, Mills, and Yorty form a third. Yet this assumes both that the second choices of lesser candidates' voters will lie within the same bloc, and that the convention voting of minor candidates' delegates will mirror the votes of delegates pledged to the bloc's front-runner—both patently untrue assumptions. Among the “minor” candidates, Muskie attracted the greatest number of voters and delegates. But while Muskie and Humphrey appear to hold similar ideological positions, potential Muskie voters were torn between Humphrey and McGovern after their hero had suspended operations; in fact, some commentators viewed the California primary not as a contest between Humphrey supporters and McGovern supporters, but as a contest between Humphrey and McGovern for the allegiance of (former) Muskie supporters. The same picture emerges from attempts to create delegate blocs. Muskie may have most closely resembled Humphrey, and he may even have asked his delegates at the convention to follow Humphrey's leadership temporarily, yet these delegates split (roughly in half) on the floor, some staying with the Humphreyites, but many going with the McGovernites.

(It might be noted, parenthetically, that even though we advise against pretending that these “ideological blocs” are real entities, we did carve up primary voters into the three blocs outlined above, to see what effect this would have on the role of primary election plans. It had almost none: Winner-Take-All was still the optimal plan for the the large, competitive states; Proportionality was still the middle-sized, noncompetitive states' plan; and Districting still favored the smaller states, whatever their level of competition.)

¹⁴ For example, Tennessee gives its first-place finisher a margin of 49 delegates under both Districted and Winner-Take-All rules; thus it retains 100 per cent (49/49) of its original political value. Comparable figures for the remaining states are: Massachusetts 100 per cent (102/102); West Virginia 100 per cent (35/35); Oregon 100 per cent (34/34); Nebraska 100 per cent (24/24); Rhode Island 100 per cent (22/22); New Hampshire 100 per cent (18/18); North Carolina 81 per cent (52/64); Florida 73 per cent (63/81); Michigan 70 per cent (92/132); Wisconsin 52 per cent (35/67), Maryland 51 per cent (27/53); Pennsylvania 30 per cent (54/82); Indiana 29 per cent (22/76); Ohio 25 per cent (39/53).

¹⁵ By mid-1975, it appeared that at least 28 of the 32 primaries in 1976 would be Districted, and both the number of primaries and the incidence of Districting may yet increase. This is already a situation much like the pure Districted arrangements shown in Table 4, or the state power picture presented in Figure 3.

¹⁶ If more of these states should add primaries to the list in the future, Districting will provide an even more inflated edge for this segment of the party, thanks to a peculiarity in delegate apportionment which Districted primaries accentuate. Because half of all convention delegates are assigned according to a state's share in the Electoral College, because more Electoral Votes go to the smallest states than they are entitled to by population, and because small states are disproportionately Midwestern-Western, this group of states receives “extra” delegates. This is hardly unfair or unreasonable; presidential voting is done through the Electoral College, so the relevant size figure is not census population, but Electoral College “population.” Since Districted rules, however, will allow these areas (if they institute presidential primaries) to cast their (extra) convention votes as a bloc, the combination of apportionment peculiarities and the distortions involved in Districting will be additive: It will further undercut both the middle-sized (Southern-Border) states, which get no bonus delegates from the apportionment formula, and the big (Northeastern industrial) states, whose delegations are fragmented by Districted rules.

¹⁷ This support is probably not directly rooted in any mass favorability toward social engineering policies, even among Democrats. Rather, it can be explained by two other factors characterizing Democratic party politics in the plains-mountain states. First, mass electorates in these areas, compared to the rest of the nation, prefer “cleaner,” more “purist” candidates on stylistic grounds, and this tendency is reinforced by a preference for potential nominees who are not identified with the industrial Northeast. Second, a continuing party leadership structure is nearly nonexistent, so that there is no regular party organization to resist a quadrennial take over by independent, liberal activists. As a result of this combination, apparently conservative states are very likely to produce left-liberal delegates.

¹⁸ Desertions by the voters have been less regular at the senatorial and congressional levels, but the old lines are seriously breached there, too. While this can hardly be chalked up to presidential primary rules, it can also hardly be slowed (much less reversed) by primary arrangements which have the net effect of favoring the Republicans at the presidential level.

¹⁹ This seems to suggest that those who desire both to obtain primary rules reform and to avoid doing serious damage to the party's electoral prospects should back Proportionality, and so they should if these were the only factors in the analysis. Many (perhaps most) reform advocates, however, are interested in the cause of rules reform because they believe it will lead to more liberal policy positions by the party's presidential nominees, i.e., because it will create a more change-oriented party. Judged by these standards. Proportionality is the least attractive primary plan—less attractive than Districting, which promises more liberal nominees and reform, but a decreased likelihood of winning elections; and less attractive than Win-ner-Take-All, which promises more liberal nominees, no reform, and an increased probability of winning elections.

²⁰ There is a growing literature in the field of American party politics which is related to the same concerns as our final section—namely, large-scale changes in electoral coalitions, and their associated shifts in public issues and candidate orientations. This concern with what have come to be called “realignments” had its roots in the work of V. O. Key, most particularly in Key, V. O. Jr., “A Theory of Critical Elections,” Journal of Politics, 17 (02, 1955), 3–18 CrossRefGoogle Scholar.

By the mid-1970s, the concepts which Key first pulled together had ramified (and rarefied) to the point where a student of American politics can find support in them for nearly any view of the future which strikes his fancy. There are those who foresee the breakup of old political alignments and the forging of a protracted Republican era, e.g., Phillips, Kevin P., The Emerging Republican Majority (New Rochelle, N. Y.: Arlington House, 1969)Google Scholar, but there are those who see the same breakup with a recoalescence around the national Democratic party instead, e.g., Davis, Lanny J., The Emerging Democratic Majority (New York: Stein & Day, 1974)Google Scholar. There are those who forecast the disintegration of old arrangements without any subsequent realignment, and who view this as a “new sophistication,” e.g., DeVries, Walter and Tarrance, Lance Jr., The Ticket-Splitter (Grand Rapids, Mich.: Eerdmans Publishing, 1972)Google Scholar, but there are those who predict disintegration with-out realignment, and who interpret it as the “new anomie,” e.g., Burnham, Walter Dean, Critical Elections and the Mainsprings of American Politics (New York: W. W. Norton, 1970)Google Scholar. There are those who think that a new majority coalition is waiting to be born, but has not yet been called into existence, e.g., Lubell, Samuel, The Hidden Crisis in American Politics (New York: W. W. Norton, 1970)Google Scholar and there are those who think that it is already in existence, but is waiting to be claimed by one party or the other, e.g.. Scammon, Richard M. and Wattenberg, Ben J., The Real Majority (New York: Coward, McCann, 1970)Google Scholar.

The only threads common to all of these studies are the “gut feeling” that “something's happening,” a willingness to make rather sweeping assertions about what that “something” is, and a belief that the social forces which are loose at the moment are so strong that institutional factors are inconsequential barriers in their path.

We have, in our final section, laid out our own set of realignment-type scenarios, each of which, we feel, has some plausibility. Our own suspicion about the realignment controversy, however, is that the state of the literature mirrors the state of its subject matter, i.e., that there has been no new coalitional movement in American society substantial enough to make one of these perspectives true, and the others false. Our view, obviously, is that changes in the rules of the game at this point are particularly likely to facilitate the resolution of social instabilities along what-ever lines those new rules favor. This is, however, merely a subproposition of our general belief that there will always be continued, minor movements in American society in nearly any direction you can imagine, and that a major determinant of whether these develop into more far-reach-ing changes in the political climate is whether the political rules of the game facilitate them, or whether they run at cross-currents to them.

²¹ Democratic conventions dominated by conservative coalitions would be a highly ironic result of reform. The rules alterations which led to a multiplication of primaries were motivated by a feeling that the pre-reform party was not sufficiently aimed at facilitating change. Thus if right-leaning coalitions subsequently became more frequent, or if such coalitions were necessary in order to win elections under the new rules, reform would have seriously miscarried. If either of these conditions were to accompany Districting—and we have argued that one or the other is quite likely to accompany it—then the pre-reform (Winner-Take-All) primary regulations could promise a greater likelihood of carrying the election, along with more generous social-welfare programs and more extensive civil rights and liberties.